International Competitiveness in Electronics. INSTITUTION Congress of the U.S., Washington, D.C

DOCU'tENT RESUME

ED 245 883 SE 044 576

TITLE International Competitiveness in Electronics. INSTITUTION Congress of the U.S., Washington, D.C. Office of Technology Assessment. REPORT NO OTA-ISC-200 PUB DATE Nov 83 NOTE 547p. AVAILABLE FROMSuperintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. PUB TYPE Reports Evaluative/Feasibility (142) EDRS PRICE MF02/PC22 Plus Postage. DESCRIPTORS Automation; *Competition; *Electronics; Employment Patterns; Engineering Education; Financial Support; Aigher Education; *Industry; *International Relations; *International Trade; Manufacturing; *Policy; Science Education; Secondary Education; Technology IDENTIFIERS *Industrial Policy; United States ABSTRACT This assessment continues the Office of Technology Assessment's (OTA) exploration of the meaning of industrial policy in the United States context, while also examining the industrial policies of several U.S. economic rivals. The major focus is on electronics, an area which virtually defines "high technology" of the 1980's. The assessment sets the characteristics of the technology itself alongside other forces that exert major influences over international competitiveness. Specific areas addressed include: electronics technology; srr,,Acture, trade, and competitiveness in the international electronics industry; quality, reliability, and automation in manufacturing; role of financing in competitiveness and electronics; human resources (education, training, management); employment effects; national industrial policies; and U.S. trade policies and-their effects. The report concludes by outlining five options for a U.S. industrial policy, drawing on electronics for examples of past and prospective impacts, as well as on OTA's previous studies of the steel and automotive induStries. A detailed summary and iltroductory comments are included. Also included in appendices are case studies in the development and marketing of electronics products, a discussion of offshore manufacturing, and a glossary of terms used in the assessment. (J1,0

*-********************************************************************** * Reproductions supplied by EDRS are the best that can be made * * from the original document. , * *************-*******p*******************,*************-:***************** Inteationa ompetitiveness In Electronics US. DEPARTMENT OF EDUCATION NATIONAL INSTITUTE OF EDUCATION EDUCATIONAL RESOURCES INFORMA.TION CENTER IER1Ci IThis document has been reproduced as recewed from the person or organaation onginatmg It Minor changes nave been made to improve reproduction quality,

Ponts of view or opinions Staled in this docd- ment do not necessarily represent official NIE position or policy.

L13. ;" CONGRESS OF THE UNITED STATES *tee* Office of Technology Assessment 41 Wastfington. D. C. 20610 Office of Technology =assessment

Congressional Board of th-. r.:41:h. Congress

MORRIS K. UDALL, Arzr:::-1,

TED STEVENS. Alaska,.

Senate House ORRIN G. HATCH E. BROV' ",JR. Utah ' California

CHARLES McC. MATHIAS. JR- ..7,":";"1-IN D. Mary land Michigan EDWARD M. KENNEDY .:`..A,R.RY WINN, Ill. Massachusetts Kansas ERNEST F.. HOLLINGS CLAYA.F.NCE E. South Carolina Ohio CLAIBORNE PELL '1..'90PER EVA Rhode Island Iowa JOHN H. G -1.717-.."N'..5

Advisory Courmif

CHARLES N. KIMBALL. Chairman JAMES C. FLETCHMt RAC) '14, Midwest Research Institute University of Pittsburgh Um, f EARL BEISTLINE S. DAVID FREEMAN University of Alaska Tennessee Palley Authority Hamlweekt & Quist CHARLES A. BOWSHER GILBERT GUDE DAVID PUTTER General Accounting Office Congressional Research Service General \P. Mors Corp. CLAIRE T. DEDRICK CARL N. HODGES LEWIS THOMAS California Land Commission University of Arizona . MemorLI Sloan-Kettering I. Cancer Center

Director

JOHN H. GIBBONS

The Technology Assessment Board approves the reh:ase of this report. The views expressed in this report are not necessarily those of the Board, OTA Advisory Council, or of individual 'members thereof. International --C in Electronics

OTA Reports are the principal documentation of formalassessment projects. These projects are approved in advance by the TechnologyAssessment Board. At the conclusion of a project, the Board has the opportunityto review the report, but its release does not necessarily imply endorsement of the results bythe Board or its individual members.

CONGRESS OF THE UNITED STATES Office of Technolnly Assessment vidsn.ngton.o C 20510 Recommended Citation: International Competitiveness in Electronics (Washington, D.C.: U.S. Congress, Office of Technology Assessment, OTA-ISC-200, November 1983).

Library of Congress Catalog Card Number 83-600610

For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 Foreword

This assessment, requested by the Senate Committeeon Commerce, Science, and Transportation, the House Committeeon Ways and Means, and the joint Eco- nomic Committee, completes a series of three reportson the competitiveness of U.S. industries. The series began with Technology and Steel IndustryCompetitive- ness and continued with U.S. industrial Competitiveness: A Comparison of Steel, Electronics, and Aufoinobiles. Today. the subject of international competitiveness hasmore visibility among the general public than ever before. it has emergedas one of the primary economic issues facing Congress. Debates over "reindustrialization" and "industrialpolicy" beginning several years ago have been renewed. Thisassessment continues OTA's exploration of the meaning of industrial policy in theU.S. context, while also examining the industrial policies of several ofour economic rivals. Electronics virtually defines "high technology" in the1980's. Th;s assessment sets the characteristics of the technology itselfa technology already ofsuch ubiq- uity that microprocessors and computers outnumber peoplein the United States alongside other forces that exert major influencesover international competitiveness. These factors range from humanresources and costs of capital to the priorities that corporate managers placeon manufacturing Technologies and the quality of their products. The report concludes by outlining fiveoptions for a U.S. industrial policy, drawing on electronics for examples ofpast and prospective impacts, as well as on OTA's previous studies of the steel and automobile industries. OTA is grateful for the assistance of the advisory panel forthis assessment, as well as for the help provided by many individuals in other parts of the Federal Government. OTA assumes full responsibility for the report.

JOHN H. GIBBONS Director

lit Electronics Advisory Pane!

Katherine Seelman, Chairperson New York, N, Y. Jack C. Acton Richard A_ Kraft Greensburg, Pa. Matsushita Industrial Co, Steve Beckman E. Floyd Kvamme AFL-CIO Apple. C.."61711Tater-inc. A. Terry Brix Geraldine McArdle Temar Ltd. Reston, rVa. Richard P. Case Charles Phipps IBM Corp. Texas Instruments, Inc. Ruth Schwartz Cowan K. M. Poole SUNY-Stony Brook Bell Laboratories William Kay Daines Benjamin t`,.t Rosen American Retail Federation Sevin .Rosen Management Co. Leonard Dietch Kate Wilhelm Zenith Radio Corp. Eugene, Dreg. Isaiah Frank Robert B. Wood Johns Hopkins University International Brotherhood of Electrical F. Willard Griffith II Workers GC International Michael Y. Yoshinu Harvard Business School Robert R. Johnson Mosaic Systems Inc.

iv 7 International Competitiveness in ElectronicsProject Staff

Lionel S. johns, Assistant Director, OTA Energy, Materials, and International Security Division Peter Sharfman, Program Manager international Security and Commerce Program

1Tt't Martha Caldwell Harris Robert R. Miller*

Contributors Robert Fisher* Phillip Otterness* Robert Rarog* Barbara Sachs*

Contractors L. W. Borgman 84 Co. Mary Ann Maguire Princeton, N.J. University of Maryland, College Park Consultant Services Institute, Inc. Mitsubishi Research Institute, Inc. Livingston, N.J. Tokyo, Japan Developing World Industry & Richard W. Moxon Technology, Inc, University of Washington, Seattle Washington, D.C., Sterling Hobe Corp. Hamilton Herman, et al. Washington, D.C. Potomac, Md. John H. Wheatley, et al. H. C. Lir University of Washington, Seattle Untverrsiry -6f-Ivta-ryta

OTA Publishinn Staff

John C. Holmes, Publishing Officer John Bergling Kathie S. Boss Debra M. Datcher Joe Henson Glenda Lawing Linda A. Leahy Cheryl J. Manning

OTAcontract p Contents

Chapter Page 1. Part A: Summary 3 Part B: Extended Summary 23 2. Introduction 55 3. Electronics Technology 67 t.trucTure and i rade in the international Electronics industry 107 5. Competitiveness in the International Electronics Industry 163 6. Manufacturing: Quality, Reliability, and Automation 217 7. Financing: Its Role in Competitiveness in Electronics 253 8. Human Resources: Education, Training, Management 301 9. Employment Effects 343 10. National Industrial Policies 377 11. U.S. Trade Policies and Their Effects 429 12. Federal Policies Affecting Electronics: Options for the United States 463 Appendix A: Glossary 505 Appendix B: Offshore Manufacturing 510 Appendix C: Case Studies in the Development and Marketing of Electronics Products 520 Index 539

vii CHAPTER 1 Part A: Summary Contents

Firiciings 5

I i i: lli*:_li,ulliptlIlll.ti l'ilSaillli ilIilltil...... :^).Lit.i.l.,:ftnlit., fliiUStr 1 1 .1' --',II.:',.1-.1-,L.;p',, ,-

15 Cr)nciusif):.! IJ

l\lartri$ ();; G./11:1/11tl'iti allf1 5 'I 11fi Ii( jIi1r)1) if 111.10.giat.lqi Cirtsuit l'ruciuct;()p. Liltv 11 t;t 1)tica!-; 72

T.)+, 1 (:(ouptit..:r ht. Value :1!:(1 11111 1.:(plip!noW. : P:\1H)ri,, 1....11;1,-, H. I ) CHAPTER 1 Part A: Summary

Is the United States in danger of losing, to ness of industries like electronics? Does the the Japanese or others, in therace to develop United States need a more coherent industrial new high-technolog; electronics products policy? ! ifth-generation computers, high-density integrated circuits, pocket televisions? Does the de- These questions and others like themare ad- cline of the American consumer electronics in- dressed in this report, which covers threepor- dustry prefigure that of semiconductorsor tions of the industry: consumer products (pri- computers? Is U.S. standing in world markets marily color television); semiconductor devices deteriorating because of poor management, such as integrated circuits; and computers. The slipshod Government policymaking,overregu- focus of the report is the United States, butcon - lation of business? Will work in automobile side: able attention goes to the electronics in- production or heavy industry be permanently dustries of Western Europe and Japan,as well replaced by high-technology jobs fewer innum- as several of the newly industrializing coun- ber and paying wages at half the level of the tries. $15 to $20 per hour earned by auto or steelworkers? To what extent can electronics stand Electronics in total employsmore than a for other technology-based U.S. industries? million and a half Americans; 1982 salesex- Which policies of the Federal Governmentare ceeded $125 billionroughly one-fifth of total inost crucial to the international competitive- U.S. durable goodsoutputandhave been

Sales Trends in the U.S. Electronics Market 120

100

C Total

60 Data processing equipment Communications equipment 40 Components

Consumer electronics

. anima.:,...'. *414.... .° . .. m.....:drm11..."r , E.

minim . jro :.....g 9. .17.., . . .. - IM mt 11.. irnp1- "- I .1 I L :''' 1- 1955 1960 1965 1970 1975 1980 Year SOURCE: Electronics Market Data Book 1982 (Washington, D.C.: Electronic Industries Association, 1982), P. 4. 12 4 International Competitiveness in Electronics

growing at nearly 15 percent per year; the sec- directly its gross domestic product, and there- tor is an export leader, with a surplus of about fore the standard of living of its citizens. $3 billion on a total trade volume of nearly $50 billion. The industry's products feed many The linkage between competitiveness and other portions of the U.S. economy. Not only employmentin the aggregate, in particular does the Nation's defense depend heavily on sectors, or in particular occupational catego- electronic technologies, but both manufac- riesis much looser. Industries can rise in turing and service industriesranging from the competitiveness while declin ng in employ- production of numerically controlled machine mentthe case in the U.S. textile industry in tools to banking and insuranceuse electronic recent years. In other cases, competitiveness products both directly and indirectly. may remain high, output may expand, but domestic employment may grow relatively slowly The competitiveness of firms and industries compared to output; this has been the case in refers to the ability of firms in one country to both the U.S. semiconductor and computer in- design, develop, manufacture, and market their dustries. Similarly, domestic employment is products competition with firms and indus- only loosely related to trends in foreign invest- tries in other countries. At several points ment or to government policies directed at con- below, shares of the U.S. market or of world trolling flows of imported goods; trade protec- markets are used as examples of trendsin in- tion has helped the employment picture in the ternational competitiveness; in fact, however, U.S. consumer electronics industry no more competitiveness is a more subtle concept. than it has in the steel industry or the automo- While market share is one possible indicator, bile industry. it is only indirectly related to competitiveness. While the competitiveness of a given sector How an industry will fare in international of the U.S. economy depends on both domestic competition depends on factors ranging from and international economic forces, the domes- technology itself, to industrial policies pursued tic contexte.g., people and institutions here, by national governments, to the human re- not overseas--generally carries the most weight sourcestechnicians to upper-level manag- in determining which industries will grow in ersavailable in a given country. In some competitiveness, which decline. As a result, cases, competitiveness is primarily a function public policies with domestic objectives exert of prices, hence manufacturing coststhem- the most influence over trends in international -selves determined by wage rates, labor produc- competitiveness. These are matters of indus- tivities, the design of both products and man- trial policy. OTA uses this .term in a neutral ufacturing processes. This is the case in con- sense to refer to tha body of regulations, laws, sumer electronics. In higher technology por- and other policy instruments that affect the actions of the industry, one firm may 7--e able to tivities of industry and the resources, including offer products that are beyond the technical human capital, that the Nation's economy de- capabilities of its rivalse.g., high-density in- pends on. The United States has not in the past tegrated circuits, advanced computer software. had a self-conscious industrial policy, in.part Where this is true, costs are less important. because it had no need for one. The lesson of From the Federal perspective, shifts in the the U.S. electronics industry, along with indus -. international competitiveness .of American in- tries like steel and automobiles that PTA has dustries have ramifications far beyond matters examined previously, is that future internation- of trade balances and foreign economic policy, al competitiveness may well depend on a more even military security. The competitive stand- coherent and consistent approach by, the Fed- ing of a nation's industries will determine quite eral Government to matters of industrial policy.

13 Ch. 1Part A: Summary 5

Principal Findings U.S. Competitiveness in Electronics This is not to say that there is little cause for concern, or that the waves of publicity given 1. 'Electronicsremains a leader among Amer- the progress made by Japan's electronics man- ican industries. High-technology firmsinclud- ufacturers over the past few years have in all ing those making microelectronic devices like cases been overdramatized. If the U.S. elec- integrated circuits and complex electronic sys- tronics industry is still strong when compared tems such as computerscontinue to be lead- to other domestic industries, its margins with ing exporters, second ,to none in technology as respect to electronics industries in other na- well as Most measures of commercial success. tions have shrunk, in some cases vanished. Although the Nation's imports of semiconduc- Moreover, the Japanese electronics industry is tor products exceeded its exports for the first one of the most productive in that nation's time in 1982 (by $160 million out of $3.8 billion economy; this high standing relative to other in imports) more than three-quarters of these domestic sectors is a major reason for the ex- imports were shipments by American-owned port strength of Japan's electronics manufac- firms; computer exports ($9 billion in 1982) far turers. In almost all categories of electronics exceed imports. productsoffice copiers and typewriters, mi-

U.S. Exports and. Imports of Computers and Equipment

r.6 0 O C 0 5

0 es 1960 1965 1970 1975 1980 1982 Year SOURCES: 1960.66 Gaps In rechnology: Electronic Computers (Paris: Organization for Economic Cooperation and Development. 1969), p. 50, 1987.81-1972, 1977, 1980, 1982 editions, U.S. Industrial Outlook, Department of Commerce. 1982U.S. Department of Commerce, Bureau of industrial Economics. 6 International Competitiveness In Electronics croelectronics, communications equipment technology is certainly less than even half a and consumer goodsthe U.S.-Japan trade bal- dozen years ago. American-owned firms mak- ance is strongly negative (see ch. 4). ing and selling semiconductor devices have faced increasingly' intense competition from 2'. Just as the competitive positions of a na- Japanese manufacturers, again primarily over hylastries will differ, with Some rising the last half-dozen years; although they have and othorSdeclining, so competitive positions lost market share both at home and abroad in within an industry like electronics will vary. Likewise, within one portion of the industry, some key productse.g., computer memory such as color television manufacturing, some chipstheir overall position remains strong. firms will at any given time be more competi- 3. It is not realistic to expect that American tive than others. semiconductor and computer firms will, in the Within the U.S. electronics industry, compet- near future and in the absence of cataclysmic itiveness in consumer products has declined changes in other parts of the world, return to precipitously since the 1960's. The Nation now the preeminent positions they held at the begin- imports many of its consumer electronic prod- ning of the 1.970's. Nor can the United States ucts, while more than 10 foreign-owned firms expect to achieve the technological and com- assemble and market television sets within the mercial leads of earlier years in other high- United States. In contrast, there are few signs technology industries. The capabilities of other of slackening competitiveness in the manufac- countries have improved; foreign electronics ture of computers, although the U.S., lead in industries have risen within their own econo-

Semiconductor Production in the United States and Japan 10.000

9,000

8,000

7.000 0 -6 6,000 l) C 25,000 E a) 4,000

3,000

2,000

1,000

0 19671968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 'I'g80 1981 1982 Year SOURCES: United Statee=1987.78A Report on the U.S. Semiconductor Industry (Washington, D.C.: Department of Commerce, September 1979j, p. 39. 197740Summaty of Trade and Tariff information: Semiconductors (Washington, D.C.: U.S. International Trade Commission Publication 841, Control No. 6 -5.22; July 1982), p. 28. 1981, 1982-1983 U.S. Industrial Outlook (Washington, D.C.: Department of Commeice, January 1983), p. 29-7. Japan-1967-80Japan Fact Book '80 (Tokyo: Dempa Publications, Inc., 1980), p. 188; Japan Electronics Almanac 1982 (Tokyo: Dempa Publications, Inc., 1982), pp. 149, 178. 1981, 1982in-Stat Electronics Reports Feb. 21, 1983, p..5.

_la Ch. 1Part A: Summary 7

mies; international economic conditions have 2. Superior product technologies may com- changed. mand premium prices in the marketplace, mak- 4. The United States can continue to be ing manufacturing costs less significant. Prod- highly competitive in electronics and other uct featuresranging from appearance to technologically driven industries, with U.S. quantifiable characteristics such as the per- firms remaining leaders in innovation, in in- formance of a computer system in running ternational trade, and in sales and profits at "benchmark" programscan contribute to home and abroad. Not only is this possible, it competitive advantage; in high-technology is necessary if the United States is to maintain fields as in low, product differentiation and its standard of living, its military security, and astute marketing can be important. if the U.S. economy is to provide well-paying Understanding customer wants and needs is and satisfying jobs for the Nation's labor force. vital to designing successful products; inte- Electronics is indispensable to a broad range grated circuits that are functionally similar, of manufacturing and service functions, from perhaps even interchangeable, may be differen- computer-aided design of the structures of of: tiated through subtle variations in perform- fice buildings to the switching of the telephones ance; advertising strategies can be built around within those buildings. claims of high quality or rapid delivery';a broad 5. Congress could take the initiative in devis- array of alternate source suppliers may re- ing programs that would actively support the assure prospective purchasers. Manufacturers electronics industry, and others of comparable of computers and peripherals. devotecon- importance. The first requisite is broad nation- siderable effort to industrial design and human al agreement on the role of high-technology factors engineering; ease of u e is vital in sell- sectors like electronics as a driving force for ing computer systems to first -time customers. future economic growth, a greater degree of Rapid technical change reates much more consensus on where the U.S. economy is now scope for product technology as a competitive heading and where it should head. The second weapon in microelectronics and computers is better understanding of how particular than in consumer electronics. For many years, pieces of legislation affect the competitiveness American semiconductor and computer man- of American industry, which in turn requires ufacturers prosper 'd by offering products that developing the capability of the Federal Gov- firms elsewhere in the world could not design ernment for analyzing the sources of compet- or build. itive strength. Industrial Policy The Role of Technology 1. OTA, takes industrial policy to be a neutral 1. One way to establish a competitive advan- term referring to the group of Federal policies tage in an industry like electronics is through that affect competitiveness, productivity, and superior technology. Better process technology. economic efficiencysometimes directly, e.g., automationcan help reduce costs. For sometimes through influences on business de- similar products, lower manufacturing costs cisions or on individuals. Industries rise and permit lower selling prices, hence a more com- fall in international comWition for many rea- petitive product. Alternatively, higher profits sons. Seldom can single causes be foundmore 'may be possible, which can help finance fur- Seldom yet simple, straightforward palicy ther improvements. Production technologies rernedies. Plainly, industrial policy offers no are particularly important in consumer elec- quick fixes for the dilemmas of the U.S. con- tronics and semiconductors, less so for large sumer electronics industry, nor any sure pre- computers. scriptions that can guarantee the future com- ,

6' 8 International Competitiveness in Electronics

the game. In the United States and in other, parts of the world, business enterprises coal pete in an environment shaped to considerable extent by government industrial policies (including elements of fiscal, monetary, tax, man power, trade, and regulatory policies). Foreign governments are experimenting with industrial policies intended to aid and support their own electronics industries; virtually all industrialized and industrializing nations single out electronics for special treatment. Amer- ican firms seeking to export or to manufacture overseas must contend with economic and social policies of host governments that are 'more complex and sophisticated than in the past. Rather than outright protectionism or other forms of overt discrimination against foreign firms, host governments now adopt indi rect subsidies for their own industriestax incentives, capital allocations, funding for commercially oriented research and development 4. (R&D). At the same time, governments bargain \vissifieisfrod with foreign multinationals using carrots and Photo credit: Mostek Corp. sticks such as investment incentives and per- Portion of a read-only integrated circuit memory formance requirements while seeking to acquire jobs and technology, or to improve their balance of payments. _petitiveness of our microelectronics or comput- 3. Although a well-designed and supportive er sectors. Just as plainly, competitivenessin industrial policy is not, by itself, sufficient to electronicsand in other U.S high-technology build competitiveness in a given sector of a na- industrieswill depend on factors including: tion's economy, government policies can, under some circumstances, tip the balance. The capable people, hence on Federal policies United States can expect no more than very dealing with education and training; limited success in negotiations with other na- capital for new business startups and for tions aimed at minimizing the impacts of those expansion, hence on macroeconomic and countries' industrial policies. tax policies; open markets for American products, For this and other reasons, a "business-as- hence on foreign economic policy; and. usual" approach is unlikely to prove sufficient the research base that supports domestic to the task of maintaining U.S. competitiveness firms, hence on Federal technology and in electronics. Better prospects for strengthen- science policy. ing the U.S. position would come with the adoption of more effective industrial. policies The job of industrial policy is to evaluate, link, of our own. The American electronics industry and coordinate the many Federal efforts that faces only a few major problems, mostly in the deal with such concerns. trade arena, that are directly--susceptible to 2. While, international competitiveness is Government remedy. On the other hand, Fed- firmly rooted in the efforts of private com- eral agencies could support the industry panies, public policies set many of the rules of directly and indirectlyin many ways. Few of

17 Ch. 1Part A: Summary 9

these would have much visibility. By the same What would be the implications of a decision token, they would not necessarily cost much. to pursue a more coherent industiial policy in Consistent and careful attention to the many the United States? First and foremost, that to smaller matters that affect competitiveness automatically equate "industiial policy" with diffusion of technology within the United a greater degree of Government involvement States, tax treatment of equipment contribu- in the economy is to view the matter from an tions to universities, the antitrust environment arbitrarily narrow perspective.Industrial for joint R&D, long-term basic researchare policy does not have to run counter to efforts the necessary ingredients in a more coherent to "get Government off the backs of business." and productive industrial policy. A supercom- Rather, it should be construed as an effort to puter project, to take a current example, may make the inevitableindeed oftentimes desir- be glamorous as well as desirable in itself, but able and necessaryFederal involvement a is no substitute.. \.. more consistently productive one. It implies an 4. The choice of policy tools, and,thr design effort to develop, both politically and institu- of individual measures, depend on overall ob- tionally, Government policies toward industry jectives; an industrial policy is the sum of many that: parts that can be put together in different`ways. explicitly consider impacts on com- Should CongreSs wish to pursue .a more ..o- petitiveness and economic efficiency; ctised industrial policy for the United States, seek to treat the problems and opportuni- it could choose from among five broalter-\ ties of particular industries in the context natives: of the economy as a whole, rather than in A protective strategy aimed at presrmving isolation; and, the domestic market base for U.S. indus- do a better job of relating policy tools to tries, along with preservation of existing policy objectives. jobs and job opportunities. Protection and/or support for a limited Policy Concerns in Electronics numbe: of industriesi judged Critical for the U.S. .:!,.7ratolny or. more narrowly, for na- Among the elements of industrial policy, the tional semirity. following are vital for the continuing com- Support for the technological base and in petitiveness of the U.S. electronics industry. Aitutional infrastructurethat underly They might have rather different places, and American industries, with particular atten- be addressed in different ways, under each of ticn to structural adjustwent (e.g., labor the alternatives listed above. ret;atuing and mobility). 1. High-quality education and training (in- Promotion ithe glolvil competitiveness of cluding retraining) for engineers, technicians, firms inc'zYstries by encouraging and other skilled workers. tand or.,,m uwnpf:-:tition in domestic Js int -7.:iatiDiiA morkets. More than anything else, the competitive position of the United States in high technology, 2 fertwi tvhorc: r3e private sector when ..-'7cefiling industrial has been built on the human resources avail- devehiy-ttent an able here. A renewed Federal commitment to education and training seems called for (see .While these five approaches to industrial pol- ohs. 8 and 9). Engineering enrollments running i(., discussed. in chapter 12, are certainly not at record levels have swamped the resources 114.f.tually _exclusive, they represent distinctly available in colleges and universities; even so, diterent thrusts, implying different mixes of the United States graduated but 63,000 engi policy instruments as well as different goals. neers in 1981 compared to 75,000 in ,Japan.

10 10 International Competitiveness in Electronics.

U.S. electronics firms have faced serious 3). Corporations, cities and regions. and peo-- problems in finding adequate numbers of en- ple must adjust to changes, many of which are gineers, as well as technicians and service per7 outside their control. Federal attention to max- sonnel with needed skills and aptitudes. Inade- imizing the positive effectives of changee.g., quaw resources in U.S. engineering schools are stimulating growth industrieswhile:amelio- harming the quality of education as well as con- rating the negative impacts, could be Opt1: of the straining the numbers of new graduates. Train- central elements in a more coherent industrial ing and retraining for technicians and parapro- policy for the United States. Policy initiatives fessionals. varies widely in quality and appro- aimed at personnel mobilitywhether \geo- priateness to emerging needs. Many people in graphic, inter-industry, or within organiza- the United States emerge from high school tionsare one example. quite unprepared to work in technology-based 4. Adequate supplies of investment capital industries. for new startups as well as rapidlyexpanding Despite fluctuations in supply and demand established firms. over the years., engineers in principle comprise one ''ofthe most employable occupational As discussed in chapter 7, venture capital groups in the labor force; it is hard to imagine markets in the United States function well, an "oversupply" of engineers or of people with although cyclic downturns are likely to recur good technical training of any of a wide vari- and risk capital is often hard to find at early ety of types in an economy like that of the stages of technology development. United States, provided that people are will- Rapidly growing companies, particularly in' ing and able to shift jobs according to demand the semiconductor industry, do face severe Within the economy and organizations are will- financial pressures. These stem from increasing to help them do so. ing capital intensity, due both to higher R&D 2. A strong technological basestemming expenses and to production equipment that has from basic research and applied R&D with gone up in cost by an order of magnitude over long-term objectives, including the diffusion of the past decade, coupled with the preference results, in-fields such as solid-state electronics, of American managers to finance expansion optical devices, communications technologies, from internally generated funds. Tax policies computer-aided design of circuits and systems, have a major influence over sovirces of financ- and computer software. ing and risk absorption. The Federal .Government could not only con- While the advantages are not as great as tinue to fund basic research, it could establish sometimes implied, large diversified electron- new mechanisms for diffusing the results of ics companies in Japan, and perhaps in some R&D to the private sector, experiment with the Western European countries, do benefit from support of commercially oriented (rather than real (i.e., inflation-adjusted).costs of capital that military) research, and strengthen tax incen- are somewhat lower than for merchant semi- tives and other encouragements for successful conductor firms in the United States. By them- innovators. selves, these differencesmatters of a few percentage pointsare not enough to weigh heavi- 3. Economic adjustment policies that smooth ly in the competitive balance. Constraints on flows of capital and labor within the economy, rates of capital spendingdue in part to the aiding growing firms in their efforts to corn- preference of American firms for internal fi- petewhilehile providing well-paying jobs for. the nancingare more likely to be a drag on the domestic labor force. competitive abilities of U.S. manufacturers. Structural change is a fact of life in American These and other factors, primarily expectations industries, driven by-the currents of an increas- concerning ;iflation, tilt the investment deci- ingly open international economy (see chs. 4 sions of American managers toward the short- and 5), as well as byjechnological change (ch. er term. Ch. 1Part A: Summary 11

Increase in Capita! Costs for HighVolume Integrated ferred dfii;ign and processing know-how to in- Circuit Production Line dustry, rallied public support behind the structural shifts th=.t were leading Japan toward an "information economy" (or at least helped diffuse counterpressures by those disadvantaged by such shifts). The goals of the heavily publicized fifth-generation computer project are similar. When many of the impacts of industrial policies are intangible, how do we countervail them? Over at least the rest of the decade, U.S. trade negotiators can expect to grapple with such issues. The prerequisite is an analytical capability by the Federal Government adequate for understanding the ways in which public policieshere and elsewhereaffect international competitiveness. S10 million American electronics firms, particularly 1910 1975 1980 1985 1990 manufacturers of semiconductors and com- Year puters, may also need the continuing support SOURCE: R. W. Broderson, "Signal Processing Using MOSVLSI Technology, VLSI Electronics: Microstructure Science. vol. 2, N. G. Einspruch (ed.) of the U.S. Government, via both bilateral and (New York: Academic Press, 1981), p, 206 multilateral negotiations, in securing access on reasonable terms to foreign marketsfor ex- 5. An international trading environment that ports and for direct investmentif they are to places U.S. firms on a more-or-less equal maintain their competitive position. Only by footing with their competitors in other coun- competing aggressively all over tho world, tak- tries, including those that have well-developed ing advantage of scale eciro .,;,dnew industrial policies intended to protect or pro- opportunities, can Atri4,riuun 1-.11St.:Xlit1Ct to mote domestic manufarturers. share fully in the growth and expansion that As discussed in chapter 11, the framework will characterize this industry into the next for international trade that emerged in the cenicry. As an example, semiconductor sales postwar era is being overrun by events. The in Japal already exceed those in all of Western ,nrust of industrial policies in many nations is Europe ',.;./ more than half; U.S. firms need ac:- toward indirect supports with effects on prices cess to Japan's market comparable to that en- and on competitivenes: that cannot be quan- joyed by Japanese suppliers here. titatively assessed (see ch. 10). Japanese in- Regardless of the overall approach and direc- dustrial policy, for instance, works in part by tion of U.S. industrial policy, Congress could breaking bottlenecks; the VLSI project of the act in support of objectives such as those out- 1970's helped train Japanese engineers, trans- lined above.

The Competitive Position of the U.S. Electronics Industry Consumer Electronics left the market. Few radios or black-and-white TVs are made in the United States. No video 1. American firms making radios, TVs, and cassette recorders are manufactured here. Col- audio products such as stereo receivers and or television production has become largely an tape recorders have been under severe compet- assembly operation, heavily dependent on im- itive pressures for years; many have failed or ported componentswhether the parent firm 12 International Competitiveness in Electronics

U.S. Sales and imports of Selected Consumer Electronic Products, 1982

U.S, sales Imports Import penetration (millions of dollars)(mAlions of dollars) (percent)3 Color television $4,253 $546 12.8% Blackand white TV 507 344 67.9 Video cassette recorders 1,303 1,032 100.03 Home and auto radiosb 1,579 1,207 76.4 Stereo systemsc 1.754 1,342 76.5 $9,396 $4,471 47.6% a8ecause many Items Imported in a given year are not sold until the following year, dividing Imports during a given calendar year by sates in that same year may give only a rough indication of Import penetration; for Instance, all video cassette recorders sold in the United States are imported even though 1982 sales figures exceed 1982 Import figures. bincludingauto tape players. cincluding audio tape units and other component equipment. SOURCE: Electronic Market Data Book 1983 (Washington, D.C.: Electronic Industries ASsociatIon, 1983), pp. 8, 19, 31. is American -owned (RCA, Zenith, GE) or for- Price Index for Televisions Compared eign-owned (Sony, Quasar, Magnavox). In tele- to All Consumer Durables vision manufacture especially, the policies of 250 the Federal Government have contributed to the plight of the industry. Dumping complaints 200 against importers going back to 1968 have 0 II never bi.en fully resolved. An industry legally tz, 150 entitled to trade protection has not received it. 0) 2. Nonetheless, trade practices illegal under 100 U.S, Inv e been only one factor in the declining competitiveness of the American con- 50 sumer electronics industry. More fundamentally, competitive advantages have shifted to 0 other parts of the worldfirst japan, now new- 1973 1974 1975 1976 1977 1978 1979 1980 1981 ly industrializing countries like Taiwan and Year South Korea. These countries 'have mastered SOURCES: Consumer DurablesEconomic Report of the President 1982 (Washington, D.C.: U.S. Government Printing Office, February 1982), the technological requirements for mass-pro- p. 294. TelovIslons.Electronic Market Data Book (Washington, D.C.: Elec- ducing consumer products such as TV sets. tronic Industries Association, 1982), p. 29. They have lower labor costs than the United States, an adequate corps of skilled workers and engineers, supportive government indus- pressed to keep up with their foreign-based trial policies, and astute corporate manage- competitors. While U.S. firms may continue to ments. innovate and to be leaders in consumer products aimed m. specialized market nichescom- American firms have been reduced to a reac- puter games have been a recent example tive posture; they have lost the lead in product broadly speaking, product leadership has been design and development while moving manu- lost. At least in the short term, prospects for facturing operations to foreign countries in taking the lead in new generations of high val- order to keep their costs competitive. American ued-added mass market products seem slim. products in consumer alectronicase.g., color television receiverscontinue to he competi- 3. The rise of foreign firms together with pro- tive in performance, quality, and reliability, but tracted trade disputes have contributed to a ma- they are no better than imports. The consumer jor shift .in the structure of the U.S. consumer electronics market is highly price-competitive; electronics industry. The number of firms-has without advantages in technology or product not charged greatly since the 1960's; but while features,.American manufacturers will behard- once there were 16 or 17 American-owned

21 Ch. 1Part A: Summary 13

manufacturers of TVs, today only 4 of 15 with In this regard, U.S. experience with OMAs plants in the United States have headquarterS restricting color TV imports has paralleled here. Still, the market shares of the traditional other cases of import quotas, for instance in U.S. leadersZenith and RCAhave not the steel industry. Although the ostensiblepur- changed much; together these two companies pose may be to give domestic firms time to re- continue to hold about 40 percent of the U.S. structure and adjust to changing competitive color TV market. It is the weaker American circumstances, most cases protected indus- manufactUrers that have succumbed. tries continue to react to pressures from abroad rather than taking strong positive steps of their 4. At the same time foreign enterprises were own; the notion that a respite from import com- investing in assembly plants in the United petition will, by itself, help corporations restore States, American-owned firms were transfer- their competitiveness gets little support from ring labor-intensive manufacturing operations events in color television. to low-wage offshore locations. In general, final assembly for the U.S. market remains here, with subassembly in Mexico and the Far East. These moves wore driven by foreign competi- Semiconductors tion. t;.S. color TV manufacturers felt they had 1. U.S. manufacturers of semiconductor little option but to move production abroad if products such as integrated circuits remain they were to cut costs and meet their competi- highly competitive in markets allover the tor's prices. world. American-owned merchant firms-- those that produce for the open market Offshore production substitutes quite directly ar'a for jobs in the United States. Nonetheless, if leaders in circuit design andprocess technol- American firms had not moved offshore, it is ogy. While their share of world sales has changed little over the past fewyears, with U.S. quite possible that they would have losteven firms and their subsidiaries still accounting for more ground to foreign-based competition, about 70 percent of worldwide output of inte- with yet more jobs lost over the longerterm. In most cases, transfers of production overseas grated circuits,. Japarfe-Se manufacturers have hare net impacts on U.S. employment and been catching up in technology. Nonetheless, on U.S. companies have the capability to maintain the U.S.. economy that appear relatively small; their competitiveness in most world markets. improvements in labor productivity, forex- amplealso- driven by foreign competition The inroads made by Japanese suppliers of have been at least as importantas a cause of commodity-like chips, notably randomaccess employment declines in television manufactur- memories (RA Ms); portend stronger competi- ing. Needless to say, the impacts on individuals tion in other types of microelectronic devices and communities where job losses concentrate but do not translate automatically into advan- are often severe and long-lasting; in 10 years tages for products such as logic chips or mil the production work force in consumer elec- croprocessor families. There is no reason to ex- tronics has been cut by more than 40 percent, pect a loss of competitiveness in advanced microelectronic products paralleling that incon- from 85,000 to 50,000. sumer electronics. 5. Beginning near the end of the 1970's, Or- Although foreign manufacturersmay some- derly Marketing Agreements (OMAs) limited times have advantagese.g., -supportivegov- imports of color TVs while encouraging for- ernment industrial policies, as in Japanor eign firms to produce here. The resultwas to Western Europethe U.S. merchant firms equalize the terms of competition and to mod- have their own strengths. Among theseare the erate employment declines in the United ability to rapidly develop and commercialize States. Otherwise, the OMAs did little to help new technologies, to anticipate and design for the U.S. industry rebuilditscompetitive shifting customer needs, and to adapt to chang- strength. ing realities of international competition by 22 14 International Competitiveness in Electronics entering into joint venture and technology Lions is a predictable trend in the, evolution of transfer agreements with both domestic and the industry. foreign firms when this is advantageous. 3. At the same time that relatively mature 2. The.structure of the merchant portion of companies like Intersilpurchased during the U.S. semiconductor industry is changing. 1981 by General Electricare being acquired, A. number of well - established semiconductor new entrants continue to repopulate the mer- firms founded during the 1960's or early 1970's chant semiconductor industry. While the .have been acquired by larger, diversified enter- downturn in venture capital markets during the prises, either American- or forein-owned. In middle and late 1970's virtually halted start- parr, these structural shifts are associated with ups, new firms are again being established. a trend toward aptivi production by end-prod- Since 1980, several dozen small firms produc- uct manufacturer;:. ing custom integrated circuits, gate arrays, specialized memory chips, ..and other niche Companies that design and build systems products have entered the industry. Aiming at ranging from computers and communications portions of the market where the knowledge networks to automobiles increasingly see needs and expertise of their founder3 can be brought for internal capability in the design, develop- to bear, some of these start-ups will be suc- ment, and manufacture of state-of-the-art mi- cessful and expand, some will remain small, croelectronic devices. The acquisition of mer- others will be acquired by larger enterprises. chant semiconductor firms by larger corpora- 4. Captive manufacturers of semiconductor. devices make vital contributions to U.S. competitiveness. Such companies include IBM, the largest producer of semiconductors in the world, and Western Electric, which moved into the merchant market in 1983an action made possible by the settlement of the GoVernment's antitrust suit against AT&Tas well as a number of aerospace and defense contractors. Com- panies that produce for internal use not only provide a major part of the technological foundations for microelectronics, they spawn startups and give training and experience to people who later move to other companies. 5. Just as important for continuing international competitiveness are firms that design, develop, and build production equipment for applications ranging from annealing silicon wafers to automated testing and assembly. While the United States maintains the lead among open-market suppliers of many types of processing equipment, notably in lithography, other countries are catching up. Gov- ernment-sponsored R&D in Japan has focused on production equipment. 6. R&Dparticularly that with relatively Photo credit: General Motors long-term payoffswill remain a critical force Microcomputer for controlling an automobile engine in support of U.S. semiconductor firms. In the Ch. 1Part A: Summary 15 past, much of tho technology base has come 2. American firms have done a much better from larger firms such as II3M and AT&T. Gov- job than their foreign competitors of balanc- ernment support for research has not been sig- ing what the available technology can do nificant in recent years, although the Very against what customers for data processing sys- High-speed Integrated Circuit program of the tems have wanted to accomplish. This has been Defense ,Department will have commercial an important element in Patterns of competi- spinoffs. tive success, which have depended as heavily The U.S. semiconductor industry can no on software that could be easily used by neo- longer rely on past approaches to R&D and phyte purchasers and was reliablei.e., free of technology development. The industry recog- "bugs"as on raw hardware performance. nizes the changing situation, and is develop- In fact, foreign computer firms have some- ing new mechanisms for strengthening its times been able to match the United States.in technical foundations; these include closer interms of hardwam; by and large, Japan's com- teractions with universities, along with joint puter marufactUrers can at present. But their Ventures and cooperative research efforts. Con- systems are still behind, mostly because the gress and the Federal Government could ac- softwareat all levels is not as good: More- tively support and encourage both basic and over, foreign firms -- whether European or Jap- applied research with longer run payoffs. This anesehave not been as adept as Americans is one of the surest ways of supporting contin- at finding new ways to apply their hardware. ued U.S. competitiveness in microelectronics. For example, U.S. firms remain well ahead in office automation, point-of-sale terminals for Computers retail merchandisers, and many other applica- 1. American manufacturers of digital com- tions of distributed intelligence. puters have dominated world markets for many years. Much as U.S. semiconductor firms 3. The ongoing structural alterations in the have demonstrated the ability to rapidly capi- data processing industry will be deeper and talize on new technological and market oppor- farther-reaching than those in microelectronics tunities, so have American computer firms or consumer electronics. pioneered most of the design concepts that have driven information processing: network- Most of the recent technological innovations ing and distributed computing, small business in consumer electronics have come from large, machines and minicomputers, time-sharing well-established firms; new products from among multiple work stations, cheap mass small companies have seldom reached mass storage, desktop microcomputers. markets. In microelectronics, while start-ups have resumed in the United Statesmany striv- There are few concrete signs that this ing to establish themselves with the aid of in- dominance by U.S.-based firms.is threatened. novative productsthe path of technological Nevertheless, relative positions within the evolution seems, for the moment, well charted; world computer' industry will continue to shift, there are few signs of sudden change that stimulated in many cases by new applications would seriously unsettle the industry. Com- of computing power. As the industry continues puter technology which depends on micro- to evolve, the technological leads of American electronics, but also on other feeders, primarily firms are likely to shrink, and competitive posi- softwareis potentially more volatile. As new tions may become more difficult to maintain. applications of computing power open win- Nevertheless, the U.S. lead in worldwide mar- dows of opportunity for firms in many parts keting of data processing systems is so large of the world, American manufacturers will that prospective challengers such as Japan can- face more intense competitive pressures. Dis- not hope for more than modest success over tributed intelligence will transform a broad the rest of the century. range of other industries as well.

29. 16 International Competitiveness In Electronics

While the err, of the mainframe computer is ness systems, personal computers, and "smart" hardly over, the increasing importance of devices that do not even look like computet.s sinaller machinesminicomputers, small busi- will continue to provide the greatest oppor-

Market Segmentation of U.S. Computer Sales by Value

1975 1980

Word processors 10%

1985 (projectec,.

SOURCE: "Moving Away From Main Frames: The Large Computer Makers' Strategy for Survival," Business Week, Feb. 15, 1981, p. 78.

25 Ch. 1-Part A: Summary 17.

(unities for growth and expansion. The multi- 4, In the computer industry, as in microelec- tude of prospective applications of computing tronics, U.S. employment is rising much less power will offer new openings for overseas rapidly than output. Although new jobsare be- firms as well as American companies. Insome ing created making, operating, and maintain- portions of the data processing equipment ing "smart machines, other jobsare being iu those still in relative in- destroyed; the net effects on U.S. employment fancy. such as desktop machines and standard- might be positive or might be negative. While ized otilc;0 automation productsforeign firms there is little meaningful evidenceon either may eventually achieve a greater presence than side of the job creation/job destructionques- :they have managed in mainframe systemsor tion, there is no question that skill requLements general-purpose minicomputers. To the extent are changing rapidly. In some cases, automa- that computers become mass-market products, tionaided by electronicslowers the skillre- manufacturers in other parts of the world are quirements associated with the rertiaining jobs: likely to emerge as more formidable competi- in other cases, "upskilling" rather than "de.- tors. skilling" results. A readily predictableconse-

N414

1% It* ,---Nam&

iro .9 .etrair )12'

Photo credit: Unirnation Industrial robot at work 18 lrternational Competitiveness in Electronics quence has been serious labor market disloca- generation computer project, Japan's industry tions; these seem bound to intensify. Even if is bent on achieving technological and com- labor market shifts cannot be predicted very mercial parity (or superiority) in machines well, the need for adjustment is clear. To the ranging from desktop processors to supercom- extent that labor market shiftsgeographical, puters. Still, Japan's rising export strength in in terms of skills, in terms of wage levelsare computers differs in a major way from the pat- unexpected (and some will always be), the im- terns visible in consumer electronics or semi- pacts will be more severe. An obvious implica- conductors: the leading Japanese exporter of tion is that policy responses must emphasize computers, by a large margin, is IBM-japan flexibility. despite the fact that it has been barred from 5. Japan's computer manufacturers will not many of the government programs that have be content with narrow or specialized markets. aided other computer manufacturers. Following strategies similar to those that have succeeded in consumer electronics anti semi- While IBM has abundant resources and tech- conductors, Japanese computer firms will at- nology to compete effectively against Japanese tempt to establish themselves in selected data computer firms, other American manufactur- processing markets and expand from there. ers may face increasing difficulty in the future. Backed by government efforts such as the fifth- AlthouE.,: the U.S. industry is not immediate-

Japanese Production, Imports, and Exports of ComputersandEquipment, including Production and Exports of U.S.-Owned Subsidiaries :,800

1,600

1,400

1,200

C >,) 1,000 0 C L2 800 co

600

400

200

0 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 Year SOURCES: 1987.78Japan Fact Book '80 (Tokyo: Damps Publications, Inc., 1980), pp. 173, 174. 1978-80Japan Electronics Almanac 1982 (Tokyo: Dempa PubKcalions, Inc., 1982), pp. 58, 59 1981, 1982 Department of Commerce, Bureau of Industrial Economics. Ch. 1Part A: Summary a 19 ly imperiled, the Federal Government could software development has gone up only help ensure future competitiveness through a slowly over the years. Financing for educa- better developed, more consistent industrial tion and training in software engineering, policy, particularly one supporting technology as well as R&D directed at computer archi- development and technical education. tectures, new programing languages,"and artificial intelligence appears appropriate. 6. As computers and their applications con- Smaller firms striving to establish them- tinue to spread through the U.S, economy, the selves in the data processing equipment in- Federal Government might act to strengthen dustryparticularly those developing soft- the competitiveness of the industry both direct- ware, peripherals, and innovative applica- ly and indirectly: tions of computing powerhave the same "Computer literacy"the ability to effec- needs as do U.S. microelectronics firms: Lively utilize smart machines and sys- not only people with highly developed temswill be a critical skill for the labor technical skills, but adequate supplies of force. Education and training in fields capital for investment in R&D and produc- ranging from traditional modes of quanti- tion capacity and access to foreign mar- tative thinking (arithmetic, algebra) to soft- kets. ware engineering deserve renewed support. Congress could provide leadership as If effectively implemented, industrial policies well as direct and tangible aid. in support of such needs could pay vast divi- Federal support aimed at critical bottle- dends throughout the-U.S. economy because necks in data processing, mostly in soft- of the multitude of ways in which applications ware, could be a vital long-term stimulus of computing power can enhance the compet- for the American industry. Productivity in itiveness of firms in industries of all types.

Conclusion A nation can never be competitive in all in- to change. The dynamic of international com- dustries at once. Not only will some rank petitiveness is continuous, and calls for a con- higher than others, but places will change over tinuing series of policy responses. time. Economies need to adjust; adjustment brings pain and distress to firms that encounter People can and will argue endlessly about the trouble, people who lose their jobs, the commu- successes and failures of industrial policies in nities affected. Even within an industry like other countries, but the primary lesson to be electronicsin the United States, highly com- drawn from foreign experience is simply this: petitive as a wholesome parts, such as con- industrial policymaking is a continuing activity sumer electronics, face a far more problematic of governments everywhere. In the United future than others. That s ch events are inevi- States, industrial policy has been left mostly table does not mean that at least some of the to the random play of events. Improvement is problems cannot be anticipated, and Some of clearly possible; policymaking can be a pur- the distress ameliorated by Government action. poseful activity characterized by learning from Moreover, the Federal Governthent can take past experience within a framework of empir7 positive actions to support the develociment ically based analysis. Developingmore effec- and diffusion of technology, human resources, tive industrial policy for the United States must the infrastructure that companies depend on begin in this spirit, While recognizing that the when pursuing' their individual competitive process is inherently political. There is no one strategies. Government policies can aid grow- thing that the Federal Government can do that ing sectors, help people and institutions adapt will make a big difference for the future com-

2 20 International Competitiveness in Electronics

petitiveness of the U.S. electronics industry, Until the Nation begins this task, American but there are many specific policy concerns firms will continue to find themselves at a dis- that deserve attention. Only by linking and advantage when facing rivals based in coun- coordinating these more effectively can the tries that have turned to industrial policies as United States expect to develop a coherent and a means of enhancing their own competitive- forward-looking approach to industrial policy. ness.

29 CHAPTER 1 Part B: Extended Summary Contents

r,t;t:. "lechnology 23 (:;)n-innic; 23 (;ein ic:onduct ors 24 :ompnters 28 Finance 31 Vt!tittn.4, Capital :31 nanc:ing Growth 32 Intematior :1 I)ifferences 32. Human Resources 34 Quantity and Quality 34 Continuing Education am: Training 35 Preparation for Work in Technology-Based Industries 37 NIIII;Welflnilt and Organization 37 Employment 39 Trade 42 Antidumping Enforcement, 42 The Environment for World Trade 42 U.S. Industrial Policies 46

Listof Tables Page Color TV Imports into the United States 24 World Integrated Circuit Output by Headquarters Location of Producing Firms 27 Predicted Growth Rates by Occupational Category in the United States Over the 1980's 41 Samico-nductor Sales in the United States, Western Europe, and Japan 44 List of Figures Page Japan's Production and Exports of Video Cassette Recorders (VCRs) 25. U.S. Semiconductor Sales by Typo 26 U.S. Production of Computer Equipment 29 Rates of Capital Spending by U.S. and Japanese Semiconductor Firms 33 U.S. Employment in'Consumer Electronics 39 U.S.Employment in Computer (and Peripheral EquiPment) Manufacturing 40 U.S.-japan 'Trade in Integrated Circuits 44 Distrilmtion of U.S. Semiconductor Sales by End Market 46 CHAPTER 1 Part B: Extended Summary

Part B of chapter 1 expands upon, without competitiveness" are discussed in detail in repeating, the findings in the St:- .mary. In parr chapter 5. The viewpoint adopted below is first ticular, the sections below highlight the role of that of the manufacturer. Private companies technology as a force on competitiveness in design, develop, produce", and market goods consumer electronics, semiconductors, and which may have more or less success in the computers, along with factors such as capital marketplace, more or less positive impact on for investment in research and expandedpro- a nation's competitive position. Later the view duction capacity, human resources and their switches to that of governments and their pol- development, and industrial and trade policies icies, which act on competitiveness directly both here and abroad. and indirectly by influencing business activities, supporting- education, subsidizing ex- The several meanings that can be assigned ports, through the climate for capital formation to the rather amorphous concept "international and economic growth.

Technology

Chapter 3 covers electronics technology in japan, but in developing countries like Taiwan some detail (also see the Glossary, app. A, for and South Korea. Manufacturing technologies explanations of technical terms). Here the in- are similar wherever TVs are built, with labor- teractions between technical capabilities and intensive operations carried out in low-wage market success are explored. developing countries by European and. Jap- anese firms, as well as American. The result Consumer Electronics is a competitive environment in which Ameri- can firms have few unique advantages. In consumer electronic products such as coil_ or TVs, both product and process technologies Differences in both product and process tech- are well-understood and widely fiiffused. Prod- nologies for televisions were greater during the uct differentiation strategies are more impor- late 1960's and early 1970's whet Japanese tant than technical differences; component firms were beginning to invade the U.S. mar- television, stereo sound, and digital chassis ket. Then, firms in Japan moved more quickly designs illustrate the frontiers of this nowiarge- than their American counterparts toward solidly routine field. Japan's consumer electronics state chassis designs. By using transistors and manufacturers have benefited from the econo- integrated circuits (ICs), they were able to im-. mies of higher production volumes and per- prove the reliability of their products, and more haps from more extensive automation, but both easily automate portions of the production prodUct and process technologies in consumer process. Automation helped compensate for electronics tend to be standardized, technical component costs that at the time were" higher change to be incremental. Companies any- for transistorized designS than for those rely- where in the industrialized world have access ing on vacuum tubes. Reliability was particu- to much the same pool of knowledgethe ex- larly important to Japanese firms because they ceptions being newer product families like did not have service organizations or, dealer video cassette recorders (VCRs). Color TV3 networks within the United States. To increase - with similar prOduct features are made not their market shares, they needed to sell through. only in Western Europe, the United States, and retail outlets such as discount chains. To

23 24 International Competitiveness in Electronics

Color TV Imports Into the United States

Number of color TVs imperted Imports from all sources by origin (thousands) as a percentage of Year Japan Taiwan Korea Total a U.S. consumption 1967 315 318 6.7% 1969 879 22 912 15.7 1971 1,191 85 1,281 18.9 1973 1,059 325 2 1,399 15.8 1975 1,044 143 22 1,215 17.9 1977 1,975 318 92 2,476 27.0 1979 513 368 314 1,369 13.6 1981 727 514 393 1,946 15.6 aincludes imports from countries not listed individually. SOURCES: 1967. 1969Television Receivers and Certain Pads Thereof (Washington, D.C.: U.S. Tariff Commission Publica- tion 436, November 1971), p. A62. 1971.1973Television Receivers, Color and Monochrome, Assembled or Not Assembled, Finished or Not Finished, and Subassemblies Thereof (Washington, D.C.: U.S. Interntnonal Trade Commission Publication 808, March 1977). pp. A-90, A99. 1975.79Color Telev'sion Receivers and Subassemblies Thereof (Washington, D.C.: U.S. International Trade Com- mission Publication 1068, May 1980). p. D-6. 1980Television Receiving Sets From Japan (Washington, D.C.: U.S. International Trade Commission Publication 1153. June 1981), p. H.21. 1981Information from Department of Commerce

achieve credibility, they had to supply TVs that Semiconductors did not need frequent service. Japan's con- Microelectronic devices, in contrast, are in- sumer electronics manufacturers succeeded in termediate products sold in accordance with this far from riskless strategy. detailed technical specifications to sophisti- If technology is now a secondary factor for cated customers who design them into final TVs, in more recently introduced product fam- products ranging from TVs and electronic iliesnot only VCRs, but video disk players, games to missile guidance systems and power- home computers, and related applications of ful computers. To be successful, semiconduc- electronic technologies to consumer goods tor firms must not only meet the current redesigns are evolving at a faster pace. Japanese quirements of such customers but do a good entrants spent many years and a great deal of job of anticipating their future needs. money on engineering development of VCRs Matsushita even reached production in 1973 Technological Factors in Competition with a design that was shortly thereafter judged not to be good enoughbefore achieving com- As explained in chapter 3, the interdepend- mercially viable products. But otherwise, com- ence of product and process technologies in petition in consumer electronics is largely a leading-edge microelectronic devicesvery matter. of prices and marketing, brand loyalties large-scale ICsis unusual even for a high-tech- and customer perceptions. While Japanese ex- nology industry. Circuit designers must under- porters have established themselves firmly in stand the nature and capabilities of the fabrica- American markets for TVs and audio products, tion processincluding proprietary details individual companies have suffered frequent to optimize the performance of a chip. Product reverses in consumer goods ranging from and process technologies advance together, Stereo receivers to CB radios and pocket cal- with process capability a restriction on devices culators, where markets have been unpredict- that can be fabricated with acceptable yields able and competition always stiff. (the fraction of circuits that function). The in-

33 Ch. 1Part'8: Extended Summary 25

Japan's Production and Exports cal of the semiconductor industry, are driven of Vireo Cassette Recorders (VCRs) by intense competition to improve process yields, reduce manufacturing costs, and cut prices to build market share. As the prices of 12 64K RAMs dropped, prices also fell for the previous-generation 16K devices, which by 1982 I sold for about $1 each. Similar patterns will be followed as 256K RAMs, in pilot production 10 in both japan and the United States during 1983, take over from64Kchips. Despite the intense price competition in these commodity-like circuits, product technol- 7 ogy continues to play a role. Not only is a good circuit design essential for high yields and low 5 costs, but a high-performance RAM can corn- mend a greater price. While the most common 5 varieties of 64K RAMs have access times (the average time to retrieve the contents of a memory cell) in the range of 200 nanoseconds (200

3 X 10 seconds), otherwise comparable circuits with lower access times sell for more; 2 during 1982, 64K RAMs with access times of 150 nanoseconds brought prices a dollar or so above those for 200 nanosecond circuits.. Nonetheless, RAMsand most other memory 19751976 1977 1978 1979 19801981 1982 chipsare in essence standardized items. As Year for consumer products like TVs, progress is incremental and predictable, at least at present SOURCE "VTR Production Demand,"Japan Report.Joint Publications Research Service JPRS 011100. Jan. 28, 1983. P. 35 although the pace is much swirder. If process technology is vital for RAMs, prod- teractionsgo in both directions. Clever circuit uct technologyi.e., circuit designcarries design can compensate for some kinds cf proc- greater weight in competition invol. ;`uther ess limitations. Among the examples are simply varieties of ICs. Foreign firms have Len less doing more with fewer transistors or other cir- successful in microprocessor families and the cuit elements and incorporating on-chip testing arrays of support chips designed to be used and redundancy. Some American firms added with the processors themselves, as well as some redundant circuit elements to their 64K RAM types of linear circuits, logic families, semicus- (random access memory) designs, a step which tom chips, interface circuits, and the many may 'pay dividends in the future as they move other varieties of specialized microelectronic to still higher levels of integration. products. In contrast to memory chipsin essence "brute force" devicescircuits that im- Competition in standardized products like plement Ionic depend more heavily on creative RAMs depends on both price and technology engineering design, on anticipating user needs, chapter 5. When 64K RAMs were first intro- and on recognizing new opportunities made duced, they sold in sample quantities for about possible by developments in either process or $100 each. From this level in early 1980, prices device technology. A well-designed micro- fell to $10 to $15 by the end of that year. After processorone with an architecture that takes, another year, 64K RAMs could be purchased maximal advantage of the circuit elements it for $5 or less. These rapid price declines, typi- employs, with an instruction set that pro-

r),)- 1 11 -fl z- 3,4 26 International Competitiveness in Electronics

U.S. Semiconductor Sales by Type Linear and (tiler ICs

Memories

ICs Microrocessorsi microcomovre.rs

Standard logic families Linear 1975 and other Linear ICs and other ICs Standard logrc families

Standard logic families Microprocessors/ microcomputers

ICs ICs Microprocessors! Memories microcomputers 1986 1980 (Projected) SOURCE: 1975El ec !tonics, Jan. 8, 1978, pp. 92, 93. 1986 ElecfronIcs, Jan. 13, 1982, pp. 124, 125. 1946El e c t ro nics, Jan. 13, 1983, pp. 128, 129; Mar, 10, 1983, p. 8 gramers find easy to use, a convenientbus technology for linear ICs, they have not been structure and input/output portscouldbe a able to match American semiconductor firms commercial success even if developed by a in design-intensive products. For instance, the company with only mediocre processtechnol- microprocessors that Japanese semiconductor ogy. Were this the 'case,however, alternate firms se'l in large volume on the world market source manufacturers might end upwith more are U.S. designs. Such patterns willprobably of the market ;' ad/or higher profits. continue to hold, although here as elsewhere the magnitude of the U.S. lead is likely to shrink International Positions in as the Japanese get better at circuitdesign, and Microelectronics Technology as Japanese semiconductormanufacturers hire engineers from other countries. While Japanese manufacturers now make and sell many types of microprocessors and In semiconductor processing, Japanese firms logic circuitry, and have always had excellent are often on a par with the United Statesand

35 Ch. 1Part B: Extended Summary 27

may. be better in some cases. One reason has U.S. firms must continue to vigorously pursue been the VLSI research project and its several new markets and American engineering follow-ons, orchestrated and partially funded schools must retain their preeminent position by Japan's Government. By 1983, Japanese in fields related to microelectronics. manufacturers were, as a group, further along in production plans for process-intensive 256K RAMs than their American competitors. Proc- Research and Development ess control also exerts a major influence over Despite the continued prowess of American quality: nevertheless, if a few years ago the circuit designers, the comfortable lead once en- quality of some types of Japanese ICsspecif- joyed by the United States in the underlying ically, RAM chipswas higher than supplied technology of semiconductor devices isnow by American firms, today any differencesare spotty at best. American merchant semicon- much smaller (see ch. 6). ductor firms devote most of their R&D efforts Semiconductor manufacturers in japan have to product and process developments with im- made great strides as well with complementary mediate application to end-products; relative- MOS (metal oxide semiconductor) circuitry, ly small companies with limited resources, they one reason being its attractions for.certain of have had little choice but to place the greatest the consumer applications in which Japanese Priorities on R&D that will help them in next semiconductor firms for manysqrs spe- y Ear's marketplace battles. cialized. In contrast, companies in Western In the United States, more basic research Europe are generally behind the United ranging from studies, of the physics of electron States and Japan in all varieties a MOS. Euro- devices to the development of process tools pean nations are making determined efforts to such as ion-beam lithographyhas been catch up, in several cases with strong govern- funded and performed elsewhere. Some of the ment support. Despite underlying technologi- work has been supported by the Department cal abilities that in many cases are excellent, of Defensee.g., research on high-speed gal- European manufacturers have not been as suc- lium arsenide devices. In other cases, large or- cessful as American suppliers at converting ganizations such as IBM or AT&T's Bell Labo- their technology into successful commercial ratories have carried much of the burden; Bell products. In circuit design, neither the Jap- Labs, in particular, has been responsible for anese nor the Europeans seem able to match many of the seminal developments in solid- wits with Americans. This is an advantagea state electronics. In the past, Bell diffused source of "technology gap" that the United these widely to both U.S. and foreign enter- States should be able to maintain. To do so, prises. Now, withAT&T enteringnew markets,

World integrated Circuit Output by Headquarters Location of Producing Firms

1978 1982a Production Share of Production Share of (millions of dollars) world output (millions of dollars) world output United States $4,582 68.3% $ 9,700 69.5% Merchant 3,238 6,450 Captive 1,344 3,250 Captive percentage 29.3% 33.5% Western Europe 453 6.7 620. 4.4 Japan 1,195 17.8 3,440 24.7 Rest of the worldb. 482 7.2 190 1.4 $6,712 $13,950 aEstimated. Includes the Soviet Union and Eastern Europe for 1978 but not 1982. SOURCES: 1978Status '80: A Report on the Integrated Circuit industry (Scottsdale, Ariz.: Integrated Circuit Engineering Corp.. 1980).p. 4. 1982Status 1982: A Report on the Integrated Circuit industry (Scottsdale, Ariz.: Integrated Circuit Engineering Corp.. 1982),p. 5. 28 International Competitiveness in Electronics including merchant semiconductor sales, and couraging basic and applied research with po- competing under new conditions, the company tential commercial, rather than exclusively mil- may no longer feel that it has the luxuryof_sup--- applications. The Fraunhofer Gesell- porting basic research so heavily; at the least, schaft in West Germany, as well as Japan's joint it will guard its technology_rnua more close- R&D programs, both discussed in chapter 10, ly (as IBM always_has). Other forces at work come to mind. The U.S. electronicsindustry, include growing software demands on micro- including but not restricted to microelectron- electronics firmsan area constrained by per- ics, could benefit from institutional mechan- sonnel shortages, low productivity, and weak isms more closely linking R&D efforts in Gov- theoreticalfoundations. Furthermore, the ernment laboratories, industry, and universi- highly competitive merchant firms have per- ties. A relatively large but decentralized sys- haps been taking advantage of new technolog- tem of centers-of-excellence, directed toward ical opportunities faster than the stockpile has commercial developmentswith ample scope been replenished. The need for new sources for local funding and entrepreneurial partici- and mechanisms of technology development pationwould fit American traditions. Given and diffusion is plain. some fraction of funding,perhaps 30 or 40 percent, from the Federal Governmentron a con- Along with continued Federal support and tinuing basis, the time horizons could be longer incentives for R&D, particularly more basic than those for R&D programs funded entirely work, institutional innovations that would help by industry. to build the technological.base for continuing U.S. competitiveness in microelectronicsas Computers well as in computer systemsappear worthy of congressional attention. U.S. competitive- If manufacturing technology is critical for ness in electronics has dep3nded heavily onthe cost control in consumer electronics, andboth technical strengths of American firms. So long process and product technologies arevital in as the United States held a substantial overall semiconductors, the computer industry exem- lead in electronics technology, smaller com- plifies reliance on product technologies. Par- panies could successfully design and develop ticularly for larger systems, manufacturing is their products and processes without doing less significant for competitiveness because much research on their own. The foundation production volumes are modest compared to provided by large companies, military spend- TVs or semiconductor devices. For small coming, and the universities sufficed. Today, not puters sold in large numbersand particular- only is this base eroding, but the overall tech- ly the desktop machines offered by companies nical edge of the Nation has diminished. In like Appleor for peripherals such as printers particular, research capabilities in American and terminals, manufacturing technologies are universities have deteriorated because of obso- of greater and growing importance. lete equipment and shortages of graduate students and faculty. A redefined Federal role in R&D could address the need for better mech- Technological Competition anisms of technology diffusion within the What are the major factors in marketing com- United States, as well as encouraging inflows puters? First and foremost, performance/cost of technology from overseas. ratio: the computing power per dollar that a A number of promising models begin- manufacturer can supply. This depends heavily ning with dothestic ventures such as the Semi- on system designboth hardwareand soft- conductor Research Cooperative and Micro- warei.e., in doing more with less rather than electronics & Computer Technology Corp. and cutting production costs. For most computer including a number of experiments in other systems, assembly is labor-intensive, costs in- countries. Some of these are aimed at enhanc- creasing with overall complexity. The coming the diffusion of technology as well as at en- pany that can design a systemoffering higher Ch. 1Part B: Extended Summary 29 performance at a given cost has the advantage. U.S. Production of Computer Equipment IBM, as in so many ether instances in the computer industry, provides something of RTh exception because its higher sales volumes mean more pronounced scale economies. A further exception has developed at the lower end of the market, where personal machines, small business systems, and micro or minicomputers sold to original equipment manufacturers are built in much greater numbers. In both cases, greater production volumes increase the significance of manufacturing technologies but in no way diminish the role of system design. Because of these characteristics, the comput- 1970 1975 1980 er industry is just as design- and R&D-intensive Year as microelectronics, but computer engineers SOURCE: 1972, 1975, 1977, 1980. 1983 editions, US. Industrial Outlook, Depart- are seldom as constrained by manufacturing ment of Commerce. 1981 and 1982 shipments estimated. processes as chip designers. They are, however, constrained by the performance charac- where others did not. Substantial penetration teristics of available components, principally by Japanese imports may eventually-follow, but ICs. Microelectronic devices are the building based more on low prices-:-stemming from the blocks for processors as well as essential ele- well-established capability of Japanese elec- ments in many other parts of computing sys- tronics firms to manufacture in high volume tems, from controllers for disk drives to semi- at low costthan unique product features. conductor memories themselves. Because sys- Nevertheless, so long as technical evolution is tem performance depends so heavily on ICs, rapid, and software one of the keys to sales, many computer firms have established captive American entrants with creative product de- microelectronics R&D and production facili- signs should have little to fear from overseas ties. While component technologies ultimate- competitors. At least at first, the more suc- ly limit what can be done, computer designers cessful Japanese personal computers will be have considerable latitude in configuring sys- based on U.S.-designed microprocessor or mi- tems; the many alternatives from which they crocomputer chips, as well as software devel- can choose are affected in different ways by oped in the United Statese.g., the popular the characteristics ol both hardware and soft- CP/M or Unix® operating systems and the ware. many applications programs that run on them. Systems Aspects This is only one example where American Although firms located in other countries are computer manufacturers have been at the fore- nibbling at U.S. market share, our dominance front in spotting new applications of comput- in computer manufacture still continues, built ing technology. Among the other examples are: largely on the abilities of American producers Small- machines suited to the needs of at system design and integration. Conceiving businesses with a few dozen to a few hun- and developing new applications of computing dred employees. power depends on engineering design and on Fault-tolerant systems that can be used understanding user needsincluding field where reliability is critical. service and software support. American manu- Specialized data processing installations facturers opened the personal computer mar- for banks, insurance companies, and Gov- ket,not through technical advances, but ernment agencies. because they perceived a potential market Dedicated processors to be integrated into

30 30 International Competitiveness in Electronics

industrial controllers, scientific instru- Sometimes market demand has driven the tech- ments, aircraft flight-control systems. nology, with the design efforts of computer Networked systems, time-sharing, satellite manufacturers shaped by perceptions of these terminals and other mechanisms for pro- needs; occasionally, more raw computing viding users with computing power when power has been available than has found im- and where needed. mediate application. Both large and small machines for specialized scientific and technical computing, System integration remains the forte of ranging from supercomputers for complex American firms, andjust as for U.S. semicon- numerical calculations in computational ductor manufacturersso long as American fluid dynamics or the development of nu- companies and American engineers continue clear weapons to array processors to be to push aggressively into new software and used in conjunction with dedicated mini- hardware applications, they should be able to computers in modeling chemical reac- maintain a technological edge sufficient to hold tions. a large fraction of the world computer market.

4ig

Photo credit: General Motors Experimental electronic map display for automobile dashboard

39 Ch. 1--Part B: Extended Summary 31

Nonetheless, this share may not be the 70 per- that have favored American firms. Improve- cent of 10 years ago. If so, the causes will be ments in the standing of computer industries multiple. as discussed in chapter 5. Rapidly ex- in countries like Japan relative to other sectors panding and fragmenting markets mean that of these country's economies may lead to no one manufacturer can cover all the bases; greater international competitiveness in com- windows of opportunity will open for foreign puter manufacturing. Most of these forces are as well as U.S. suppliers.. Manufacturers in outside U.S. control. Given the circumstances, other countries may benefit from supportive it becomes particularly important that the Na- industrial policies, as well as drawing on grow- tion avoid unnecessary saciifices in competi- ing pools of capable computer scientists, sys- tiveness through missed opportunities by tems engineers, and managers. The result is American firms or defective policy choices by likely to he a narrowing of the technology gaps the Federal Government.

Finance Well-developed r:apital markets have been a have brought to American growth' industries major source of strength for entrepreneurial have seldom met with success. high-technology firms in tha United States. Bottlenecks in y'.s. capital markets are more Under most circumstances, both new start-ups probable and more significant when it comes and young, rapidly expanding companies have to financing rapid 'expansion in sectors like been able to find the money needed to grow microelectronics, where capital intensity is s.vith their markets. Still, this has not been escalating along with sales, than in funding universally true; in recent years, some elec- new ventures. Nevertheless, venture funding tronics companiespreferring, in common has not always been available, particularly seed with most of American industry, to fund ex- capital for developing new ideas well before pansion with internally generated cash flows production is in sight. When venture funds have found themselves lacking the financing dried up in the middle 1970's, new start-ups needed to keep pace with market opportunities. in electronics manufacture virtually halted. Perhaps of greatest significance, volatile inter- Around the turn of the decade, after the reduc- ost rates in the united States reinforce other tion in capital gains taxes that took effect in factors that bias decisionmaking by corporate 1978one of many forces affecting venture managers toward short-term undertakings. capital suppliesthe market revived. Most of those supplying venture funds look for capital Venture Capital appreciation over a 3- to 5-year period, with typical target returns being 35 to 50 percent per Over the past quarter century, venture capital year. Plainly, capital gains tax rates' are impor- in its various forms has spawned many of the tant both to individual and institutional sup- new entrants in the U.S. electronics industry: pliers of risk financing. However, for reasons companies supplying softwar', instrumenta- that are poorly understood, venture capital, on, semiconductor devices, computers and funding is notoriously cyclical: factors other peripherals. Some of theSeDigital Equipment than tax changes, also contributed to the revival Corp., Intel have become mainstays of U.S. of the market. By mid-1980, a veritable boom competitiveness. Other nationsWest Ger in venture funding was underway, with much many, the United Kingdom, even Japanhave of the money going to electronics. Prospective rnight to build some of the characteristics of entrepreneurs, many in the Silicon Valley re- J.S. venture capital markets into their indus- gion of California, saw opportunities in micro- trial policies. These attempts to generate the computers and other applications of micro- vitality and dymirnism that venture start-ups processors, in software and computer periph- 32 International Competitiveness in Electronics erals, in semiconductor chips themselves. Cap- At times over the past decade, it would have italiSts saw the technological windows in much been difficult to issue either bonds or stock. the same light. More than 20 new microelec- Nonetheless, the U.S. electronics industry ex- tronics firms alone were established during the hibits a pattern of consistently low debt/equi first two years of the venture capital resur- ty ratios contrasting sharply with foreign man- gence. ufacturers. Aversion to borrowing may have constrained the growth of some American elec- Financing Growth tronics companies over the past decade. The changes in U.S. `Ax law implemented by As chapter 7 points out, finding capital for the Economic Recovery Tax Act Of 1981 have continued expansion has been a greater con- increased cash flows for electronics firms cern for many U.S. electronics companies, par- along with other businesses in the economy. ticularly given the high growth rates in much High-technology electronics manufacturers of the industry. While there are few if any signs benefit particularly from the R&D tax credit. cf overall shortages of capital for investment Accelerated depreciation is a different matter: iii`the United States, financing growth is a com- although more rapid capital cost writeoffs are mon problem for young companies anywh..re now available to virtually all U.S. corporations, in the economy. ElectronicS firms, especially the benefits are much greater for numerous those producing ICs, face unusually steep hur- other sectors. Because electronics firms, par- dles. The first is simply th9 need to keep up ticularly in the high-technology portions of the with markets that in some years have grown industry, have always been able to,depreciate at 25 percent or more. Firms trying to increase at fairly rapid rates, they have not been helped their shares of such markets have to add pro- as much as sectors like primary metals. In duction capacity at rates that can severely earlier years, many such industries faced strain financial resources; needless to say, the lengthy capital cost recovery periods. The re/- investments must precede the added revenues ative position of electronics has suffered under they bring in. At times, U.S. semiconductor the 1981 Tax Act to the extent that companies manufacturers may have been unable to secure in other lines of business have an easier time the funds needed to keep up with market securing external funds. growthor, more likely, have judged the conditions imposed by prospective suppliers of capital unacceptable. International Differences _ Rising capital intensity for semiconchctor Why do American companies limit their use processing creates a second hurdle. Denser ICs of external funds? Most managers would an- require more expensive fabrication equipment. swer by citing high costs of capital, whether A state-of-the-art manufacturing facility, which debt or equity, as reflected in high U.S. interest cost perhaps $5 million a decade ago, now rates over the past few years. American busi- might run $50 million. High levels of R&D nessmen have claimed that they face costs of spending, mandatory for companies that hope capital perhaps twice those of their'competitors to compete in markets for advanced devices, in Japan. In fact, although costs of funds in the .contribute a third hurdle. Thus capital demand United States are higher than in Japan, the is mounting even more rapidly than the market differenceswhen adjusted for inflationary ex- has been growing, compounding the already pectationsappear relatively small, certainly difficult financing problems of U.S. semicon- less than 5 percentage points. While not insignificant, the resulting advantages for Japanese ductor firms. electronics manufacturers are hardly over- In common with most of American industry, whelming, even when the tax benefits of the U.S. electronics firms have been reluctant to higher debt/equity ratios characteristic of Jap- rely 'heavily on external fundseither debt anese corporations are taken into account. (loans, bonds) or equityfor financing growth. Lower costs of capital in Japan make no more Ch. 1Part B: Extended Summary '13 than minor contributions to differences in ket for merchant products, and to aggressive- manufacturing costs. A much more potent ly add new production capacity. It remains to source of advantage for large, diversified Jap- be seen how American firms like Mostek or In- anese electronics firms, particularly in periods tersil, which are now parts of large conglom- when markets are expanding rapidly, stems erates, will behave over the longer runand from their ability to allocate funds internally, how Western Electric will fare, now that it is using moneys generated in other lines of busi- entering the merchant market. ness to finance high rates of spending on R&D and new production capacity. U.S. semicon- While the contrasts between financing prac- ductor firms, especially those that remain in- tices of American and Japanese electronics cor- dependent and have a limited range of prod- porations are manyas are those with Euro- ucts, will always be hard-pressed to keep up pean enterprisesthe net advantages that with diversified companies, Japanese or Amer- Japan's companies receive from government ican. A major difference between diversified. guidance applied to investment funds are Japanese and American electronics firms is the small. Japanese industrial policies continue to evident willingness of Japanese semiconduc- influence capital allocations and costs of funds, tor manufacturers to compete in the mass mar- but the high leverage characteristic of Japanese

Rates of Capital Spending by 11.S.and Japanese SemIconduOtor Firms.

0) a)

0 C a) 2 a) a.

CC)

a)10 a. cn

7:3 c,- 0

1973 1974 1975 1976 1977 1978 1979 1980 1981 Year aintegrated circuits only, 1973.1979; weighted average of 12 manufacturers 1973-1979, 11 manufacturers1980, 1981. b1973-1980. Weighted averages for 11 U.S. merchant semiconductor manufacturers;1981 estimated: SOURCE: Unitsd States-1973-1977, Bureau of Census; 1978-1981, Department of Commerce and Semiconductor Industry Association. Japan-1973-1979, Japan Fact Zook '80 (Tokyo: Dempa Publications, 1980), p. 203; 1980, 1981, Japan Economic Journal, 42 34 International.Competitiveness in Electronics corporations, as meLaured by the ratio of debt than, on the average, between the electronics to equity or debt to total capital, helps primarily industries of the United States and Japan. in terms of taxation. In Japan as in the United The apparently high costs of capital in the States, interest can be written off as an expense United Statesas reflected in high interest, (while dividends paid to shareholders cannot); rates, stemming in the past from expectations therefore higher proportions of debt reduce of continued price inflationdo have a serious corporate tax bills. That banks in Japan lend consequence: High and uncertain interest rates willingly to highly leveraged firms places these in the United States tend to skew investment banks in a position more like that of equity- decisions toward short-term undertakings. holders in the United States: Japanese banks Although no one knows how to measure or ag- absorb higher risks than American banks, but gregate the time horizons of business execu- the impacts of this, by itself, on the competi- tives in any meaningful waymuch less com- tiveness of Japanese companies are small. Fur- pare those of American executives with their thermore, leverage ratios of Japanese firms counterparts in West Germany or Japanall have been slowly declining over the years else the same, interest levels that fluctuate un- one example among many of the gradual move- predictably will act to shoe ten time horizons. ment of the Japanese economic system toward Investments with longer payback periodsfor convergence with other advanced nations. example in basic research or in advanced pro- Likewise, the unusually high rate of personal dudtion equipment will appear less attractive. savings in Japan has impacts at the aggregate To the extent that capital markets in the United level which are only loosely coupled with costs States continue to mirror expectations of high or capital for individual firms. These costs vary and uncertain interest rates, the future compet- widely across the Japanese erktronics indus- itiveness of American industries like electron- try, just as in the United States. Indeed, cost ics may suffer. and availability of capital differ more from firm to firm within the U.S. electronics industry

Human Resources

Business enterprises depend on capable peo- in programing and related jobs have degrees ple for tasks ranging from assembly line work in other fields. to service and repair of their products, design and development, and general management. Quantity and Quality From the standpoint of international competitiveness, the larger the pool of qualified peo- For several years, during which engineering ple a firm or an industry can draw from, the graduates in all disciplines were in short sup- better. An ample supply of engineers and tech- ply, the U.S. electronics industry experienced nicians means that companies will have the a scarcity of entry-level electrical engineers and luxury of picking and choosing, while from the computer scientists. In the short term, demand employee viewpoint, salaries may be de- has droppedlargely because of recession pressed. A small pool means potential short- while the supply continues to rise, fed by ages, most likely of specialists, perhaps driv- swollen undergraduate engineering enroll- ing organizations to move people laterally to ments. The longer term pictureincluding meet their needs. Soaring demand for comput- prospects of continuing shortages of software er professionals, for example, has drawn in engineers, integrated circuit designers, and many people without formal training in the others with specialized skillshas not changed. discipline; about two-thirds of those employed Moreorier, the supply of grey-collar workers for

4d: Ch, 1-:-Part B: Extended Summary 35 the electronics industrytechnicians, drafters, Deteriorating laboratory facilities create a and designers, field Service repairmen, labo- second bottleneck. Engineering education is ratory aidesmay also be short, although quan- expensive; curricula include numerous labora- titative data on supply and demand for such tory courses, as well as heavy use of computing jobs are scarce. There is, needless to say,no facilities. Keeping laboratories relatively cur- shortage of unskilled assembly workers; the rent, so that students get some experience with heart of the problem in this, as in a number of up-to-date equipmentinstrumentation, small other American industries, is an excess ofun- computers, applications of microprocessorsis skilled workers coupled with sporadic short- a long-standing problem that has grown_m_qrse ages of those with higher levels of education in recent years. and training. If the trends outlined above continue, serious The scarcity of recent U.S. graduates inen- harm to the competitive prospects of, many gineering has been real, extending to virtually American industries could result. all specialties. Its origins lie in low enrollments during the early and middle 1970's (see ch. 8). Since then, engineering enrollments havere- Continuing Education and Training bounded to record levels. Educational resources have not kept pace, with the result that In contrast to constraints on supplies ofnew a substantial number of engiheering schools engineering graduates, the United States has have had to limit the numbers of students ad- hundreds of thousands of mideareer engineers Milted. Not only is supply constricted, but the already in the labor market. If the half-life of quality of engineering education is suffering. -a college education in engineering is,say, 10 years, upgrading these peoples' skills offers Shortages of teaching faculty have con- vast opportunities both for individuals and for strained engineering education more thanany U.S. industry. In some cases periodic short other factor. Despite undergraduate enroll- courses or self-study may be enough to boost ments that have nearly doubled over the past people along chosen career paths; in others, decade, trends in graduate engineering study they may wish to move laterallye.g., froman- have been nearly flat. In particular, students alog to digital circuit design, from hardware have been reluctant to enter doctoral programs. design to software. As pointed out in chapter Fewer Ph. D.'s in engineering were graduated 8, little data exists on the frequency with which in 1982 than in 1972. Nearly half thosenow re- engineers take advantage of opportunities for ceiving Ph. D.'s from American engineering continuing education and training; it appears schools are foreign nationals. Recent Ph. D.'s that most who do are recent graduates, and that have been avoiding teaching careers, for which those with the 'greatest needi.e., people "10 the doctoral degree is today virtually manda- years or more out of schoolrarely pursue con. tory. Not only are salaries low relative to in- tinuing education beyond the occasional (and dustry, but new teachers can anticipate heavy seldom very challenging) short course. Several course loads as a result of high undergraduate implications follow: 1) the rewards of.pursu- enrollments and the faculty shortages that ing continuing education and training iren- already exist. Coupled with uncertain pros- gineering could be lowe.g.,..employersmay pects for research support and a lack,of pros- not support such activities extensively, prefer- pective graduate students of their own, univer- ring to hire new graduates with the skills they sity teaching is no longer an attractive prospect need at lower salaries; 2) the programs avail- to many who in earlier years would have been able may not be attractivei.e., working engi- prime candidates. The result is 1,400 to 2,000 neers may perceive them as academic and un- unfilled vacancies onthe faculties of U.S. en- related to their jobs; 3) the quality of programs gineering schools. may vary quite widely, so that those who have

4 4 36 International Competitiveness in Electronics

el or hear about bad experiences are reluctant to More broadly, programs of all types aimed try again. at vocational education in technical fields appear to need reexamination and modification The paucity of information on this subject if the quantity and quality of graduates is to is in itself disconcerting, but it appears that grow.. In the United-States, as many as-8;000 all of these factors are at work, and others as public and private schools offer vocational- well. Certainly, existing incentives seem high technical education and training (compared enough to motivate only those with unusual with roughly 300 engineering colleges). The ability or perseverance. While some companies quality of the courses and programs offered by have devised effective programs for encourag- these institutions varies widely. Activities are ing employees to maintain and improve their fragmented, with little detailed information skills, others do little or nothing. The picture available even to form a baseline for analysis. is likewise mixed among educational institu- One point is clear: the fraction of the labor tions; some engineering schools have devel- force in U.S. manufacturing industries with oped aggressive outreach programs aimed at formal training in technical fields (through ap- providing high-quality coursework for tech- pi'enticeship programs or schooling) and/or nical professionals, those who are seeking ad- credentials (e.g., certification granted after ex- vanced degrees and those who are not. Contin- aminations) is far lower than in a number. of uing education programs offered by profes- other industrialized nations, including West sional societies as well as profitmaking orga- Germany and Japan. While correlations with. nizations vary considerably in quality. The on-the-job ability may be imperfect, the preva- quickest, surest way of providing the numbers lence of such programs in other countries is of qualified engineers needed to maintain the good evidence of a commitment by' individuals, competitiveness of American industries like governments, and business enterprises to build- electronics is to make high-quality continuing ing a labor force that will help maintain the engineering education more widely available competitive ability of technologically based in- and attractive to midcareer professionals. dustries into the future. So far this commitment Despite recent difficulties, engineering edu- has been lacking in the United States. cation in American universities remains best in the world. In part because schoolsanda nd Congressional 1 o,atiership could have a ma- universities in some countries do relatively jor impact. As the pace of technical advance poorly at preparing their graduates for careers in industries like electronics continues or accel- in industry, foreign companies resort more fre- erates, workers at all levels will face new de- quently than U.S. firms to internal and on-the- mands on their capabilities. Given the increas- job training. Extensive company-run training ing disjunction between the skills of the U.S. programs are prominent in the Japanese elk- labor forcewhat 'people are capable of do- tronics industry, where continuing education ing=and the skills that industry needs, the is widespread among blue-collar and grey-col- American economy seems bound to face .in- lar employees as well as white-collar profes- creasing problems in meeting its manpower sionals. One way for the United States to in- needs, as well as controlling unemployment, crease its pool of skilled grey-collar workers unless progress can be made in training and would again be to develop a more effective ap- retraining. A company might, for example, proach to continuing education and training. lend an employee the money to cover vocation-' Such programs will be more effective _where al schooling, retraining, or an advanced degree closely coupled with prospects for upward mo- program, with ropayment forgiven if the em- bility within organizations. At present, the ployee remains with the firm for an agreed probability that an unskilled worker in an elec- period. Tax policies and other instruments of tronics firm will be able to move tip to a higher Government support could increase the incen- paid position is small. tives for both corporations and individuals.

fit, Ch. 1Part B: Extended Summary 37

Public policies might be designed to lessen the tinues to shift from manufacturing toward serv- risks that companies ::;ponsoring education and ices, and toward more knowledge-intensive in- training for their employees would lose their dustries, the American labor force will need 'investments when people switch jobs. The Fed- to be prepared for technology-based jobs or risk eral Government could provide incentive doing without. Even those performing un- grants to the States, to be raitched with cor- skilled work will be in a position to make porate support. greater contributions to productivity and competitiveness, while enhancing their own job se- Preparation for Work in curity and job mobility, if they are comfortable Technology-Based Industries .with numbers and quantitative reasoning, and have a basic understanding of the physical At the root of many of the present and vo- world. spective difficulties outlined above lies pour Part of the problem is again a shortage_pf-- preparation in science and mathematics pro- teachers; secondary schools are being stripped-- vided by the public schools. Leaving aside the of their science and mathematics instructors large number of functional illiterates among by the attractions of higher paying jobs in in-: U.S. high school graduatesan illiteracy rate dustry. But the fundamental point is this: a stu- that some estimates place as high as 20 per- dent who opts out of scienceand particular- centand the one-quarter of this age group that ly mathematicsat an early age has made a vir- does not even complete high schbol, many tually irreversible decision, foreclosing a wide good students get little education in science or range of options in college and in his or her mathematics once they reach the upper grades. career. If American students continue to turn The number of students electing courses that away from mathematics and science at second- are prerequisites for careers in technical fields ary and high school levels, the United States is low and still falling; even those who choose will find itself with an even greater fraction of science often shy away from physics and chem- technological illiterates in the adult population. istry in favor of biology or geology. Technol- Already, the Nation finds itself with few lead- Ogy, as opposed to science, is invisible within ers in industry or Government who grasp the the public schools. As the U.S. economy con- workings of technology. Management and Organization Patterns of organization and management in 1 business enterprises mediate between the skills aleloommie and abilities that employees bring with them to the workplace and outcomes in terms of competitive firms and industries. How well companies utilize the talents of the people they hire is quite as important as how good these people are to begin with. Thus management, style and philosophy becomes a second critical element in human resources for the U.S. elec: tronics industry. While American management includes a "human relations" or participative management tradition, employee involvement tends to be honored in theory more than in practice. Techniques flowing frbm scientific management, the other main tradition, remain Photo credit: Digital Equipment Corp. dominant in U.S. industry. The recent vogue Computer-aided engineering design for Japanese management practices repres9nts 46 38 International Competitiveness in Electronics a swing of the pendulum toward the human re- and demanding tasks of integrating robotics lations pole, offering paths to improved com- and other forms of programmable automation petitiveness for some U.S. fh ms, though no into the factory environment; companies that sure cures. learn to utilize programmable automation most effectively will reap substantial competitive Within a given countrywhether the United dividends in the future. States, Japan, or one of the European nations electronics firms show a good deal of diversity Stress on human relations is certainly not in manageMent style. Nonetheless, successful unique to Japanese organizations, but is more electronics companies in the United States and consistently visiblenotably among large com- Japan exhibit more similarities than differ- panies characterized by low labor mobility and ences. Despite the current fascination with the "lifetime" employment (ch. 8). That workers at "secrets" of Japanese management, uniquely all levels tend to spend much if not all of their Japanese traits are rare even in the cruder ster- careers within a single organization creates eotypes. If the differences between firms with- strong incentives for internal training, job rota- in each country are often greater than the dif- tion, and other steps aimed at improving peo- ferences between countries, and clear-cut dis- ple's skills and preventing stultification. A tinctions between management styles in the:- number of successful American electronics United States and Japan less common than firms also go to considerable lengths to retain often assumed, two features of Japanese man- their employees, even in periods of business agementdo,stand out: first, reward-structures downturn. Rather than treating the labor force in Japanese companies create incentives for tal- as a variable cost, such firms, in both countries, ented people to huild careers in manufactur- regard their workers as a resource to be re- ing; second, Japanese organizations tend to tained and nurtured although economic condi- stress human relations more consistently and tions might seem to point toward layoffs. Ac- more effectively. complishing this implies more than keeping Generally speaking, manufacturing and pro- people at work and providing education and duction engineering get more attention and training. It also implies opportunities within more status in Japanese companies than Amer- the corporate structure to fully utilize present ican. This is one reason consumer electronics and potential skills without sliding into pater- and semiconductor firms in Japan could move nalism or coercion. A number of the highly swiftly to create perceptions that theinproducts publicized techniques associated with Japanese offered better quality and reliability. Often, as management could, in fact, be as fairly termed discussed in chapter 6, those perceptions, were manipulative as participative. firmly grounded in reality; although American firms have largely caught up, the strong institu- A renewed commitment to the development tional commitments in Japanese corporations and utilization of the human resources avail- to production engineering mean continuing able to American firms could make a major pressure in this area. Furthermore, that manu- contribution to the future competitiveness of facturing managers in Japan carry more weight the U.S. electronics industry, indeed may be in corporate councils means in at least some critical for the future prospects of this as well cases faster shifts into automated production. as other high-technology sectors of the Ameri- The importance Japanese companies place on can economy. Management practices in suc- manufacturing in their internal decisionmak- cessful organizations, whether American or ing also translates into a greater share of re- Japanese, tend to be associated with attention sources for developments such as the complex to human relations and employee participation.

4 ./ Ch. 1Part B: Extended Summary 39

Employment Continuing inadequacies in U.S. education change. In the United States, it is fair to say in science and mathematics will aggravate that jobs in electronics are becoming more skill- structural unemployment caused by technolog- intensive. Only in the manufacture of con- ical dovelopment and shifting competitiveness sumer products like TVs, a portion of the in- among American industries. In the past, tech- dustry that has been relatively stagnant, has the nical change has generally created more jobs ratio of blue-collar to white-collar employees in the aggregate thIn have been destroyed. Un- reniained high. In both computers and semi- fortunately, ;!here are no guarantees that con- conductors, the fraction of white-collar work- tinued technological changeespecially that ers is much greater and increasing. resulting from applications of microelectronics But a division into skilled and unskilledor and computerswill in the future lead to ag- white-collar, grey-collar, and blue-collar, not at gregate increases in job opportunities.In all the same thing given the high levels of know- Europe, the term "jobless growth" has come how associated with some but not all jobs in to describe the widespread phenomenon of each of these categoriesis too simple. It high unemployment despite expanding output. masks the increasing stratification and special- This may or may not have been happening in ization characterizing technologically based in- the United Statesthe evidence either way is dustries, not only electronics but the sectors scantybut structural unemployment is a reali- it feeds. The journeyman machinist may go the ty here. way of the tinker as computer-controlled ma- Shifts in the composition of the work force chines replace engine lathes. The skilled me- in electronics illustrate one of the conse- chanic who could rebuild such a lathe can quences of technological and structural probably no more fix the electronics of a

U.S. Employment in Consumer Electronics

120,000

u1 100,000 a) I.% e ...... ee . . e - N...... N. 80,000 N N.

MO. C)-.) . ... 60,000 ..1116 E z 40,000 All employees Prodaction workers

20,000

0 1960 1965 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981. 1982 Year SOURCES: 1960 -1965 -1977 Census of Manufactures. 1972-82Bureau of Labor Statistics.

A r--; 40 International Competitiveness in Electronics

U.S. Employment In Computer (and Peripheral Equipment) Manufacturing 450,000

400,000 All employees Production workers

350,000

a) 300.000 0 Cl E 0250.000

200.000

150.000

100,000 so* " 50.000

0 19551960196519671968 1970 1972 1974 1976 1978 1980 1982 Year SOURCES: 1955-65-1977 Census of Manufactures. 186682Bureau of Labor Statistics.

numerically controlled machine than program Many unskilled jobs are migrating over- the computer that controls it. Specialists not seasin electronics, mostly to low-wage coun- only design the parts to be made and program tries in Asia. Moves offshore by American cor- the computer, but pick the feeds and speeds, porations have attracted widespread attention, specify tool materials and cutting fluids. Gag- and opposition on the grounds of "exporting ing and inspection may be automated, rather jobs." Offshore assembly has been.much more than the responsibility of practiced hands with prevalent in semiconductors and consumer dial gage and surface plate. As skilled jobs electronics than in computers; even so, in both changeand at least some skilled work disap- sectors, other factors have often made greater pears along with unskilledpeople who have contributions to declining blue-collar job op- no skills to start with will face still more trou- portunities (see ch. 9 as well as app. B). Among ble in finding satisfying, well-paying employ- these factors, improvements in labor produc- ment. Those who cannot learn new skills may: tivity, many .stemming from investments in find themselves outside the labor pool. Upward automated manufacturing equipment, have mobility in the United States may. decline. generally had the greatest impacts. Moreover,

4a Ch. 1Part B: Extended Summary 41

transfers of production offshore tend to hi,: Predicted Growth Rates by Occupational Category driven by competitive pressures, domestic as in the United States Over the 1980's well as foreign in origin, which are largely out- Predicted increase side the control of individual firms. For inin employment stance, once a few U.S. semiconductor manu- Occupationa (1980.90) facturers began assembling chips in low-wage Paralegal 109% countries to cut costs, other suppliers had lit- Data processing machine mechanic 93 Computer oper,Itor 72 tle choice but to follow suit. When thepres- Computer systems analyst 68 sures are international, moves offshore may in Business machine service technician 60 some cases save domestic employment oppor- Computer programer 49 Employment interviewer 47 tunities over the longer term by helping main- Computer peripheral operator 44 tain U.S. competitiveness, though sacrificing Psychiatric aide 40 jobs in the shorter term. allon(nclusive; fastest growing occupations In electronics are listed together with selected occupations outside of electronics On italics) for comparison. Manufacturing by American-owned as well SOURCE: "Testimony Before the senate Subcommittee on Employment and Pro. ductivity, March 26, 1982, by Ronald E. gutscher, Assistant Commis. as foreign-owned companies has become wide- stoner, Office of Economic Growth and Employment Projections, Bureau of Labor Statistics," Productivity in the American Economy, ly dispersed internationally. But this is only one 1982, huarInge, Subcommittee on Employment and Productivity, Corn cause of unemployment in the United States. mittee on Labor and Human Resources, U.S. Senate, Mar. 19 and 26; Ongoing structural and demographic shifts are Apr, 2 and 16, 1982, p. 327. causing serious and persiStent adjustment difficulties. Peoplf.; with few skills or with obsolete skills will find diminishing job opportunities in many of the older U.S. industries. Ten mil- panies need good people to remain competi- lion and more Americans have been out of tive. Nonetheless, there has as yot been little work at a time when American industry has concrete discussion of what is needed to pre- been short of as many as a million employees pare people for future job opportunities; the with specific skills and abilities. In the ag- organizations and institutions that deal with agregate, and even considering multiplier' ef- such concerns tend to be dispersed and o op- fects, a million new jobs only dents the un- erate indepen tly of one another. Although employment problem facing the United States. the past few years have seen considerable crit- Yet from the standpoint of the individual, each icism of training programs said to be prepar- job counts. Policyrnakers may find themselves ing people for jobs that have already disap- 'unable to predict the causes and consequences peared, little usable information on such sub- of structural unemployment with any preci- jects in fact exists. Local control of secondary sion. This does not mean the problems cannot and vocational education is the traditional pat- be attacked. It means that adjustment measures tern in the United States. Educators and should aim to enhance job mobilityintra-firm schools of education seldom interact extensive- as well as inter-firmwithout depending on dely with industry or organized labor. Vocational tailed predictions of supply and demand byoc- education and training has little visibility at the cupational category and industrial sector. Federal level. Over the past two decades, schools have turned away from providing mar- The total number of jobs created over the ketable skills. Company-run training programs next decade in electronics and other high-tech- are limited in number and tend to be emergen- nology industries will' not be large. After all, cy responses to hiring shortfalls rather than the entire U.S. electr:/nics industry employs only everyday features of corporate organization. A about P12 million people today, and employ- thorough re-thinking of the American approach ment has not expanded as rapidly as output. to education and training, particularly for blue- Still, many of the fastest growing occupational and grey-collar workers, seems called for. Con- categories-hrthe economy will be found in this gress could decide to take the lead in reinvig- sector. The people who fill the new jobs will orating the traditional American commitment benefit; at the same time, U.S. electronicscom- to education and training.

99-1 1 1 0 83 4 5tl 42 International Competitiveness in Electronics

Trade

Trade policies pursued by the Federal Gov- time, uncertainty created by lengthy and in- ernment have affected the several portions of conclusive legal proceedings means that do- the U.S. electronics industry in radically dif- mestic firms could not know whether they ferent ways. Consumer electronics has suffered might eventually Iv' able to raise prices as a from uncertain enforcementsome would I stilt of antliltillWi.6 1.tties ied imports. nonenforcementof antidumping statutes, al- These added duties might have totale(" well though other parts of U.S. trade law ! ave been over $100 million; higher pricesgenerating called on to protect domestic firms L'om im- higher profits could, in principle, have aided port competition. American manufacturers of embattled American firms in revitalizing their semiconductors and computers have benefited businesses. from U.S. leadership over the postwar period Eventually, U.S. color TV manufacturers and in creating an open environment for interna- their suppliers did receive trade protection, in tionaltrade and investment. One of the the form of negotiated quotas on imports from 'strengths of American semiconductor and Japan, Korea, and Taiwan. Under the name Or- computer firms has been their global approach derly Ma rketing ereemohL; As),the to markets, a strategy aided by reductions in quotas followed escape clause proceedings barriers to trade and investment over the pa: filed in the wake of sharp rises in color TV im- three decades. Even though, semiconductor -imports during tL ff 1970's. 1 i'lfair trade practices ports frotn Japan have increased rapidly clor were t at The iJi,f: , speeded struc- ing the last few yearS, more than three -qtL ,t.. tural shifts already underway in the U.S. con- of U.S. semiconductor imports continue to c ,in- sumer electronics industry. Bylimiting im- sist o. interdivisional shipmeinia of American ports, they created incentives for foreign firms companies. In Japan, the largest exporter by to invest in assembly plants within U.S. bor- far among local computer manufacturers is ders. OMAs did little to revive the American IBM-Japan. consumer electronics industry; they did accelerate foreign investments, many of which Antidumping Enforcement would eventually have been made in any case. In effect, weaker American TV manufacturers Dumping complaints leveled at h -Torte's of driven from the market by imports have been Japanese TVs as early as 1968 have never been replaced by subsidiaries of foreign firms. While fully resolved. Dumpingselling imported TVs these subsidiaries help to maintain domestic at prices below those charged in Japan was employment, half or more of the value added proven under U.S. law, but legal challenges and interagency disputes have delayed final typically remains overseas. collection of duties for years. The Environment for World Trade During the 1960's and 1970's, Japanese TV manufacturers, followed by those in South A number of American computer firms that Korea and Taiwan, relied heavily on price cut- began as makers of office equipment, including ting to force their way into U.S. markets. None- IBM, maintained foreign operations before the theless, dumping-Is-as neither the sole cause nor war. During the postwar period, overseas in- even a primary- cause of competitiveshifts in vestments by American computer manufac- consumer electronics. The worldwide success turers expanded; subsidiaries of U.S. firms be- of Japanese consumer electronics firms amply came the backbone of computer industriesin demonstrates their ability, not only to manu- most parts ofthe,industrialized world. Today, facture at loW cost, but to produce reliable-TVsalong with-the-older companies whose product with good performance and product features lines still center on general-purpose main- that American consumers want. At the same frames, the major American manufacturers of Ch. 1Part B: Extended Stimmary 43 minicomputers also operate all over the globe. and guaranteed procurements. European na- Likewise, U.S. semiconductor firms began to tions, in particular, have sometimes used in- invest overseas at an early stage; these invest- vestment incentives to attract American elec- ments, beginning around 1960, were made for tronics firms in the name of jobs and tech- two reasons: 1) to supply foreign markets via nology. local production in industrialized nations; and 2) to cut costs by moving labor-intensive man- Over the past half-dozen years, spokesmen ufacturing operations to low-wage countries. for the U.S. semiconductor industry have frequently complained that the trade practices of Direct and Indirect Barriers some foreign enterprises have been unfair, while also voicing concern over government Foreign investments by U.S. computer and industrial policies in countries such as France semiconductor manufacturers, along with their and especially Japan. (U.S. computer firms continuing high levels of exports, have been fa- have seldom been as vocal over trade practices cilitated by a relatively open environment for or internal subsidies benefiting their foreign international trade and investmentchapter rivals.) Among the restrictive practices that still 11. Created largely under the auspices of the exist in many parts of the world, the relatively General Agreement on Tariffs and Trade high tariffs levied by the European Communi- (GATT), in which the United States has played ty (EC) on semiconductors=17 percentare a major role, this opening of opportunities for perhaps the most visible. One consequence has multinational firms via relaxation of direct bar- been to encourage investments within the EC riers to trade and investmenttariffs, import by American firms, but the European market qudtas, restrictions on flows of capital outward is in any case large enough that these invest- as well as inwardhas, on the whole, been of ment patterns could have been anticipated. In great benefit to the U.S. electronics industry. neither semiconductors nor computers have At the same time, relaxation of direct barriers European suppliers been very successful in ap- to trade has been accompanied by a simulta- proaching the EC market as a whole. With only neous increase in less direct obstacles and a few exceptions, local firms have exhibited controls. relatively fragmented patterns of production As the industrial and trade policies of foreign and sales. In contrast, American-owned enter- governments have evolved, they have swung prises have often done a better job of treating toward more subtle combinations of indirect Europe as a unified regional market. But import barriers, performance requirements, in- whereas the trade practices of Western Euro- vestment incentives, and subsidies. In some pean nations may have ended by harming the cases, these measuresdescribed in chapter ability of local firms to compete with the 10are intended to influence investment and Americans more 'than they have helped, the exporting. In others, t!;-- objectives are primari- situation has been vastly different in Japan. ly matters of domest_olicy: national security, employment, regional. developmc:at. Gov- Japan ernments intent on pursuing industrial policies For many years the Japanese Government ef- that will support local industries while attract: fectively protected the country's electronics in- ing U.S. dollars and/or technology can choose dustry, including manufacturers of consumer from a well-stocked arsenal: trade barriers products such as TVs, through controls over range from paperwork obstacles to "buy na- foreign investment as well as restrictions on tional" rules; performance requirements may imports. With only a few exceptionse.g., entail transferring technology, purchasing sup- IBM-JapanAmerican-owned computer and plies and materials locally, or exporting a semiconductor firms have had no more than prescribed fraction of production as a condi- modest success inselling their products tion for investment; common forms of subsi- through either exports or local production. A dies include R&D funding, capital preferences, complex of factors ranging from chauvinism 44 International Competitiveness in Electronics

U.S.-Japan Trade in Integrated Circuits Given the rapidly improving technological

400 abilities and competitive postures of Japanese electronics manufacturers, investment in Japan 350 by American firms appears vital for maintain- U.S. imports from ing U.S. competitiveness; while many in Jap- 300 Japan anese Government and industry will no doubt continue to oppose such investments, Japan's 0 .50 Government has officially endorsed liberalization many times, and should be held to these U.S. exports to statements in practice as well as in principle. Japan The Japanese market for electronics products -ra; 150 is now second only to that of the United States; for many types of products, sales within Japan 100 exceed those for all of Western Europe. Not only will local manufacturing help expand markets for American firms, enabling them to 0 -r compete more effectively with Japanese com- 1975 1076 1077 1978 1979 1980 1981 1982 panies in third countries as well as inside Year Japan, it will accelerate flows of technology SOURCE: 1983 U S. Industrial Outlook (Washington, D.C.: Dapariment of Corn from Japan to the United States. As in indus- merCe. January 1983, p. 29 -5. 1982 figures estimated. tries such as steel or automobiles, American electronics companies can now learn from their Japanese counterpartsand not only in to explicit government policies has impeded consumer products. U.S.-owned R&D labora- both exports and investments in Japan by tories in Japan could help compensate for per- American electronics suppliers. The negative sonnel shortages here, as well as improving ac- impacts on U.S. competitiveness have been far cess to the results of subsidized research pro- greater than those visible anywhere else in the grams such as the fifth-generation computer ef- world. foil. Full participation in the dynamic Japanese American companies have been able to sell electronics market is critical to the continuing products that the Japanese could not make for competitiveness of American computer and themselvesadvanced integi.atedcircuits, state-of-the-artsemiconductor fabrication Semiconductor Sales in the United States, equipment, some types of computers. But prod- Western Europe, and Japan ucts available from Japanese suppliers tend to be purchased locally, in part because of deep- Sales (billions of dollars) ly ingrained "buy Japanese" attitudes. Struc- 1974 1982 tural differences also play a role, particularly United States Discrete semiconductors . $0.88 $1.3 in microelectronics: the half-dozen large com- Integrated circuits 1.2 6.3 panies that produce most of Japan's semicon- $2.1 $7.6 ductors also consume perhaps two-thirds of Western Europe these same semiconductors; such a market is Discrete semiconductors .: $0.77 $0.77 difficult to attack from the outside. While Integrated circuits 0.52 1.5 $1.3 $2.3 foreign investment is in theory much :less re- Japan :_stricted today than in the past, only a few Discrete semiconductors _ $0.55 $1.2 American semiconductor and computer firms Integrated circuits P.59 2.4 have as yetestablished wholly owned opera- $1.1 $3.6 SOURCES: 1974Electronics. Jan 8, 1976. pp. 92. 93. 105. tions of any size within Japan. 1982Electronlcs. Jan. 13, 1983, pp. 128, 142, 150; Mar. 10, 1983, p. 8. Ch. 1Part B: Extended Summary 45

semiconductor manufacturers; Congress could the labor force; these make outcomes difficult lielp ensure that the Federal Government ac- to evaluate. Subsidies directed at commercial tively supports such endeavors by American technologies and typically rationalized as do: firms, which are fully consistent with this mestic support measures rather than export country's historic commitment to open trade promotion policies have few counterparts in and investment. Competition in,Japan on terms the United States. It is the justification in terms, perceived to be fair will yield dividends within of domestic objectives rather than strengthened the United States by creating conditions under export competitiveness that makes such poli- which American companies can better main- cies a problematic subject for international tain their competitiveness. negotiations. Public funds for R&D, the use of government procurement to favor domestic industries, and the many related instruments of Recent Developments industrial policy detailed in chapter 10 have Broadly speaking, the Tokyo Round multilat- become part of the conventional approach by eral trade negotiations, completed in 1979 and foreign governments. Countries in many parts.. implemented shortly thereafter, in the United of the world pursue such measures in hopes States by the Trade Agreements Act of 1979, of building their competitiveness in high-tech- are having generally positive though small im- nology sectors like electronics.. pacts on the American electronics industry. Given the growing prevalence of planned Continuing tariff reductions will help U.S. ex- programs of industrial development, virtually ports; accelerated duty reductions on semicon- all of which give electronics a prominent place,. ductor products and computers by Japan are it seems unlikely that continued U.S. attacks especially significant, though perhaps as sym- on such policies as "export subsidies" will have bol more than substance. much effect in arresting the trend. This is par- Nonetheless, tariffs are no longer the prin- ticularly true given the indirect and intangible. cipal barrier to international trade in elec- impacts of programs directed at infrastructural tronics. They are being replaced by indirect support, precompetitive technology develop- and nontariff barriers, including a wide range ment, or human resources. Many of Japan's inof implicit and explicit subsidies. In particular, dustrial policy initiatives have' been directed American electronics firms continue to com7 at overcoming structural obstacles such as plain over government-funded R&D programs limited labor mobility and a less than stimu- in Europe, Japan, and a number of developing lating working environment for technical pro- countries. Although the Tokyo Round yielded fessionals. Totaling the monetary value of such a new subsidies code intended to deal with this subsidies, even where possible, is an exercise and related issues, the prospects for substan- that holds little meaning. And, in the end, most tial progress seem slim. foreign governments will regard such efforts as too important to give up; certainly they do Taken one at a time, individual programs in not welcome them as legitimate topics of in- foreign countriesincluding such prominent ternational negotiations, bilateral, or multi- examples as Japan's VLSI R&D effort, and, lateral. While it is clearly in the interests of the prospectively, the fifth-generation and super- United States to press for clarification and computer projects now underwayhave often agreement on the "rules of the game," it may had no more than modest impacts on interna- not be very productive to devote a great deal tional competitiveness. At the same time, their of effort to combating on a case-by-case basis goals often include intangibles such as technol- what have become standard tools of industrial ogy diffusion or improvements in the skills of policy in other. countries.

54 46 International Competitiveness in Electronics

U.S. Industrial Policies As many observers have noted, industrial Faced with increasing competition in many policy in the United States has been a largely industrial sectors, slower economic growth, ad hoc construct of unrelated measures aimed and a multitude of adjustment problems, the at diverse object' vr,s; not infrequently, policy question for the United States has become: Can measures have vcaLl..:.ed at cross purposes or led the country continue with a de facto industrial to unanticipate i outrnmes. Seldom have they_ policy or is a new approach needed? The sur- represented core,.-.4,inv.,

4. Policies intended to promote the global the problems of a domestic industry into the competitiveness of American industries. context of the world economypolitical con- 5. An orientation that would defer wherever siderations can more easily predominate. To possible to the private sector when choices begin coordinating such decisions more closely concerning the development of industry carries two quite different implications: 1) are to be made. greater reliance by the Federal Government on These policy directions, examined in detail in empirically grounded analysis of industrial chapter 12, are by no means mutually exclu- competitiveness, productivity, and economic sive; they might draw, for example, on similar, efficiency; and 2) risks thatbeyond'influenc- policy tools in areas such as international trade ing policy decisions on a case-by-case basis, as or technology development. Nonetheless, they' happens alreadypolitical pressures will skew represent distinctly different thrusts: the goals the policy approach as a whole. The first is one differ even where the instruments are ali'7e. of the potential advantages of a mare coherent industrial policy, the second, one of the pit- The five alternatives, outl: -led below in the fallsa pitfall because companies and indus- context of the electronics industry, carry impli- tries in trouble, and their employees, have a cations for the entire economy, as well as the more obvious stake in policy decisions, hence political environment where any policy would bring more pressure to bear, than sectors of the have to be implemented. These broader dimen- economy on the upswing. sions are emphasized below because focusing too strongly on specific policy tools e.g., those The first two of the policy orientations listed addressing problems visible at the moment in above carry the greater risks of political deflec- elect7onicswould simply repeat the ad hoc tion. Ensuring the domestic market base for approach to U.S. industrial policies now U.S. industries could easily amount to nothing current. more than a protectionist response to trade pressures and the rise of competitive enter- Each of the five approaches has positive and prises in other countries. Basically an inward negative aspects. They can be usefully con- looking, defensive strategy, it equates import trasted in terms of differential effects on sec- penetration with damage to U.S. interests. An tors of the economy as well as susceptibility industrial policy centered on safeguarding to the political forces that corporations and American markets and American jobs would their employees bring to bear on the policymak- be largely congruent with the political forces ing process. The intrinsically political charac- that will always advocate protectionist meas- ter of this process, now or in the foreseeable uresfirms and industries in competitive de- future, has often been couched in terms of Gov-, cline, their employees, the communities and re- ernment's ability to pick and choose among gions in which they are located. "winners" and "losers." Early debates over industrial policy in the United States tended to Decline may be temporary and reversible, or focus on such questionsrather pointless given it may be the consequence of deeply rooted that many. Federal policies have always had shifts in the international economy that, over this effect. The Economic Recovery Tax Act the longer term, are iikely..to force contraction of 1981 treats some industries much more fa- regardless of public policy responses. A market vorably than others; trade protection has in re- protection strategy implies, first of all, deter- cent years been extended to manufacturers of mining whether protection is needed because color TVs, automobiles, and clothespins; politi- of short-term problemswhich might range cal pressures routinely affect decisions on from macroeconomic dilemmas to misjudg- public works and defense projects. ments by corporate managements. For such reasons, temporary protection is sanctioned by When industrial policy decisions are made international trade la ''u under circumstances on an ad hoc basiswithout linking one sec- as described in chapter 11. Indeed, temporary tor of the economy to others, without setting trade restrictions might find a place in any in- 48 International Competitiveness in Electronics dustrial policy alternative. Longer term decline ican political system, but has consequences brings a different set of issues to the policymak- that are largely undesirable if a basic objective ing process; the options may range from man- of industrial policy is to improve U.S. competi- aging decline (via adjustment measures in- tiveness. Industrial policies are most likely to tended to ameliorate the most immediate prob- be productive and effective when they comple- lems)-to wholesale protection and subsidization ment ongoing changes in the world econ- (as several European nations have occasional- omye.g., by aiding structural adjustment. ly attempted). When industrial policies oppose long-term shifts in comparative advantage, they are gen- A critical industries strategywhether erally doomed to high costs, inefficiencies, and "critical" refers narrowly to military strength marginality if not ultimate failure. or carries sc,ne broader economic connota- tionwould also lead to a great deal of political This could well be true, for instance, in the jockeying among firms and industries bent on case of Federal actions that would steer capital demonstrating their criticality. Snch an out- to selected industries. Such policies have fre- come is virtually inevitable because few objec- quently been advocated by those favoring "re- tive criteria exist that would allow essential or industrialization," as well as a critical in- critical industries to be identified beyond the dustries orientation. However, targeting of in- broadest and most general level. Under virtual- vestment in a consc:ous waya key element ly any criteria imaginable, electronics would in many foreign industrial policiesseems an be judged vital to "economic security" as well unlikely prospect for the United States, if only as military security. Even so, when the industry because capital markets here work much bet- is disaggregated, judgments at finer levels im- ter than in most other economies. Moreover, mediately become difficult. the Federal Government's experience with in- Electronics would probably not suffer under vestment, leaving aside sectors such as hous- either a protectionist or a critical industries ap- ing, has been restricted mostly to aggregate proach; although backlashes by other countries measures and to a few well-known bailouts of are always a possibility, nations that import loubled corporations. Finally, the records of high-technology electronics products usually foreign countries that have tried to channel in- need them badly enough that they would pick vestment into industries intended as mainstays other alternate targets for retaliation. Even so, of economic growth and competitiveness are a number of other U.s. industries would be like- decidedly mixed. ly to benefit more. For at least some compenies, Everyone knows what the future growth in- the lob'iying involved would be business -as- dustries will be. The current list includes usual.. Large and powerful corporations experi- computer-aided manufacturing and robotics, enced in dealing with the Federal Government, biotechnology, new nonmetallic materials, defense contractors, and firms in heavily microelectronics, computers and communica- unionized industries would tend to have an tions. U.S. capital markets have been "picking" edge over smaller, technology-based concerns. these winners quite effectively. An industrial The more aggressive and outward looking high- policy intended to support future U.S. growth technology portions of electronics couldnot ex- industriesunder a critical industries rubric pect as much positive support as they might or some other policy approachwould have to get' under other policy decisions. do more. Specifically, it would have to search Under any of the five alternatives, political out cases where markets were not performing forces would bear heavily on policy outcomes. consistently well. These do exist. The time Firms and industries will always have strong horizons of markets may be shorter than desir- incentives to press for direct and indirect sub- able from a public standpoint (there are many sidies flowing from Federal decisions. This is examples in R&D, most notably in basic re- quite understandable, and built into the Amer- search but also in the development of generic Ch. 1Part B: Extended Summary 49

technologies that could benefita wide range Government can help build competitiveness by of firms -While being difficult to protector promoting evolutionary shifts in the i7.onomy, monopolize). Bottlenecks are always possible as well as by easing the negative impacts of ad- (the un,linhiguous successes of foreign indus- justment on particular groups and regions. trial policies often involve breaking bottlenecks). Response times can be excessively loag In terms of R&D, the chief difference be- (as in the case of the educational system, heavi- tween private sector and Federal Government ly dominated by government bodies which decisions lies not in the ability to evaluateop- create inertia and slow responses, while also portunities but in the longer time horizons that suffering from cloudy perceptions of future Government can bring to such questions. Mo- needs and opportunities). Such examples point tivated by social rather than private returns to to approaches that would not be explicitlysec- investment, unconcerned with capturing im- toral. mediate rewards, public policy initiativescan be formulated with a longer term view than pri- The third of the policy orientations consid- vate corporations take. This is as true forma- ered by OTApolicies that would providegen- ture industries like steel or automobiles as for eralized support for technology and infra- "sunrise" or growth sectors. One of the tasks structural development, cuttingacross sectors of an industrial policy oriented toward adjust- of the economywould reduce the leverage ment and infrastructural support would be to that special interests could exert by avoiding, find such opportunities and develop appropri- where possible, policies with strong sector- ate responses. To develop an industrial policy specific thrusts. Instead, the tools of first choice capable of attacking problems of this sort, the would have more aggregate objectivesnot Federal Government would need to understand only R&D and its diffusion, but education and industries and their workings ona concrete, training, open competition, structural adjust- practical levelthe level of the shopfloor and ment. At the same time, sectoral policies would the R&D laboratory as well as the boardroom. not be totally ruled out. The Government does not now have thisca- A variety of instruments are available: pability, a capability it would need in order to manpower training and retraining; new institutional mechanisms for technology development (emphasizing, forex- ample, cooperative efforts among Govern- ment, business, and universities); incentives as well as direct funding for research and development; the infrastructure for diffusing available technologies as well as new R&D results through the U.S. economy (including technologies from overseas); and policies aimed at stimulating capital formation and investments in new and productive technologies. By supporting the technological and humanresources underlying competitive industries, interest groups anglingfor special favors would have fewer obvious and attractive targets, at least in terms of immediate financial rewards. Primarily future-oriented, this policy orienta- Photo credit: Intel Corp. tion is based on the assumption that the Federal Silicon wafer after chip fabrication 50 International Competitiveness in Electrobics

implement with reasonable effectiveness any reciprocally keep the domestic market open, consistent and explicit industrial policy. while vigorously pursuing antitrust enforcement in the name of competition. Rather than As pointed out above, competitive industries resorting to bilateral trade negotiations, the depend on the human resources available, United States could continue to work toward while training and retraining are essential to multilateral agreements aimed at reducing bar- economic adjustment. But what, specifically, riers to tradein the current climate, primarily should people be trained to do? What kinds of nontariff and indirect barriers. Tax incentives skills will be needed 20 years from now? What could be used to reward competitive, export- are the best ways of reaching peoplealready oriented firms. While more direct forms of ex- in the labor force? Are institutional changes port promotion might also find a place, direct needed? Should the United States continue to measures always carry the danger of becom- leave training and retraining largely to local ing subsidieswhich, in the name of competi- initiative, or is a continuing but redefined Fed- tion, this policy orientation would seek to eral role needed? These are among the ques- avoid. Instead of protectionist measures for tions with which this third approach to indus- aiding troubled industries, the Government trial policy would have to come to grips. They might attempt to manage decline and en- illustrate the need for a well-developed analyti- courage restructuring. cal capability within the Federal Government. If interestrgroups in the United States see the Like the first of these five policy alternatives, Nation opening its ;own markets to foreign preservation of domestic markets, the fourth goods and foreign investment an intrinsic promoting the global competitiveness of U.S. part of the global approachwithout corre- industriescenters on trade issues. However, sponding openings in other parts of the world, promoting competitiveness implies an outward this option could invite a strong backlash. Even looking, export-oriented stance, an emphasis if the United States persuaded its trading part- on openness in international trade coupled ners to join in a thoroughly open and com- with stimuli for emerging, competitive sectors petitive world market system, the accelerated of the economy. Taking as its starting point the processes of domestic change might generate dynamics of international competitiveness c,f rong sentiments in favor of protection as well the rise and decline of industries over time as adjustment assistance. Open world trade has the global trade alternative would seek out, many attractions as one element in a more even accelerate, processes of change, attempt- cohesive U.S. industrial policy,- but by itself ing to keep American industries technologi- might not offer advantages great enough or cally and commercially ahead of their foreign visible enough to attract the political support rivals. To the extent that such policies hastened needed for implementation. the decline of other portions of the economy, adjustment measures aimed at speeding re- The last alternative is built around giving in- source floWs out of these sectors, as well as dustry a free hand, where possible, in decisions that affect productivity and international cushioning the impacts of decline, might also competitiveness. This alternative fits the recent be called for. mood in the United States: that Government The global approach to industrial policy i:ivolvement in economic affairs is counterpro- builds naturally on the traditional U.S. attitude ductive, that business activities should be de- that international trade benefits all parties and regulated,. that markets work best and indus- should be encouraged. It is the option furthest tries compete best when the Federal presence __removed from the common notion of industrial is minimized. Like the global trade alternative, policy as necessarily working against Ofienness it could mean more rapid- rises-and declines in trade. The Federal Government might not within the U.S. economy. Unlike that alterna- only continue to press for access for U.S. ex- tive, it implies less attention by Government ports and investments in other countries,but to structural adjustment and less support for 5., Ch. 1Part 6: Extended Summary 51

the efforts of American firms to export and/or of the economy are an unavoidable conse- invest overseas. Nor would protection against quence of any industrial policy, and it may be import competition be looked on with favor. better to confront such issues directly than try to avoid them, as this last alternative would in Such a policy approach would have to con- general do. While true neutrality can never be frOnt and resolve the following dilemma. Amer- achieved, an industrial policy ostensibly in- ican businessmen direct many of their com- tended to "get Government off the backs of plaints at foreign industrial policies that in- business" would more likely end up rewarding tervene in markets by, for example, encourag- those who could bring the most political pres- ing mergers or allocating capital. Spokesmen sure to bear. These interests would probably for U.S. industry often hold, on the one hand, be able to perturb the policymaking process that industrial policies in other countries are tax policy being only one exampleto their not only unfair, but serve to tilt the competitive own benefit, aided by the illusion that the balance by strengthening or even creating com- Federal presence was diminishing. Industries parative advantage. On the other hand, these with less political strength or sophistication same spokesmen frequently argue that Govern- would, in a relative sense, fare less well. ment policies could not do so here. Some of these statements may simply express a desire Indeed, it seems wishful thinking to argue for unfettered competition among all corners; against Government involvement in economic in other cases, they appear to imply that affairs, although not against counterproductive government actions are counterproductive in or excessive involvement. The fact is, of course, the United States but not overseas. In any that governments here and elsewhere do in- event. the fundamental question is: Given that tervene; it is part of their job. Moreover, as foreign governments are not likely to abandon economies grow more complex and more heav- their industrial policies so long as they consider ily dependent on advanced technologies, the them useful, can the United States counter forces that governments seek to modify or con- diem simply by avoiding policy interventions? trol may become more powerful, the need for government action greater. When, where, why, More positively, then. this fifth policy ap- howthe circumstances in which govern- proach might include tax incentives for capi- ments intervene, the effects of the involve- tal formation and investment, deregulation, mentare the crucial questions. and free competition. Control of inflation and What does this mean for industrial policy in macroeconomic stability wo! .-,ertainly re- the United States? First, more effective policies main a Federal. responsibilityi toser examina- toward industry in the United States will re- tion of re^.ent changes in tax' policy' points to quire relatively broad agreeMent on objectives. one of the. central issues raised by this alter- Second, the Federal Government would need native: Cali Government really be a neutral ar- to develop an analytical capability adequate to biter of economic competition? Past experience the task of reaching these objectives. Bothare giVes little evidence in favor of the proposition. efforts to which Congress could turn its atten- The 1981 Tax Act seems on balance to have tion. The first is largely a political task, the been a move away from neutrality in treatment basis of the argument that our standard of liv- of the various sectors of the economy. Noting ing depends on the international competitive that accelerated depreciation has varying con- ness of industries like electionics. The second sequences for manufacturers of consumer elec- demands that Government go beyond the large- tronics and semiconductorsand that these ly static and abstract economic perspective that 4wo.parts-of-the electronics industry are, treated.. in many agencies is now called on 'to justify quite differently than producers of heavy elec- policies adopted for-other red-sells: tricalmachinery, much less nonelectrical The political, environment in the United machineryindicates some of' the potential States makes movement toward a more con- problems. Differential effects on various parts sciously developed industrial policyfollowing 52 International Competitiveness in Electronics any of the five alternatives outlined above ulated by countries experimenting with indus- not only slow and painful, but an endeavor that trial policies, and learning to use them more risks being turned to ends far removed from effectively, a pragmatic orientation. , by the economic efficiency. (This is not to implyat United States, grounded in empirical analysis, economic efficiency is the only goal of indus- could be viewedby Congress and the Federal trial pOlicy, but that one of the purposes of a Government as a whole, and by both parties more coherent approacli would be to bring this as a vital support for our own economy. Such and related objectives closer to the forefront.) an attitude toward industrial policy would help But even where decisions are made largely on to ensure that the U.S. electronics industry and political gioundsas will frequently be the other high-technology sectors would get their casea more explicit industrial policy could fair share of the resources neededlo compete help frame the questions, bound the responses, effectively in world markets. It isalso the best increase the probability that individual policy hope, in the longer run, for older industries instruments function as expected and in- ranging from primary anetals to machine tools tended. Given an international economy pop- and textiles. CHAPTER 2 Introduction

62 Contents

1.:kctrunics as a high-Techntdogy Industry 56 11)fviduintient 56 Clania!titiut! 57 Structural 1.:itangt 58 Thc Importance of Competitiveness 60 industrial P 61

. 62 CHAPTER 2 Introduction

electronics industry provides exam- ble harbinger of their own fate if the U.S. Gov- ples that can support almost any perspective ernment does nothing to support them in their on competitive trends in the American econ- competitive battles with foreign (i.e., Japanese) omy over the past decade. That portion of the rivals. At the same time, American semicon- industry manufacturing computers has been a ductor firms share with our computermanu- champion of U.S. economic strength both facturers a deserved reputation as worldwide domestically and internationally. Here and leaders in technology, innovation, and entre- abroad, American computer firmsparticular- preneurial zeala reputation which the 1980- ly IBMhave been symbols of technological 83 round of new startups in Silicon Valley can prowess, market power, and multinational only enhance. marketing and production. In Europe, U.S. These three portions of the electronics indus- computer manufacturers have been models to trycomputers, consumer electronics, and be emulatedfor in,ii:;enotis companies like semiconductorsare the focus of this report. ICL in Great Britain, or for joint ventures such But other parts of the industry could illustrate as CH-Honeywell Bull in Franceand targets many of the same themes. Electronic compo- to be displaced with the aid of national in- nent productionswitches, resistors and ca- dustrial policies. In Japan, American computer pacitors, printed circuit boardshas been firms have been explicitly depicted as the en- moved to offshore locations as part of the re- emyIBM as a stateless, global giant, with sponse to competitive threats from imports. Japanese firms urged to mount fierce efforts Professional and industrial equipmentinstru- against it. Meanwhile, in the United States, the mentation, industrial process control, medical Department of Justice had in 1969 begun an an- electronicsis a continuing U.S. strength, but titrust suit aimed at dismembering IBM, a suit again the technological leads that American that was finally dismissed 13 years later. firms once held have narrowed. In telecommu- American consumer electronics firms have nications, American firms have lost out in ;;ev- been pictured much differentlyparticularly eral promising developing country markets. the old-line manufacturers of televisions and While boundaries between information proc- other home entertainment equipment, such as essing and information transmittal have been Zenith and RCA.-Many firms in this part of the blurring for years, and communications is cer- industry have seen themselves as victims of un- tainly one of the central electronics-related por- fair trade practices by overseas rivals, primarily tions of U.S. industry, this report touches on Japanese. Foreign firms selling TVs in this communications only in passingnot because country have been accused of dumping (and this portion of the industry is unimportant, but found guilty of this), attempted monopoliza- only to keep the study to manageable proportion, and of receiving subsidies from their own tions. governments. To other observers, the U.S. con- The breadth and diversity of the electronics sumer electronics industry has been a victim industry contrasts with industries such as steel, of management failures, has lacked the will to which are often pictured as monolithic. Even compete internationally, has.ceded some seg- here, however, specialty steel Sand noninte- ments of its markets too easily to imports, and grated "minimills" have prove% notable excep- has lagged in adopting manufacturing methods tions to the commonly accepted no'ion of de- that could have cut costs and increased the clining U.S. competitiveness., Steel is an old, quality and reliability ui its products. 'Technology and Steel Industry Competitiveness (Washington, D.C.: U.S. Congress, Office of Technology Assessment, OTA- Semiconductor manufacturers in the United M-122, Juno 1980); U.S. Industrial Competitiveness: A Corn. parison of Steel, Electronics, and Automobiles (Washington, States have, over the past several years, pointed D.C.: U.S. Congress, Office of Technology Assessment. OTA- 56 International Competitiveness in Electronics established industry compared to electronics; tor firms real, or are they better considered vet the electronics industry has roots going natural consequences of the growth and matur- back to the early part of the century, in con- ing of these portions of the industry? How have trast to biotechnology and genetic engineering policies adopted by the Federal Government for xvhich international competition has affected the industry? How do public policies hardly begunthough here, too, there are roots here differ from those of foreign governments, in fields like plant breeding and pharmaceuti- both in their forms and in their effects? To cals. Emerging industries like biotechnology what extent have foreign industrial policies are important for future economic growth; succeeded in strengthening the electronics inelectronics is critical right now. Moreover, dustries of other countries, in affecting the in- lessons learned from electronics might apply vestment and export strategies of American to older, "mature" industries such as steel, as firms, in replacing tariff and nontariff barriers well as to nascent sectors like biotechnology. to international trade with less visible but no less effective constraints? `:an governments What .can be learned from electronics, pa'- create comparative advantage? If the United ticularly the last 10 or 15 years? That is one States were to pursue a more consciously de- the questions this report attempts to answer. veloped industrial policy, what should be the Is the apparent decline of the American con- objectives in the context of a high-technology -sumer electronics industry irreversible? Are industry like electronics? How might the policy the threats to U.S. computer and semiconduc- tools be formulated and implemented?

Electronics as a High-Technology Industry

The electrical equipment and electronics in- ity will rise, the skill mix needed by thework dustries have been known for technical leader- force continue to shift. In service industries, ship and innovation since their beginnings at office and workplace automation will also dis- the close of the 19th century. While progress place people while creating new jobs needing in electrical equipmentthat which produces new s'cills. or utilizes electric poweris now mostly incremental, electronicsreferring to devices and Patterns of Development systems that operate on the information content rather than the power transmitted by an The portions of the electronics industry electrical signalremains a technology in rapid where American firms remain preeminent are flux. Developments in electrical machinery just those where the pafle of technological such as practical applications of superconduc- change continues to be most rapide.g., com- tivity can still promise significant gains in the puters and semiconductors. The United States efficiency of energy conversion and power has been a leader in both the technology and transmission; advances in electronics will have the science that underly these sectors: elec- effects that reach further, and affect the Amer- tronic properties of solids and the materials sci- ican economyindeed, society as a whole ences more generally; electrical engineering; more deeply. An obvious case will be the con- computer science and software engineering tinuing applications of distributed computing. and also in the development of new and suc- The impacts will be broad as well af deep cessful commercial products. Nonetheless, al- manufacturing industries as a whole will be though Americans have been among the lead- transformed by applications of electronics to ers in the technology and science ofelectrical automated production equipment. Productiv- machinery and electronics, many of the impor- Ch. 2Introduction 57

tant prewar developmentse.g., understanding like Taiwan, South Korea, Brazil, and Mexico of band gaps in solids and the dynamics of con- now factors in at least the lower technology ductionoriginated in Europe. segments of the electronics industrywere, The Second World War pushed electronics before 1960, simply irrelevant. to the forefront of engineering science, creating a momentum that still exists. Developments in Rising Competition radar and computing, both analog arid-digital: Much of the impetus this strong postwar start proved especially significant.2 Agaii many of gave to the U.S. electronics industry has now the advances came from Europe, particular- dissipated. Gloomy predictions for the future ly the United Kingdom, where considerable competitiveness of even the strongest sectors, strides were made in radar technology.3 How- such as semiconductors, have been heard. The ever. American industry was in a far superior business press reminds us incessantly that Jap- position to capitalize on these new technol- anese firms captured 40 percent of the U.S. ogies in the aftermath of the war. By the late market for 16 kilobit random access mem- 1950's, the United States had what appeared ory circuits (integrated circuits called 16K to be an unchallengeable lead in fields such as __IrtAMs), more than 50 percent for 64 kilobit cir- digital computers and semiconductors. cuits. Market analysts predict that Japanese Hindsight shows the more temporary nature manufacturers could have 30 percent of the of this lead, the result of an infrastructure for world computer market by the end of the technology and science that emerged from the 1980's.4 war not only intact, but strengthened, coupled In the past, competitors in countries like with an industrial base that was likewise far Japan relied to considerable extent on electron- stronger than in countries that had been either ics technology first developed by U.S. firms; allies or enemies a few years earlier. The push now they have independent capabilities and _created by new technologies, coupled with the need not follow paths broken here. As Japanese pull of war-starved markets in the United electronics companies have become less de- Statesmarkets that were eager recipients of pendent on American technology, their exports the products of these technologies, rather of microelectronic devices to the United States than devastatedcreated an environment for have grown faster than their imports of U.S. growth and innovation unmatched in the rest semiconductors. And, where once they ex- of the world. Meanwhile, trading partners and ported mostly discrete semiconductors and the potential competitors such as Japan, Great Brit- simpler integrated circuits, now firms based ain, and West Germany had to rebuild. Nations in Japan are exportingor assembling in the United Stateslarge-scale integrated circuits =H. H. Goldstine, 'lir. Computer From Pascal to von Neumann (Princeton. N.J.: Princeton University Press, 1972), especially Part Two. ------.- Kraus. "The British Electron-Tube and Semiconductor In- 4"No. l's Awesome Strategy," Business Week, Juno 8, 1981, dustry, 1935.62," Technology and Culture, vol. 9, 1958, p. 544. p. 84.

4111111111W

'Photo credit: Smithsonian Institution Harvard Mark I electromechanical computer, 169-44 58 2 International Competitiveness in Electronics

(ICs) at the leading edge of the technology.* Technological and Structural Change Japanese computer firms are not yet exporting large numbers of systems to this country, but The rather complex structure of the electron- clearly iwend to try. The government-sup- ics industry in the United States is described ported fifth-generation computer project is only in more detail in chapter 4. The diversity of the one recent signal of the seriousness of Japan's industry has already been pointed out; there efforts. are more than 6,000 electronics firms in the United States. Only a small fraction could legit- Needle 7,F to say, this resurgence by America's imately be called "high-technology" com- competitofs has not been an overnight phe- panies. But this smaller fractioncompanies nomenon, nor should it be unexpected. The building computers, designing and manufac- Japanese presence in consumer electronics be- turing large-scale ICs, supplying capital equip- gan to be felt in the 1950's with the transistor ment such as microprocessor development sys- radio; by the late 1970's, firms based in Japan tems or plasma etchers, developing software held, strong positions worldwide in audio packages for computersis a driving force for equipment, digital watches, calculators, and the rest of the industry, as well as for much of TV receivers. Their burgeoning capability in the rest of the economy. high-technology electronics builds naturally on earlier developments. By the standards of computers or microelectronics, consumer electronics cannot be con- Interactions within the industry often stimu- sidered high technology. Yet the manufacture late technological and commercial develop- of cathode ray (picture) tubes is a sophisticated ments; ICs have made possible new families process, and TV receivers are now designed of consumer products, such as hand calcula- around ICs, some of rather advanced design. tors. as well as cheaper and more powerful Digital TV and digital audi,) are on their way computers. Semiconductor devices are becom- to commercialization, while consumer prod- ing indispensable for the products of more and ucts are providing some of the first applications more industries outside electronics; emissions of speech synthesis; the same will be true of control systems for automobile engines depend voice recognition. Solid-state displays as re- heavily on microprocessors and related de- placements for picture tubes are a demanding vices; more than half the cost of an airplane technical challenge. Indeed, the low costs re- can be electronics. As one result, electronics quired for practical consumer applications technologyand particularly microelectron- create technological constraints that are, in icshas come to be widely regarded as critical their own way, more severe than those im- to a modern, competitive economy, hence ac- posed on designers in portions of the industry cess to this technology a vital strategic.weapon more commonly associated with high technol- of national industrial policies. Government at- ogy. At the same time, consumer electronics tention to computer industries goes back to the products such as table radios or conventional early years of this technology; a number of TV receivers are simple enough that they can countries began in the 1970's to subsidize semi- be manufa7.tured and marketed competitively conductor research and development; other:2 by firms in industrializing countries such as have, felt applications of ICsrather than South Korea and Taiwan; the same is true of capability for designing and manufacturing the many types of discrete semiconductor devices circuits themselvesto be more important, and and small-scale ICs. have channeled government funds to this end. As such examples indicate, even consumer electronics is changing more rapidly than in- ICs Mcorporate of the, order of hundreds of cir- dustries like steel or automobiles. Despite the cuit elements, large-scale ICs of the order of thousands to tens of thousands, very laixe-scale ICse.g., 64K RA Ms, 16 bit micro- pace of technological change, electronics is not processorsof the order of a hundred thousand. See ch. 3. only much larger and better established but Ch. 2Introduction .59 more stable and predictable than, for instance, Although many of the major technological in- biotechnology. But again, the industry is far novations in electronics have originated in the from monolithic. Consumer electronics has ori- United Statese.g., color TV, computer time- gins in the 1920's, when radio broadcasting be- sharing, most of the important developments came widespread. The computer and semicon- in semiconductor devicesAmerican firms ductor sectors are basically post-World II have not always been the leaders when it phenomena, though many of the leadni8 com- comes to product innovations that depend, not panies--e.g., IBM, Western Electric, Me:4:01:11a necessarily on new technology, but on product have prewar origins. Thus, the three porl;r1f. planning, engineering design, production of the industry on which this report concen skills, and marketing. Although transistor tratlis include examples of both well-estab- radios were developed in the United States, it lished, "mature" sectors, and more volatile, was Japanese products that reshaped the en- rapidly growing, technology-driven sectors. tire audic, market.5 Analogous strategiescon- There are lessons to be learned from each. centrating on product design and engineering, One of the lessons that even a superficial look originally perhaps imitative, rather than high at the computer industry teaches is the impor- technologyhave led to success by the Japa- tance of marketing, sales, customer support nese in fields such as cameras and automobiles. and service, and related nontechnical factors Japanese firms, aided by their skills at low-cost even in a technology-driven industry. IBM hasnia:aufacturing, have recently done much bet- been a dominant force worldwide in comput- ter at this than companies in the European na- ers since the beginning of the 1930's. But IBM's tions with which the United States also com- strength has beennot only hardwarebut petes. marketing, software, and customer support. In many cases, IBM's competitors have offered It White, "Management Criteria for Effective Innovation," considerably more computing power for a giv- Technology Review, February 1978, p. 15. en price, but IBM has only slowly lost market share because of its many strengths beyond hardware technology. In some contrast, other U.S. electronics firms have sometimes seemed to rely primarily on advanced technology to win markets. As other countries catch up in technical capability, a technology-based marketing strategy may no longer be enough. In microelectronics, for example, the ability to pack many circuit elements onto a single integrated circuit chip is still important, but competition is more and more a matter of the systems which the ICs comprise or can be integrated into. Moreover, as microelectronics technology continues to evolve, one path to competitive success will be the creation of new end-products incorporating ICs. The skills required for this differ from those needed to establish and maintain leadership in the underlying technology, as shown -by_the_examples of pocket calculators or digital watchesand also by the failure of the West German electronics industry, which has access to excellent fundamental technology, to develop into a strong in- Photo credit: Bell Laboratories 60 International Competitiveness in Electronics

Thus, overall, the cushion that greater tech- ment. Electron-beam lithography is now essen- nical capability once provided U.S. products tial for making the masks that are, in turn, used is eroding. And of course, in some technologies to fabricate large-scale ICs (in a few cases elec- the United States has never had an advantage. tron-beam lithography is applied directly in In optical communications, for example, Jap- fabricating the chips). AlthOugh the equipment anese companies have always been near the has its roots in technology developed for scan- .forefront. ade rship_in_ele c tronic s_e quipm e nt_ ning- olectron-microscopes.---virtually all of used for certain types of scientific research has which are designed and built in Europe or long resided overseasone example being elec- Japanthe United States has not thus far been tron microscopes. That this need not always handicapped. Several U.S. firms are, in fact, be a handicap is shown by current develop- leaders in electron-beam lithography. ments in electron-beam lithographic equip-

The Importance of Competitiveness

OTA's earlier comparison of steel, electron- which for many purposes must be further dis- ics, and automobiles provides background and aggregated. illustrations for many of the questions concerning competitiveness, economic efficiency, and" Considering the electronics industry itself, industrial policy that remain of concern to Con- competitiveness is one of the factors that deter- gress, to employees of the U.S. electronics in- mines, among other things: employment pat- dustry, and to the public at large.° The prac- terns within the industry (size of the work tical meaning of "competitiveness" in the con- force, wage levels, skill mix); balance of trade text of electronics is discussed in chapter 5. In for electronics products; and the value that pur- essence, the term refers to the ability of elec.': chasers of electronics products receive for their tronics firms located in one country to design, money. Electronics products are used by many develop, manufacture, and market their prod- industries-7whether components such as semi- uctsdomestically and by exportingin com- conductors costing a few cents, 'or capital petition with foreign enterprises. (For some equipment that sells for hundreds of thousands, purposes, subsidiaries of foreign firms that pro- even millions, of dollarsand can affect their duce and sell electronics products in the competitiveness. Computers are the most United States are considered part of the U.S. prominent example, but are far from alone. Nor industry, but in general the consequences of do they always fill the role of capital equip- foreign direct investment must be treated on ment; many smaller computers are integrated a case-by-case basis.) into more complex electronic systems. Indus- The:competitiveness of an induitry like elec- trial process control, scientific equipment, of troniCs is important not only intrinsically, but fice machines, and communications apparatas also because of interactions with other parts are further examples where electronics or elec of the economy. Still, there is no -meaningful tronics-related products can affect the compet- itivenessmore generally, the economic per- way of measuring the competitiveness of an en-.. tire economy. Competitiveness must be exam- formanceof other parts of the economy. ined on an industry-specific basis, although it can also be difficult to generalize about an in- On the bioadest levels, then, the competitive -. dustry as large and diverse as electronics, ness of the electronics industry affects aggregate employment 'levels, trade balances (more Industrial Competitiveness: A Comparigonof Steal.Elec- importantly, the ability to pay for imports), and tronics. and Automobiles.' op. cit. living standards. How this industryfares in in- 'Ch. 2Introduction 61

ternational competition influences thetypes of ing operations are carried outin other coun- jobs available, the country'smilitary strength, tries to take advantage of low labor and overall rates of economic growth. costs, is In turn, common in electronics). The ability to raiseex- the health of the aggregateeconomy, the qual- ternal capitalwhether equityor debtand to ity and quantity of employees availableto firms generate capital for reinvestment through sales, in the industry, the market provided by themil-- is crucial to firms inany industry, but particu- itary, are among the factors that determinethe larly when marketsgrow as fast as those for competitiveness of American electronicsfirms. semiconductors and computers. Aswith off- Ultimately, however, the competitivenessof shore manufacturing, which is favoredby U.S. any industry in the United Statesor tarifi laws, sources andcosts of capital for elec- where=depends on thefforts ofindividual tronics-firms-are affected by public policies firms. Policies adopted by the FederalGovern- tax policies and many others, includingthose ment influence these efforts inmany ways, aimed at controlling inflation. often indirectly. Foreign industrialpolicies are With respect toconsequences of shifts in part of the same.context. Among themore important domestic measures competitiveness, employment receives the are those dealing most attention in this reportboth interms of with taxes, Government spending, andmone- job opportunities and in tary policy, as well as research and develop- terms of the skills needed. This and many other topicsare dis- ment (both basic and applied), international cussed, where possible, in theconfect of inter- trade, and many types of regtilatorypolicies. national comparisons drawnbetween the Sc:.neticries Federal policies affect onlyone or United. States and its tradingpartners and a .kiWidustriese.g., regulation of TV broad- rivalsusually one and thesame. Japan, at casting,Others are broader. Tax treatment.of present, is the home of the strongestcompeti- income from overseas investmentsaffects tors, in electronics as in many other industries. firms with multinational operationsregardless Japanese firms are likely tocontinue to be the of industry. Some policies affect theentire chief rivals for U.S. electronicsmanufaCturers economymacroeconomic policiesor those over the remainder of the century. But several dealing with education.',"= European nations have strongtechnological Generally within the province of individual bases in electronics,as well as supportive'gov- firms are factors associated withmanufactur- ernmental policies. And rapidlyindustrializing ingincluding costs, the quality andreliabili- countries will rise in competitive strengthin ty of finished products, and decisionsto man- the future; TVs from Taiwan andSouth Korea ufacttire domesticallyor overseas (offshore as- are growing factors in the U.S. market. sembly, wherein some but not allmanufactur-

-IndustrialPolicy

Public policies that affectcompetitiveness The United States does not at can be considered elements of "industrial pol- present have icy."' The term is intended a coherent or consciously developed industrial to embrace Federal policy, in contrast to nations suchas Japan or Government policies of whateverorigin that France. This is not to imply that industrial affect the activities of private industry, pol- particu- icies like those of the Japaneseare necessarily larly its competitiveness, productivity,and economic efficiency. effective in promoting internationalcompetitiveness, but simply that the United StateShas not attempted to develop a coherent industrial 'Ibid., ch. 8. policy. Instead, policies affectingindustries 62 International Competitiveness in Electronics and their competitivenesshave been formu- be relaxed, government procurements chan- lated and implemented on an ad hoc basis. As neled to favored companies, which in some a result, industrial policy inthis ,country has cases may be publicly owned.Nationalized been fragmented, sometimes contradictory, enterprisesan increasing presence in sectors often inconsistent and lacking in continuity. like banking or energy production although not a major factor in electronics--coupleindustry These characteristics of U.S. industrial pol- and government even more tightly.8 American icyreflecting our pluralistic political tradi- businessmen increasingly complain of the diffi- tionshave sometimes served the American culties involved in trying to'l compete with such economy well, lendingflexibility and the po- ventures, which need not make profits, or may tential for innovative response to changing cir- have unusually long profit horizons. cumstance. But the OTA report cited above concluded that this approach to industrial pol- The variety and complexity exhibited by icywhile it might have been well-suited to an present-day national industrial policiespartic- earlier period when U.S. industries were rela- ularly the difficult questions of when govern- tively, isolated from foreign competition, and ment support measures should be judged subsi- possessed advantages in technologyin more dies that distort international tradehave ham- recent years has too often contributed tode- pered efforts by international organizations clines rather than improvements in competi- such as the General Agreement on Tariffs and tiveness. Trade (GATT) to fit remedies for many of the possible means of "unfair" competition into the Foreign industrial policies often include di- body of international trade agreements. As one rect subsidies to industriesperhaps tomain- result, bilateral agreements are becoming more tain employment, or for reasons of national commonexemplified by the Orderly Market- security. Export incentives and protection for ing Agreements negotiated by the U.S. Govern- domestic industries are common. Foreign in- ment to control imports of TV receivers from vestors may face a complex set of carrotsand several Far Eastern nations. sticks. Cooperation among nominally competing firms may be encouraged. Governments in For a survey, see R. P. Nielsen, "Government-Owned Busi- some countries haveengineered "national nesses: Market Presence, CoMpetitive Advantages and Ra- champions" in attempts to increase competi- tionales for Their Support by the State," American Journal of tiveness. Restrictive business regulations may Economics and Sociology,"vol. 41, 1982, p. 17.

Issues

As emphasized above, a vast number of Fed- fit within the overall framework of U.S. eral Government policies in some way affect foreign economic policy, the meaning of the international competitiveness of the U.S. "reciprocity" for an industry like elec- electronics industry. Among the more impor- tronics, barriers to investment in foreign tant are: electronics industries; Government policies affecting capital for- Government support for commercial (as mation for the economy as, a whole, and, opposed to military) R&D, ranging from firms in the tax policies intended to increase levels of more directly, the ability of research spending or encourage commer- electronics industiy to generate and attract cialization to direct support; capital for expansion; trade policies dealing with exports as well regulatory policies that may affect the as importse.g., the ways inwhich meas- competitiveness of the U.S. electronics in- -tha fin° caiorfrnnirC inch ictry dustrye.g.. antitrust enforcement; Ch. 2introduction 63

the availability of enough people with ade- to firm. The remainder of this report attempts quate levels of education and training, par- to deal with such complexities; at thesame ticularly engineers and skilled_workers time, of course, public policies continue to such as technicians, as well as the Govern- evolve and changewitness the 1981 taxact, ment role in supporting technical educa- or the expiration in July 1982 of the Orderly tion; Marketing Agreements covering imports of col- economic adjustment policies intended to or TVs from Korea and Taiwan. The objective encourage shifts of resources from declin- is not to be exhaustivebut selectiveto try to ing industries to thOse with better pros- differentiate the factors influencingcom- pects for future competitiveness, and to petitiveness in electronics that are primarily aid workers, communities, and regions under the control of managements of individ- that have suffered because of shifts in in- ual firms from those that are strongly affected ternational competitivenesse.g., in con- by the Federal Government, and to examine the sumer electronics. latter in the context of a high-technology in- These examples all involve complex issues, dustry that has been one of the mainstays of with effects that may differ among various U.S.competitiveness during the postwar parts of the electronics industry, and from firm period. CHAPTER 3 Electronics Technology Contents

l'ago Overviow 67 List of Figures Consumer Electronics...... 68 _Figure No. Semiconductors. 71 1. Simplified Diagram of TV ReceiVer Transistors 72 Componentry 69 Integrated Circuits 72 2. Comparative. Feature Sizes of System Design and the Microprocessor 73 Microelectronic Devices Such as Microprocessors and Nlemory. 74 Integrated Circuits 71 Learning Curves anti Yield ... 7fi 3. Controller for a Microwave Oven Based Integrated Circuit Design 77 on Single-Chip Microcomputer 75 Manufacturing Integrated Circuits 78 4. Increases in IC Integration Level 75 Fut ore Developments 5. Schematic Learning,,Curve for the Computers Production of an Integrated Circuit or Types of Computer-, 82 Other Manufactured Item 76 Computers as System 83 6. Ranges in Cost and Time for Design and Computer Son \,vare 85 Development of Integrated Circuits , 77 Growth of.Computing Power 87 7. Steps in Designing and Manufacturing an Applications of (:oinputers 90 Integrated Circuit 78 8, Step-and-Repeat Process of Summary and Conclusions 92 Photolithographic Pattern Formation on, Appendix 3A-Integrated Silicon Wafer 79 Circuit Technology 93 9. Typical Microcomputer Intended for Personal and Small-Business Applications 84 Appendix 3B-Computer Memory 99 '10. Typical Minicomputer Installation Appendix 3C-A Process Including a Pair of Terminals and a Control Example 102 Printer 84 11. Data Processing Installation Built Around General- Purpose Mainframe Computer 84 List of Tables 12. Simplified Block Diagram of a Microprocessor System 84 Tahlo Page 13. Elements of a General-Purpose Digital 1. Examples of Solid-State Technologies Computer System 85 Used in Information Processing 72 14. Relative Hardware and Software Costs 2. Cost Increases for Fine-Line Lithography 80 Faced by Users of Larger Computer 3. Characteristics of Drferent Categories of Systems 86 Digital. Computers 83 15. Minicomputer Price Trends 89 4. Increase in Computing Power 16. Parallel Decreases Illustrate How Costs of Over Time 88 Computers Depend on Costs of 5. Comparison of IBM 650 and Fairchild Semiconductors 89 F-8. Microcomputer 88. 3A-1. Silicon Wafer Showing Patterns 6. Characteristics of Generations of Formed by Photolithography 97 Computer Technology 90 3A-2. Integrated Circuit Chip Before and 7. Typical Applications of After Final Encapsulation 98 Computing Capability 91 3B-1. Performance of Computer Memory 3A-1. Common Termina'ngy and Alternatives 100 Classifications of Integrated Circuits 94. 3C-1. Portion of Process Control System for 3A -2. Principal Types of Memory Circuits 96 a Chemical Plant 103 313-1. Computer Memory 101 CHAPTER 3 Electronics Technology

Overview

This cllapter outlines the technology on of discrete voltage levelse.g., any voltage be- which the consumer electronics, semicon- tween 2 and +1 volts might represent a ductor, and computer industries depend, cov- binary "0," any value from +2 to +5 volts, a ering them with enough depth to provide back- binary "1 " ground for discussions later in the report concerning the role of technology as a force on Regardless of the simplicity or complexity of competitiveness. Except for occasional ex- the information content in a signal, either ana- a mph competitiveness itself is left to later log or digital technology can, in general, be chapters. employed. The choice turns on the practical The primary function of electronic compo- advantages and disadvantages for a given ap- nents and systems is td manipulate and trans- plication. A complex system may use analog mit information in the form of electrical sig- circuitry for some tasks, digital for others. In nalseither analog or digital. The transmission geophysical exploration, for instance, an ana- and utilization of electric power are integral log signalessentially a mechanical pressure to these processes, but constitute only second- pulse or sequence of pulsesis transmitted into ary functions of electronic equipment. Even in a geological formation. The reflected pulse the case of a 50,000-watt radio broadcasting sta- from the subsurface strata is sensed by trans- tion, the high power simply increases the area ducers analogous to microphones. These trans- coverage of the information in the signal. The ducers respond to the mechanical energy of the information manipulated and conveyed via an reflected pulse by generating a proportional electronic system can range from a simple analog electrical output. Analog-to-digital consequence of numberse.g., a zip code, or the verterstypically integrated circuits (ICs) balance in a checking account to the sounds then convert these signals to digital informa- conveyed by radio, images such as television tion that can be processed and analyzed by a pictures or weather maps, or the information digital computer. contained in radar or sonar signals. Both analog and digital technologies have a 'Changes in electrical voltage are the most place in the three sectors of the electronics in- common carrier of information in elecAronic dustry covered in this report. But while most systems. In an analog system, the signal takes consumer electronics equipment is still based the form of a voltage or some other electrical on analog technologyradio and TV receivers, parameter that varies continuously over a phonograph records, magnetic tape players-- range, while digital information is encoded in virtually all computers process information in the form of a string of binary "bits." Each bi- digital form. At the same time, computer pe- nary hit can take on one of a pair of discrete ripherals such as terminals and printers con- values, again usually voltages. The magnitude tain analog circuitry, while digitally based con- of these is unimportant, so long as they can be sumer products are becoming more common. distinguiShed from one nother.A bit can be Semiconductor devices come in both analog visualized as having values of "0" or "1," or (often termed "linear") and digital varieties. A " +" as opposed to " In a digital circuit, few ICs combine analog and digital circuitry the signal normally takes the form of a string on the same "chip." 68 International Competitiveness in Electronics

Consumer Electronics

The most common consumer electronic prod- ics markets is now in new generations of products are radios, TVs, and audio eqUipment such ucts, notably VCRs. Still, in many respects as "stereo" systems. Electronic toys and games, .e.g., the relatively standardized design ap electronic watches, pocket calctilators, and proaches and critical importance of production home computers are other familiar examples. costscolor TV continues to typify consumer These are all "systems" in the sense that they electronic technologies. contain more than a single electronic compo- Television signals 'ire broadcast via ampli- nent, but some are much more complex than tude modulation of a high-frequency carrier others. An electronic watch may, consist of lit- signal, much like AM radio. But the bandwidth tle beyond a single IC and -a displayitself a requirements for TV are far greaterabout 6 solid-state deviceplus a battery. Television MHz, versus 10 KHz for AM radio. Bandwidth, receivers contain several hundred componentS. which is expressed in terms of frequency-6 Video cassette recorders (VCRs) are complex MHz being equal to 6 x 10° cycles per second, inechrmically as well as electronically. 10 KHz to 10 x 103 cycles per second--is a Radio broadcasting provided the foundation measure of the rate at which information can for the development of the consumer elec- be conveyed, hence must increase with the

. tronics industry. Despite a real cost much amount of information in a signal. The pictorial higher than today, there were well over 10 image in a TV signal has a much higher infor- million radio receivers in use in the United mation content than sound, hence television's States by 1930, with annual sales exceeding $1 high bandwidth requirements. In principle, the billion.' analog information in either a radio or a TV signal could be conveyed in digital form with- Research and development (R&D) on televi- out changing the bandwidth requirements sion began in the 1920's, with limited broad-. greatly. casting prior to World War II. Large-scale commercialization had to await the end of the war, A home antenna receives the amplitude-mod- but by 1949 5 million TV sets were sold in the ulated TV signal at a microvolt level (1 micro- United States all black-and-white. Color tel- volt equals 10-8 volts). To produce a visual im- evisionfor which most of the early work was age, this signal is amplified to control an elec- performed by RCAfollowed the next year, but tron beam which scans the front of the picture color TV sales in the United States did not pass tubeor cathode-ray tube (CRT)forming a the 5 Million mark until 1967, and first ex- new image 30 times each second (the number ceeded black-and-..vhite sales in 1972.2 -..sf frames per second can vary abroad). The cir- With more than 11 million color sets sold in cuitry in a TV receiver is quite complicated (fig. 1982, and about half as many black-and-white 1) and now entirely solid state (the chassis in- sets, the TV receiver remains the largest sell- cludes both discrete transistors and ICs) except . ing consumer electronicsproduct, accounting for_the picture tube. The CRT is the most ex- for nearly half the dollar value of consumer pensive single component in the set, account- electronics sales in the United States (ch. 4). ing for about 40 percent of the cost. Produc- Monochrome TV sales have been rather static. ing picture tubes fs a highly specialized activi- for a number of years, with the market for col. ty; smaller firms oiler, buy CRTs from manu- or sets expriding only slowly; cable TV and facturers such as Zenith or RCA. direct satellite reception may spur future sales, Conventional picture tubes are not only but much of the-growth in consumer electron- bulky and expensive, but account for much of the power consumed by TV sets; in Japan and Fiirrinnnartiridaricrrrin evi mar altar+ rrit,i,c Ch. 3Electronics Technology 69

Figure 1.Simplified Diagram of TV Receiver Componentry

Antenna,

elalma Microvolt signal Speaker

Tun17er Audio 1 signal 0 Amplifier 1111. I 0 Amplifiers Video detector

Video signal

Horizontal Video and vertical amplifier(s) scanning circuits, color Power decoder supply

Picture tube

SOURCE: Adapted from K. (ed.). Radio Engineering Handbook, 5th ed. t Yew York: McGrawHill:1959).ch. 22.

only reduces consumer electrical bills, butcan vances have come from Japan as weli.3 Televi- lead to more reliable operationone of the ad- sion technology is now well diffused interna- vantages that TVs imported from Japan have tionally, and no one countryappears to enjoy enjoyed over the years (ch. 6). The drawbacks a technological advantage. Product innovations of conventional picture tubesprincipally continue to come from U.S. firms, but also bulkhave stimulated considerable R&D aimed from other parts of the world. Europeancon- at flat-screen television displays. Flat screens sumers, in particular, are often attracted by are not yet practical, but continued progress in solid-state technology will doubtless lead to rabid., p. 27; also "International Technological Competitive- , eventual success. ness: Television Receivers and Semiconductors," Charles River Associates Inc., Boston, Mass., draft report under NationalSci- Before 1960, most of the significant technical ence Foundation grant No. PRA 78-20301, July 1979,app. 2A. 70 International Competitiveness in Electronics new and different product featuressuch as While the commercialization of VCR technol- multiple image displays (several channels ogy by Japanese manufacturers is onesign that shown simultaneously on one screen). -These Japan may be taking over product leadership have become important to product develop- in consumer.electronics, video disks thus far ment strategies of firms in Western Europe. present a mixed, perhaps contradictory; pic- Continued progress in large-screen projection ture. The optical video disk systeindeveloped TVs has been stimulated by competition be- in Europe by Philips reached the consumer tween Japanese and American producers for market first. In the Philips system, a laser reads what could be a large new market. Japanese the digitally encoded signal on a spinning disk; firms have been leaders in reliability, and may microscopic depressions in the disk represent have tended to emphasize R&D directed at binary "Os" or "1s." While an elegant technical automationand at rationalization of the man- achievementand one with potential for high- ufacturing process in generalmore than density digital data storage of other types (e.g., European or American producers (ch. 6). In in conjunction with computer systerns)the particular, Japanese TV makers were leaders optical video disk sold in the United States by in adopting transistorized chassis designs in Magnavox has not been a commercial success. the late 1960's and in automating the insertion RCA's video disk, introduced early in 1981, of discrete components such as transistors, ICs, functions on analog.-principlesmore like a capacitors, and resistors into printed circuit phonograph record. Yet a third system, devel- boards. oped in Japan by JVC, may eventually reach the marketplace. As compared to VCRs, disk sys- Although TVs still account for much of the tems are cheaper but can only play back, not consumer electronics market, they are a ma- record. While the technology is evidently in ture product in the sense that most American hand, it is too early to tell how large the market homes already have one or more. Thus, the for home video disk players will bee.g., great proportion of saleS are now supplements whether it will rival that for VCRsor which or replacements. The marketfor VCRs, in con- systems will survive in the marketplace. trast, is expanding rapidly (ch. 4). U.S:' sales of VCRs during 1980 Nere less than a million A number of trends in consumer electronics unitsall importe i; sales nearly doubled in e.g., recent introductions DI "component" 1981. TVs analogous to component stereo systems, along with games and low-end home comput- Video recording on magnetic tape' was pio- ers that use a conventionaltelevision as the neered in the United States by Ampex.4 Al- displaypoint toward the eventual develop- though Ampex and RCA continue to manufac- merit of more-or-less integrated home enter- lure video tape recorders for broadcast applica- tainment systems. Such systems might in, tions. consumer VCRs were developed largely corporate TV and audio recepti .-'n and repro- by Japanese firms (ch. 5)which now build duction, including various kinds of informa- about 95 percent of the world's. V.CRs. In tion services, along with applications ,Df com- Europe, Philips has a few percent of the mar- puting capabilitynot only games, but record- ket. hut all the VCRs sold in the United States keeping, home security systems, control of come from Japan. household appliances, and regulation of heating, ventilating, and air-conditioning systems. Such developments do not depend heavily on ""Interactions of Science and Technology in the Innovative, technological advances except as low produc- Process: Some Case Studies.-. final report. Battelle Columbus Laboratories. National Science Foundation Contract No. NSF-C tion cost is necessary for mass market accept- 667. Mar. ch. 12. ance. Ch. 3Electronics Technology 71

Semiconductors

Strictly speaking, the term "semiconductor"' conventional semiconductor physics. Such de. refers only to the mat;/lids from which semi- vices might function on magneticor optical conductor devices are made. Such materials principles, rather than being strictly "electron- have electrical conductivities intermediate be- ic," although the materials involvedare some- tween good conductors like copper and insula- times semiconductors.* Boundaries between tors such pis glass. Silicon is the most common electronic, magnetic, and optical technologies semiconductor materialvirtually all ICs, and tend to blur as device technologiesmove to- most discrete transistors, are based on silicOn. ward microstructural and submicrostructural In this report, the term "semiconductor" will size ranges. (Microsauctural sizes are large be low,* applied to the products of the "semi- compared to interatomic distances but small conductor industry" as wellas to the materials compared to objects that can be easilyseen and that are the starting points for these products. handled, like an IC chip itself; the feature sizes The broader designations "microelectronics" of microelectronic devicesare currently in the or "Microe lectroi tic devices" include Semicon- range of 1 to 10 micrometers, or less than a ductor productswhich have replacedvacuum tenth the diameter of a human hairfig. 2.) Be- tubes in nearly all applicationsas wellas cause solid-state devices based on magnetic or other types of solid-state devices thatprocess, optical principles are often usedas co m- manipulate, or display information. ponents in systems that are broadly electronic The most familiar example ofa vacuum tube in nature, no fine distinctions will be made. application that-solidtate technology has not yet been able to match is the CRTnot only the common TV picture tube, but the display Optical data transmission can give bandwidths much higher than electronic signals; this, along with the low raw material screens of computer terminals. In other cases, cost. is one of the advantages of optical fibers. Systems based' solid-state devices are not only much smaller on laser light sources: with optical fibers for signal transmission than vacuum tubes, but cheaper,more rugged; and thin-film integrated optical devices for signal processing. and longer lasting. They also offer higherop- could tepla.:e many types of electronic circuits andsystems. erating speeds; indeed, on almost any measure of performance, microelectronic devices offer Figure 2.Comparative Feature Sizes of order-of-magnitude improvements over the Microelectronic Devices Such as Integrated Circuits components they have, replaced. Modern digi- Representative !al computers would be quite impossible with- feature sizes Dimension in out semiconductors._ Although a computer (design rules) Examples micrometers otherwise like current models could, in prin- Human hair 100

ciple. function with tubes instead of ICs, such `led blood cell 7 a machine would fill a building and probably 1978 5 micrometers . 19802 micrometers . not execute a single program without one or Yeas: cell 1110L'i!tubes failing. Solid-state circuits have 1985' 0.5micrometer -Smallestbacteria 0.2 made practical many electronic systems that 1990'. 0.1 micrometer would earlier have been too big, too costly,or otherwise in fact unthinkable. Poliovirus 0.01 Although virtually all commercial microelec- 0.001 tronic products are now made from semicon- Mom 0.0002 ducting materials. considerable R&D has been 1 micrometer = 106 meter devoted to classes of solid-state technologies Projected. with tmlential for transtrnittina and nrnoccirm SOURCE: AJaated !'rom G B. Larrabee rharartnriyntinn <".1" 72 International Compotitiveness in Electronics

Among these solid-state devicestable t regulate the flow of electricity in a circuit. As are: a result, they can amplifyelectrical signalsa function that earlier could only be performed transistors,ICs,light-emittingdiodes by vacuum tubes. (LEDs), all of which are semiconductors; bubble memories; which depend on the Transistors come in many varieties to serve magnetic rather than the electronic prop- different functions, just as for the vacuum erties of materials; tubes they superceded. To make a transistor, liquid crystals (as in alphanumeric dis- a semiconducting material such asgermanium plays for watches or calculators), chemi- or silicon is "doped" withsmall amounts of cals that change colors when their temper- elementsarsenic, boron, phosphorusthat lo- ature changes; cally affect its conductivity. The transistor, to integrated optics, in which information is the naked eye, is then just a small piece of, say, transmitted and processed in the form of silicon with two or three wires attached. In light. Integrated optics and prospective fact, however, the purity, chemical composi- future technologies such as organic semi- tion, and perhaps crystal structure have been conductors are not commercially impor- carefully tailored on a microscopic level. tant at present, but could have impacts on future success in internation& r:ompeti- Integrated Circuits' Hon. An IC is made by fabricating several circuit elementstransistors, capacitors, and such- Transistors on a single substrate. Integrated circuits were independently developed in the 1950's at Texas While most of the fundamentals of semicon- Instruments and Fairchild Camera and Instru- ductor physics were known prior to World War ment, still two of the leading selniconductor II, the transistor itself was developed at Bell manufacturers in the United States (Fairchild Laboratories after the war, and first demon- is now French-owned). The two companies ap- strated in late 1947.5 proached the problem quite differently during In contrast to passive electronic devices such 1958-59, but their developments shared the as resistors, capacitors, andinductorswhich common characteristic of an ICtwo or more can only respond to electricalsignalsactive distinct transistors fabricated on a single circuit elements like transistors control and substrate.° Thus they were monolithic circuits. °M.F. Wolff, "TheGenesis of the Integrated Circuit,"IEEE W. Shockle, ihe Path to the Conception of the-Junction Spectrim. August 1976, p.45: Braun and MacDonald,op. cit., Transistor." !EEE Transactions on Electron Devices, vol. ED-23, ch. 8. The depths of the transistors fabricated on a chip are small 1976. p. 597; E. Braun and S. MacDmald. Revolution in M:nia- enough that ICs can be considered two-dimensional. Often the ture: TI .0 History and Impact of Semiconductor Devices (Cam- silicon substratea few millimeters on a sideandless than a bridge, Mass.: Cambridge !University Press, 1978). ch. 4. millimeter thickis called a chip, as is the resulting circuit.

Table 1.Examples of Solid-State Technolcgies Used inInformation Processing

Technology and examples Description Current status Semiconductor,,electronics: Depends on electronic properties of semicondectin,s Production Transistors materials such as silicon, germanium, gallium arsenide. Integrated circuits Bubble memories in limited Magnetic devices: Depends on magnetic rather than electronic properties of Bubble memories materials. production. LEDs widely used for displays; Solid-state optics: Depends on electro-optical properties of r-aterials, some of which are semiconductors. integrated optics (sometimes called experimental. optoelectronics) . Light-emitting diodes LEDs are lighted whan a current passes. Ch. 3Electronics Technology. 73

Other types of ICs can also be builte.g., MOS. Finns that were slow to master MOS hybrid or thin-film circuitsbut monolithic tended to fare poorly in sales growth and prof- devices comprise the bulk of production. itability over the past decade. For the foreseeable future, competition in ICs will continue At present, the market for ICs is more than to center around MOS devices. four times the size of that for discrete semiconductors. Because of this, and because very large-scale circuits pace the industry and are System Design and the Microprocessor a major focus of international rivalry, this report gives much more attention to ICs than In designing a digital system, the engineer to other microelectronic devices. has several options: 1. to assemble a number of standard logic ch.:- Appendix 3A discusses IC technology in cuits like those of the transistor transistor some detail. The significance of the technology logic family described in appendix 3A; itself for international competitiveness resides 2. in cases where large production volumes largely in the commercial advantages that can are anticipated or performance;require- accrue from innovative and/or widely accepted ments are specialized and demanding, to chip designs (ch. 5), as well as from mastery call for one or more custom ICs; of processing technology. Being first on the 3. to use a standard microprocessor or micro- market with a new design gives a firm the op- computer with software written for the portunity to build market share before competi- particular application. tors can offer similar products. The advantage of a particularly well-accepted design is that Assemblies of standard logic circuitstypi- it may become a de facto industry standard. cally small- or medium-scale ICsmay be eco- A manufacturer whose design becomes such nomical in limited production volumes, despite a standard not only has the assurance of a rela- relatively high design and development costs. tively large end stable market, but also the pros- In such cases, the system is implemented in pect of a broad range of licens:ng and/or sec- hardwarei.e., its functioning can only be ond-sourcing agreements. The firm may also altered by changing the circuit components get a headstart in the competition to design the and/or their interconnections. next generation.replacement. Processing capa- Specially designed custom circuitsanalog bility is just as important to competitive suc- as well as digital, bipolar as well as MOShave cess as design, because advanced circuit de- a place in high-volume applications ranging signs are often limited by what can be built at from consumer products like TVs and electron- acceptable yields. (The yield is the fraction of ic watches to telecommunications systems. "good" circuits coming off the production line.) They are also employed where performance re- In semiconductorkadvances in circuit design quirements such as operating speed cannot be and in processing capability are iniArdepend- met in other wayse.g., in some military and ent to a greater extent than in almost any other aerospace applications, or in mainframe com- industry. puters. Custom circuit design is expensive: hundreds of thousands of dollars, sometimes The first of the two major types of ICs to be running into millions. Here the designer is also developed, bipolar, has been replaced for many working in hardware, but new custom hard- applications by MOS (metal oxide semiconductor, app. 3A). Over the course of the 1970's, ware rather than standard ICs. MOS technologywhich is denser, cheaper, In contrast, when a system based on a micro- and consumes less power, but which does not processor or microcomputer is designed, the ofier speeds as high as bipolar--became domi- logic is implemented largely through software nantfnr larop-cralp intporatinn fT.ST1Vpry Pa rnmnntpr nrnoram ctnrpri in mpmnry 74 International Competitiveness in Electronics

microcomputers can function as central proc- live for systems that could be built with as few essing units for general-purpose computer sys- as two or three dozen standard logic circuits. 4emssuch as the small machines sold by Ap- Figure 3 illustrates a typical application of a ple or Radio Shack---they were originally con microcomputer, control of a microwave oven, ceived as replacements for custom ICs to cir- where the chip contains memory for program cumvent the high costs of designing, develop- storage along with a processor. in such an ap- ing, and producing custom parts. The first plication, production volumes might be high commercial microprocessor was introduced by enough to justify a custom LSI chip design Intel Corp. in late 1971 to implement the arith- tens of thousands of identical parts, at a mini- metic functions in an inexpensive calculator. mum, are normally called for. But the micro- Faced with the request of their Japanese cus- processor/microcomputer alternative has a big tomer for igroup of custom chips to be used advantage in flexibility; design changes are in a line of calculators, Intel instead proposed simple, different models simply need different a simple 4-bit microprocessor chip.' Rather programs. And, in the microwave oven exam- than hard-wiring the opera ions required for ple, there is no need for high performance. Iho different calculator modelsaddition, mul- Before the advent of the Microprocessor, sys- tiplication, printing, and so onsoftware pro- tern designers had only two choices: assemblies grams permanently stored in memory imple- of standard parts or custom ICs: The micro- mented these functions. Money was saved in processor/microcomputer introduced a third design and production compared to the custom option, one that proved highly attractive. As IC alternative. a result, several hundred.different models of Subsequent experience .ias shown that mi- microprocessors and single-chip microcom- croprocessors may prove the low cost alterna- puters are now marketed (many differ only in details), and custom microprocessors are some- 'R. N. Novice and M. E. Hoff, Jr.. "A History of Microprocessor times designed for special applications. Development at Intel Corporation."IEEE MICRO.February 1981, p. 8. Parallel developments took place at Texas Instru- ments. Microprocessors and Memory Microprocessors cannot be used by themselves; they must be supported by other chips, at a minimum for program storage. Memory circuits aredescribed in some detail in appendix 34, particularly table 3A-2. Memories, in fact, comprise the largest single market cafe- gory'for- ICs; the majority go into general-purpose computer systems, but large numbers are also used in dediCated applications of micro-

I processors and microcomputers (i.e.. applica- r", tions where the computing function is invisible to the ,user of the system).

,:-.!it ,:r.-:.:. tvirv-,.':-:-.i-c'c-v-,, Technological progress is easier to measure i..t-,,;,,,,t.,-,-,-ii. ..- . s_, , . for memory circuitse.g., RAM chips (random t-vv-?,..f77,,,,r . ___ ,-..,,-' t.----.:77,..r:, t-, access memory)than any other type of IC. ,,,,,.7754,,,,-,43., ,, 1, .....:.... i-,. ;,..,..:,..-;: irf, k..1a ; 11121._ Densities have increased steadily over time /figure 4while the cost of a chip has remained /'roughly the same. As a result, the_cost per bit ..1411iigkaii. rigiej II' of information stored goes down. This opens Ch. 3Electronics Technology 75

Figure 3.Controller for a Microwave Oven Based on SingleChipMicrocomputer

Display Touchpanell

Status hod Display indicators drivers

8-bit A-D con- verter Line power PowerClock supply

Alarm driver

Oven heating] Magnetron Alarm triac driver element triac

Door- interlock

SOURCE, IR Walker, "Analog/Digital LSI," 1979 Elecfro Professional Program, New York, Apr. 24.26, 1979.

Figure 4.increases in IC integrain:Level rely on memory-7-notably microprocessor systems. LikeWise, as semiconductor memory be- for 10 comes cheaper it will continue to substitute 32-bit microcomputer alternative storage media such as magnetic tos 256K RAM- disks in general-purpose computer systems. It 64K was widely noted in the early 1970's thatthe 16K RAM. RAM 16-bit cost per bit of memory had fallenbelow the 4K microcomputmicrocomputer c 104 RAM cost of a jelly bean. According to some esti- 0 1 chip .8-bit mates, a jelly bean (at 10 may buy as many as 10 caiculator microprocessor 1,000 -(1K) bits of memory by 1990. .5) -.. Meinory and microprocessors are the most Medium-scale integration visible products in domestic and international competition. While it would be wrong to con- Small-scale integration sider these,the only important categories, they do constitute half the 0total market for ICs. 0.1 Moreover, significant advances:in both device '1960 1965 1970 1975 *!980 1985 Year technologies and proCess technologies have often found their way into production via cir- SOUttcr. Oueyssac, "Projecting VLSI's Impact on Microprocessors," IEEE :lpecltum. May 1979, p. 38. .cuits of these types.- 83. 76' International Competitiveness in Electronics

Learning Curves and Yields Figure 5. Schematic Learning Curve for the Production of an Integrated Circuit or Other Making ICs is demanding, more so at higher Manufactured Item Itu, els of imegration. Forty or more processing steps may be required for a VLSI chip, a figure that will continue to grow. Designing VLSI circuits is also complex, and becoming steadily more time-consuming and expensive, while product design and process design go hand-in- hand. For consumer electronic products like TVs, decisions on product features often hinge on the costs of production; Mr a new IC, the first question is: Can it be made at all? Once a semiconductor firm has designed an IC and carried it through the pilot production stage they can normally assume that produc- Cumulative number produced tion costs, even if initially high because of low SOURCE: Office of ".::_nnologt Assessment. yields, will decrease over time. Figure 5 is a schematic learning curve, sometimes called an experience curve, showing cost declines with the many steps in the production process, cumulative production volume. Learning smoothing of work flows, and perhaps changes curves typical of IC manufacture show that in the design of the part itself. Control of the when cumulative production doubles, costs process is particularly important, and often de- decrease by about 28 p-rcent.9 pends on subtle variations in parameter,influencing phenomena such as diffusion, etching, Learning curves as in figure 5 apply to man- or polymerization (of photoresistsIC fabrica- ufactured products of many kinds, but their im- tion steps are described .in more detail below). pacts on pricing decisions have been par- In inany cases, the physics and chemisiry of ticularly noticeable among sernimnductor such phenomena are only poorly understood, --firms; they are a major factor in forward-pric and cannot be modeled theoretically; process ingsettinp prices below the initial cost control models tend to rely on empiricism, production to gain market share. Be,-:ause firms hence on accumulated experience. Denser and feel confident that costs will decrease as pro- more complex ICs call for a better grasp of duction experience accumulates, forward-pric- processing fundamentals. ing has been .a common competitive tactic. In general, learning improvements show up These cost declineswhich can be consid- as increased yield, the percentage of chips that ered equivalent to increases in productivity pass final test and function satisfactorily. When stem from much more than simple learning or a new IC goes into production, the yield is gen- experience by the labor force; other causes in- erally low perhaps only a fraction of a per- clude better equipment performance and utili- centbut rises as experience accumulates and zation, greater understanding and control of processing can be better controlled. As a rule- of-thumb, products are seldom marketed, ex- 8"Boom Times-Again for Semiconductors," Business Week. cept for sampling purposes, until yields have Apr. 20, 1974. p. 65: A Report on the U.S. Semiconductor Indus- try (Washington, D.C.: Department of Commerce, September. risen to about 10 percentwhich may take as .1979), pp., 48-50. The 28 percent figure is an average from which'' much as a year of production-line experiences the cost experience for a given IC can deviate substantially. Pro- For mature products, yields can rise to well duction volumes typically rise rather slowly at first, because it takes time for customers to design the new part into their over 50 percent. Increased yields are a power- systems. In comparison to other types of manufactured prod- Ids, learning curves for semiconductors are not particularly °R. Bernhard, "Rethinking lila 256.kb RAM," IEEE Spectrum,. steep, but continue to fall over very long production rims. May 1982, p. 45. Ch. 3Electronics Technology 77

ful force in driving down the costs and prices Figure 6.--Ranges In Cost and Time for Design and of ICs; in effect, doubling the yield hay--is the Development of Integrated Circuits production cost. 35 16-Oit microprocessors Because processing capability is so critical to commercial success, the specialized equip- 30 8-bit ment used in fabricating ICs is one of the keys microprocessors to competitive ability. Much of this equipment is designed and built by independent suppliers 25 4- to 3-bit many of them American firms =- selling to microcomputers customers throughout the world (ch. 4). The tech::Jlogical capability and competitiveness 20 of the portion of the U.S. electronics industry that designs and manufactures equipment has 15 Custom circuits $20million for been just as important to the international posi- Intel 8086 (1979) tion of the United States in semiconductorsas 10 the efforts of semiconductor firms themselves. Because of the interrelations of device technologies and process capability, and the depend- 5 ence of costs and yields on process control, a number of the more important steps in producing ICs--beginning with circuit designare de- 5100.000 Si million SIO million $100 million scribed in more detail below. Development cost (dollars) SOURCE: P. M. Russo, "VLSI Impact on Microprocessor Evolution. Usage, and System Design,"-IEEETransactions on Electron Devices, vol E1127, Integrated Circuit 1114sign 1980. p. 1339. The task cf the circuit designer is to define an arrangement of circuit elementS7---transistors, capacitors, logic gates, interconnections man-yearrs for desirna and development using that will satisfy the functional requirements of the conventional methods of the past decade.10 the IC. The more complex the circuit and the Computer-aided designi.e., the use of special- higher the level of integration (a 64K RAM con- ized computer programs by the design engi- tains more than 100,000 circuit elemP-Its) the neersis the principal hope for cost reduction, and is increasingly necessary just to handle the more difficult the designer's job, and . :,;gher the cost of design and development,.lure 6 logical complexity as the number of devices per illustrates ranges of development time andcost chip goes up. Designing-a microprocessor-with--- including hardware, software, and peripher- 100,000 transistors would be irr:,-actical with- al chipsfor several types of ICs. out computer aids. R&D aimed a: more regular even modularchip designs is also under. - Asa rule-of-thumb, circuit design costs his- . y, again intended to reduce the time, hence torically have been rather stable at about $100 the cost, of IC design. Modular approachesare per gate. Thus, a microprocessor with 10,000 particularly attractive for custom logic circuits. gates will have a hardware design cost of perhaps $1 million, and may represent 10 man- As pointed out above, the microprocessor years of effort. Sof:ware costs add to this. It itself originated as a way to reduce the costs may well be possibe to make chips with 1 mil- of custom circuit design; in essence, choosing' lion gates within a few years, but the costs of a microprocessor means replacing hardware designing them will be prohibitive unless the design by software design, .and in many cases design costs on a per-gate basis can be reduced; lowers costs. But as logic complexity goes one estimate has been that an IC with a densi- I°C. L. Hogan, cited by F. Ogden, "Audience Gives Mixed F,e- ty of a million devices would require about 200 ception," Electronics Weekly, Mar. 28. 1979, p. 5.

85 78 In ternat o nal Competit;veness in Electronics

_ . the software costs for grog: . ning the micro- enough to be familiar with the performance processor escalate rapidly. in part, this simply characteristics of off-the-shelf devices. At the reflects the more sophisticated processors same time, designerstypically electrical e, g., it 18-bit rather than a 4- or 8-bit device engineersmust be at home with the logical needed for demanding applications. Although concepts and ciftl.vare orientation of the com- software production can also be automated, the puter scientist; the need for software skills will 0.vo basic paths toward implementing logic arm., ICs come more and more to resemble hardware via custom chip design, or software integrated syst6ms. This melding of hardware tia a standard microprocessor with a program (including process technology) and software embodying the logicwill continue to com- skills, and the rapidity of technical change, pete. Design cost, flexibility, and performance make unusual demands on the people who fill are all factors. But if computerized design aids such jobsone reason that the electronics infor hardware and software advance sufficiently dustry has been experiencing shortages of qual- far and in tandem it may eventually make lit- ified engineers (ch. 8). Although neither the cir- Ile difference. perhaps even to the designer, cuit designer nor the process specialist can be whether -the-logic is embodied in hardware or fully conversant with all aspects of IC design ftwa re: and manufacture (fig. 7), some knowledge of each is needed. Although much of the engineer's work revolves around the logic that the circuit will implement, IC design also calls for intimate Manufacturing Integrated Circuits12 knowledge of processing and fabrication."fig- The design process culminates in a pattern tire 7. Designs that can be implemented in or layouta large drawing, several hundred n-MOS might be impossible in c-MOS (see app. times the size of the circuit itselfthat must 3A, table 3A-1, for an explanation of the types be translated into the "tooling" for producing of MOS devices). One company might have the chip. In simple terrils, the procedure for n -MOS process capabilities beyond the reach making an IC resembles a series of photograph- of another. The desirn team must consider fac- ic processeslithographic patterns are created tors such as the spacing between transistors in layers on a silicon wafer. Each layer is made and the widths of the lines that interconnect by exposing a polymeric chemical called a pho- them. In contrast, when designing a system to toresist to light or other, radiation, the light. be built from discrete components it was often 11P..W. I. Verhofstadt, "Evaluation of Technology Options for "See, in general, F. W. Vollmer, "Manufacturing Process Tech- LSI Processing Elements," Procee&r;gs of the IEEE, vol. 64. nology for MOS VLSI," VLSI Electronics: Microstructure Sci- 1976. p. 842: C. Mead anti L. Conway. Introduction to VLSI Sys- ence, iol. 1, N. G. Eins.pruch led.) (New York: Academic Press, tems (ReadinA, Mass.: AddiSon-Wesley, 1980). 1981), p. 1.

Figure 7.Steps in Designing and Manufacturing an IntegratedCircuit

Processing Science off technology ManufaCture Test KII Design Design technology (product and proces.$)

Product definition Design t verification

SOIJIICE: G. Moore, "VLSI: Soma, Fundamental Challenge!," (ZEEspect-um,April 1979, p.30: 80 Ch. 3Electronics Technology 79

passing through a grid-like mask. as shown technology of lithographic processing that schematically in figure 8. More than a dozen makes this possibie. such- masking steps may be needed tc build up a VLSI part. Many other processes besides I lographic line widths in production ICs lithography are involved in IC fabrication, with have been reduced an order of magnitude over more detail given in appendix 3A, but lithog- the past decade, from about 20 micrometers in raphy is critical for future increases in circuit the early 1970's to 2 to 4 micrometers current- density. While advances at many stages in the ly. (A micrometer is about 40 millionths of an manufacturing process take place in an inter- inch.) Thinner lines give higher operating dependent vaye.g., laser annealing is replac- speeds as well as denser packing. The 16K ing furnace annealing because it does a better RAMs designed in the early to mid-1570's were job of restoring the crystal structure of the based on 5 micrometer "design rules, -64K silo ::on which is disturbed by ion implantation RAMs on 3 micrometer rules (design rules, lithography is the major factor in determin- which are directly related to lithographic line ing howEmailindividual devices and intercon- widths, comprise- the full set of geometric con- nections can on a production as opposed straints that designers follow). The next-gener- to laboratory basis. Already, transistors can be ation 256K RAMs are based on 1.5 to 2 m. packed much more closely in an IC than neu- crometer design rules.13 Continued progress in rons are packed in the human brain; it is the reducing line widthsmore generally, feature sizeis thus a major driving force in moving further into VLSI. Forihis reason, a principal Figure 8.Step-andRepeat Process R&D target of the Defense Department's Very of Photolithographic Pattern Formation High-Speed Integrated Circuit (VHSIC) pro- on Silicon Wafer gram has been lithographic technologyan ob- Filament' Light source jective paralleling commercial R&D efforts, one re nson the program is likely to have a positive effect on nonmilitary portions of the semicon- Radiation ductor industry. The VHSIC program goals include two stages of lithographic improvements: the first stage calling for line widths of 1.25 micrometers; the second, lines of 1 micrometer Lens and below. While feature sizes of 1/2-to 1 micrometer are Mask well above the range for which the physics of electron devices will begin to cot:strain performance, such feature sizes co demand significant developments in lithographic capability, particularly for mass production," In the past,

""Rethinking tie 256-kb RAM," *n. cit. On design rules, see Mead and Conway, cp. cit., p. 47.

"1. E. Sutherland, C. A. Mead, at-7 E. Everhart, Basic Limita- Lens tions in Microcircuit Fabrication hnology, Defense Advanced Photoresist Research Projects Agency Report R-1956-ARPA. November 1976; R. W. Keyes, "The Evolution of Digital Electronics Towards VLSI," IEEE Transactions on Electron Devices, vol. ED-26, 1979, p. 271. The ultimate limit to reductions in the sizes of electron S'iicoil wafer: devices will perhaps be thermal noise, althougi: a variety of practical concerns may intrude first. Feature sizes are likely to de- Stage traverse crease to the range of 0.01 to 0.1 micrometer before fundamental SOURCE: Office of Technology Assessment. physical limitations are encountered. , 80 Internationa; Competitiveness ,n Electronics

visible lightgenerally ultraviolethas been usedtoexpose ohotoresists (fig. 8). But even deer.) ultraviolet, which has a wavelength of hoot'2 micrometer, cannot produce line widths much below a micrometer because optical considerations limit the lines to about twice the tvavelength of the radiation. ro achieve 1Ancrometer lines with visible light requires very sophisticated lithographic equipment--positioning and layer-to-layer reg- istration of the sequential masking steps must held to a small fraction of a micrometer. A single machine of the direct-step-on-wahr type diagramed in ligure 11 now costs about half a million dollarstable 2. As the table shows, the earlier generations of equipmen;placed by direct-step-on-wafer machines we.e much less expensive. Filter patterns will be still more costlywhether the technology of choice is electron-beam lithography (table 2), X-rays, or ion beams (see app. 3A). The rapidly rising costs of IC processing equipmentwhether for lithography, for ion implantation, or for testing along with higher costs for design and development, are the mo,4 important causes of the rapidly increasing capital intensity in the semiconductor industry (ch. 7). Entry costs are now roughly $50 million, versus $5 million to $10 million in the early 1970's. Photo credit Ga.! Corp Some have already moved to electron- beam lithography for ci itical circuit laYers. Electron-beam lithography system Electron-beam lithography is also a routine tool for making masks. X-rays and electrons have wavelengths much less than light, and so offer greater resolutionat the expense of high first technology, R&D on high-resolution lithograph- cost for the equipment, and low p:fedv.',:tion it-, techniques (see app. 3A) has been a principal drivinl.: iC target er g,-,vern merit-funded programs in other rates.15 Because of its importance countriese.g., Japan's VLSI projectas well ,,(;. lt. br,.,r.lio, Resolution Lithograph,..- as the VUSIC program funded by the U.S.mil- Tocimningy Aficroolectronic Fabrication. G. R. 1-.et.,-r. (New Vtrk: Acarlun ;r:Precs, 111110), p.I. itary.

Table 2.Cost Increases for Fine-Line Lithography. Approximate capital requirements Line width Throughput Approximate costfor production capacity of 1,000 Lithographic,system (mi(:.nmeters) (wafers hotr.) per system water starts per week Light Contact printing 10 60 $15,000 $30,000 Projection 2-5 60 $240,000 $400,000 Direct -step on-wafer 1-2 30 $480,000 $1.6 million Electron;beam 0.5-1.0 6 '1)25 million ,OUNCE Adapter' tram A J Stein, J Marlay, and R. Ma::r), -Toe Impact of VLSI on the Autcmobile of far,: r).», Elef.:tronic M,crost-t;cturo vo!: 2. N G Eatsprtica fed ) (New York' Academic Press, 1981), p 295.

4. Ch.3Electronics Technology 81

More cost conies with the "clean rooms.. Iso illustrate the importance of speed, as well needed for VLSI processing. Even micrometer- as power consumption, consider the technol- size dust particles can ruin the lithographic pat- ogy embodied in a current-generation main- terns: cleanliness is vital to high yields. In a frame computer. The central processor for one clean room. the air is filtered and people must such computerthe Amdahl 470-Vemplbys . wear special clothing. As circuits become dens- 1,680 ICs, with a total of about 100,000 logic er. and feature sizes smaller, not only is cleanli- gates. As is typical in large .computers, the . ness moreimportant, but the whole range of chips use silicon bipolar technology to give processing equipment used in making ICs be- high computing speeds. Replacing these bipo- comes more sophisticated and expensive, add- lar chips with .2.-Jlium arsenide may offer the ing to the capital requirements for semiconduc- potential for increasing corn.:.:atat:ional speeds tor manufacturing, a matter discussed in chap- by a factor of 10 to 100, and ra,.iocing the power ter 7. consumed by the processor from about 3,000 watts to perhaps 30 wattsless than most light Future Developments bulbs.I6 Comparable improvements other applications of ICs carry implications for coMpet- -;einiconducto,. devices need not be based on itive trends in many industries if s.;,-,iae com- silicon. One alternative is gallium arsenide, a panies or some countries manage a headstart material that offers considerable potential for in reducing such technologies to practice. improvements in packing density and speed (me to two orders of magnitude compared to siliconhut is still largely a laboratory technology. Whether gallium arsenide circuits will become commercially important depends on "R. C. Eden. H. M. Welch. R. Zur., and S. I. Lang, -The Pros- rates (*improvement compared to silicon, and pects for Ultrahigh-Speed VLSI GaAs Digital Logic," IEEE also on developments in other prospective -t.ich- Transactions thl Electron Devices, vol. ED26. 1979. p. 299. About 1 percent of 'he electricity consumed in the United States now n:..;logies.e.g., Josephson junctions. Josephson goes to computers, mosily for .cooling"CBEMA Predic- devi,:esalso experimental, and much further tion: Say Energy Crunch Could Cut EDF Growth Pate 50%," from demonstrated practicality than gallium Electronic News, Mar. 17, 1980, p. 32. On Josephia... junctions. see JMatisoo, "Overview of Josephson Technology Logic and arsenide ICspromise still better speed Memory,"HIMJournal of Research and Development, vol. 29. 1980, p. 113.

CerriolreeS

Computer tecin. ;logy has many roots, cgy progressed, the lead swung decisively clticli.gmilitar, needs during the Seconc American firms, much as happened overover Wirlc VVar for fire control tables and other roughly the same period of time for semicon- complex a ndior repetit:,fe computations. The ductors. Unitcd States had no great advantage over other nations during the early development of The Bureau of the C".7.' =us was an early non-. computers; significant innovations also origi- military custorner American computers, nated in several European countries, particu- census ds,procssing requirements remain- larly Great Britain.17 But as computing technol- ing a typical example of computer ru:ilications. When a Univac I NAP,s delivered to the Bureau r*Gaps in Techou'oey: Electronic Computers (Paris: Organiza- of the Census in It some observers pre- tion for i:ooperation and Development. 1909), p. 61. dicted that the market for digital computers Ft. a concise :ilimmory of develetnnentsduring thefirst three izeneratioi. of com;.citing technology, primarily in United might eventually total a dozen; a few years States, see S Rosen, "Electnnnic Computers: A 11 iitorical Sur- later, when :;ales to private industry began, the vey." Computing Surveys, vol. 1, 1969. p.1. estimatez were tle.t the potential market in the 82 International Competitiveness in Electronics

United States consisted of perhaps 50 corpora- of new circuits containing greater numbers of tions." devices, now VLSI, might not appear. Such Needless to say, as computer technology ad- fears seem to have vanished from the computer vanced manyflu':> firms became customers. business, though not the perennial questions and the ranks at computer manufacturers of which firms will get the largest share of the swelled. Among the entrants were a numbee new markets. of companies that had become established in the office. equipment market. International Types of Computers Bueiness Machines Corp., Burroughs, and Na- As pointed out in the earlier section on "Sys- tional Cash Register (now NCR) joined firms tem Design and the Microprocessor," the es like Univac that had been set up specifically sential elements of a small digital computer (ex- to manufacture digital computers. By the end clusive of power supply and input/output de- of the1960's.. computer applications had vices) can be placed on one IC to create a sin- spread well beyond numerical computations gle-chip microcomputer. A microprocessor is and data processing. The great part of com- more limited in function, but when combined puting power is still devoted to data processing with the necessary memory and peripheral for business and governmentaccounting, chips cn a printed circuit board becomes a sales, production, inventories, recordkeeping single-board microcomputer. From such prod- of all kindsand to scientific and engineering ectse---selling for around MO without cabinets calculations. in addition, many individual com- and other auxiliariesdigital computers range puters, mostly microprocessors and microcom- upwards in size, speed, and cost to "supercom- puiers, now perform "invisible" functions in puters." Intended for complex scientiLe and applications ranging from alto raft flight con- technicalcalculationse.g., modeling the trol systems to the microwave oven example Earth's atmosphere, designing airfoils shown in figure3. nuclear weaponssupercomputers are mhde The spread of computing power has some- by only a few manufacturers, and cost in the times been technology-driven, sometimes driv- vicinity of $10 million each. en by user demands. Technology-driven devel- In between board-level microcomputers and opments arose when more computing capabil- supercomputers come a number of broad cate- ity was available than people knew how to use gories of machines: personal and small busi- productivelyi.e.. before the eiiplications were ness computers like those made by Apple; min- wellefined. Under these circumstances, the icomputers of various types; general-purpose availability of more powerful machines or mainframes. The latter, typified by many of greater performance per dollar tends to gener- IBM's larger models, can handle many differ- ate new applications, or. more broadly, serve ent tasks at once. Table 3 outlines some of the needs earlier unmet. As in many instances of conventional distinctions among eeese cate- technological change, what the technology gories. The differences are not always clea:- could do. at any given time and for a given cost, cut and will blur even moE2 as microcomputers evolved in conjunction with applications, with become more powerful, competing power still one or the other ter-poi arily in the lead. Much cheaper, and computers of all types more ver- tlie same has been true for ICs. In the period' satile. when demand from military and 'space programs in the United States was high, the The central processing unit (CPU) for -im- market deeve the technology; but leaders in the jcomputerin fact, just a -ticroproce: "semiconductor industry heve periodically wor- is shown schematieally in fif:et.e 12. A micro- ried tbh.t applications for the full capabilities prcressor functions like the CPU-a: any com- puterit brings information (in the form of bi- "I.. M. liransco111, .'f.',1ect conics and Ccanputf.,, An 0:ivr- nary bits) into an arithmetie logic unit, manipu- .ccience, VOL I?, 1qn2, p. 755 lating the bits in accordaree with instructions. Ch. 3 Electronics -Technology 83

Table 3.Charac!,1ristics of Different Categories of Digital Computers

Family" Distinguishing features Microcomputer. The central processing unit (CPU) consists purpose data processing. Often mainframes were needed of a microprocessor or single-chip n.,:.focomputer, some- in such application's only a few years ago; minicomputers times several. The most common microcomputers use an tend to supplement rather than displace them. 8-bit word and sell, without peripherals but otherwise com- At the lower end, it is increasingly difficult to distinguish plete and ready to operate, for under a thousand to a few minicomputers from the more powerful microcomputers. thousand dollars. They will typically fit on a desk top Many less expensive minis now rely on a single-chip proc- fig. 9and do not require special training to operate. Ex- essor. However, he smaller minicomputers typically-use amples include popular models sold by Apple and Radio 16-bit wordse.g.; the currently popular POP-11 models Shack, along with the IBM Personal Computer. - made by Digital Equipment, or the Nova series of Data Gen- Machines based on microprocessors or microcomputer eral. Larger, more powerful machinessometimes called chips with 16-bit word iengths are beginning to appear, par- "superminis"normally have a 32-bit word length. Exam- ticularly for business applications. These are nearer in per- ples of superminis are the Data General Eclipse series or formance to low-end rninic-3mputers than to the 8-bit micro- the VAX models of Digital Equipment. In the 1960's, 32-bit computers originally ,oped for the hobbyist and per- words were found only in mainframes. sonal computer markets. Most minicomputers carry prices in the $10,000 to Minicomputer: Microcomputers, by the definition above, $100,000 range. A principal distinction between minicom- could not have existed before the development of the puters and mainframes is that minis seldom. require either microprocessori.e., before the early to mid-1970's. Mini- operators with a great deal of training or specially con- computers, in contrast, stem from the 1960's. A popular structed facilities. Mainframes, in rnt!'.?. usually early mini introduced in 1905 the PDP-8, built by Digital be permanently installed; some large computers dissipate Equipment Corp.was the first tow-cost, mass-produced so much heat that air-conditioning is needed even in mid- computer of any type. It was designed around a 12-bit word winter. and discrete transistors. Mainframes: The CPU for a mainframe typically contains sev- Minicomputers are small compared to mainframes, eral thousand logic chips, usually bipolar for speed. Word which can fill a room. As fig. 10 indicates, minicomputers lengths are commonly 32 to 64 bits. Mainframes often sup- are often about the size of a desk. Although not as portable port multiple terminals and peripherals fig. 11 and gen- as micros, many minicomputer models can be mt.:led rela- erally require trained personnel onsite. tively easily within an office or factory environment. While i BM is the world's largest producer of mainframe Minis found much of their market as dedicated proces- computers, more than a dozen other firms build machines sors designed into more complex systems, or in special comparable in computing power. Mainframes can sell for iz-ad data processing taskse.g., industrial controllers, $10 trillion or moreexclusive of peripheralsbut the data acquic.iiiion systems for laboratory research, inventory more popular general-purpose machines typically cost management in factories. While such applications remain under $5 million. common, minicomputers are also dely used for general.

SOURCE 014,ce of Technology Assessment from a program stored in memory, and ends Peripherals such as disk and tape drives, ter- information back to the memory or to an out- minals, and print7-3 are made by large num- put device. The stored program, whichmade bers of independent vendors, as well as by com- the modern digital computer possible by pro- puter manufacturers. Nei:11;y 90 American viding a means for telling the computer what firms were producing terminals as of 1980, try cic without the need for hardware cl .;nges, about half of these "smart" by virtue of em- accounts for a good deal of the information that bedded microprocessors, while nearly 30 had enter:; and leaves the CPU. Even rather simple announced their intention to build 8-inch Win-. computer systemsfigure 0typically are sev- chester disk drives, a product just beginning eral types of memory. .wh'are desci bed in to reach the n;larketplace at that time.19 The more detail in appendix 3B. market dynamics associated with the computer industryrapid growth, intense competition, Computers as Systems new entrants with new productscharacl,rize peripherals and &ware as well as processors. edition to CPU and memory, a computer and outp,q devices (fig. 13), kteeds ""The DigithlAge," Electronics. Apr. 17, 19:A. p,.:487; G. Sluf.. srnal comput==rs, all the components ma} sker, "28 Rivals Eye 8-InchDisks ButNoneLands F. OEM he integrated inV, a siugie' self-donteined unit. Yet,'Fitic::r!,:nic News, Jan. 21. 1960, p. .1C. 84 International Competitiveness in Electronics

12) re S.Typical Microcomputer Intended for Figure 11.Data Processing InstallatioL atilt Around Personal and Small-Business Applications General-Purpose Mainframe CompAer

Photo credit. Control Data Corp.,

Figure 12.Simplified Block Diagram of a Microprocessor System

ROM 0t*-

Photo credit: Apple Computer, Inc. Addiess bus Data (to call bus instructions and (to and from data) memory) Figure 10.Typical Minicomputer Installation Including a Pair of Terminals and a Prinier RAM}

r --I

CPU 1

1 --Controi- unit : logic

Power I 1 Regist bus Clock I 1 J1 I/0 (input/output) bus (external control, D /A. A/D)

D/A = Digital/analog conversion A/D = Analog/digital conversion ROM = Read-only memory 1.110f0 credit: Digital Equipment Corp. RAM = Random access memory SOURCE Office of Technology Assessment. Cn. 3Electronics Technology a 35

Figure 13.Elements of a Gonoiiil.Putpose Digital "dunib' terminal (an input/output device with Comp, Aar ::,ystorn no fu .coon other Liao. to communicate with I the central processor)is to link a centralcom- Terminal( puter to satellite machines which can share the Tape. disk drives processing load. Many sir h distributed processing schemes are possibie, among themore common being a mainframe -',.i.pported by minicomputers. A mainfranie or mini caralso communici,::- with "smart" terminals..:at Cache carry out memory computations, compile pro- Central grams, and u'inerwise relieve the central or host processing computer of some of the work. Point-of-sale ter- unit minals found in retail stores often functionas ic..;PU) parts of distributed systems. The terminal not Main memory only acts as a cash register, but sends dataon purchases to a central computer that canman- age,aventory, compare sales volume by brands, and provide other information tomanagers. Automatic banking machines are another fa- Auxiliary storage miliar example; each automatic teller functions Ls a smart term; linked to the bank's central computer(s). These systems nay inch: fie hundreds of machines spread over several Output States.

Pr -.3r Networking is a related term, referring to dispersed machines that communicate withone another but are each autonomous. Anyone ma- Tt,,h119:oy,y chine can transmit data to any other; control of the network may be distributed over thesys- Technological change in computing has been tem or may reside in a designated processor. rapid since the }ieginning of commercialpro- In some but not all cases, netwo7'.ed computers duction in the 1950's, but now the industry is not only communicate and share control, but perhaps facing the most comprehensive set of also share the processing load. Local networks changes yet. These stem from "distributed in- serve a limited group of users, such as a single telligence," the dispersal of computingpower office. At the other extreme, a multinational to many farflung locations. In some respects, corporation might link computers located in this trend began with the development of time- many countries to form a worldwide network. sharing in the early, 1960's. Time-sharingper- mits users at i'emote terminals to interactdi- Computer Software rectly-with a central processor, extending the capabilities of a powerful computer tomany Physical equipment, o. .rd'.vare7--ranging people simultaneously. It alsouses the proc- from ICs, to disk drives, to netNti orks--has been essor more efficiently. Even during big jobs the the p 'Inary subject above.Butmodern com- CPU maybe idle much of the time; with time- puters. depend just as heavily on software. The sharing, system software keeps the CPU busy programs /hat stand between 'user and CPU tell by dividing its processing poWeramong many hardware what to do. Arrayed in several people, each of whom is unaware of the others. levels, they range from applications software written in langua!!os such as Fortran or Cobol Conceptually, the next step beyond time- the only type of program that the typical user shcring---for which each user needs onlya ever sees--to operating systems that supervise

9? -56 in!er.lat;onal Com,-)etitiveiess rr, Electronics arid iinate both hardware' and soft.vare a result, the largest part of thetotal cost to the ;Aim', !us, It is the softwarearchitecture, oper- user of a large computer system. is nowsoft- at in e system, compilersthat allows complex ware, rather than hardwarefigure 14.The networks- of computer and communications chart appli is to both purchased softwarefrom components to control steel mills, regulate air computer manufacturers or independent ven- traffic, determine the path of a guided missile. dorsand to user-developed programs; soft- distribute secial security checks. Hardware ware maintenance is alsoincluded. At one and software in conjunction determine system time, many computer manufacturers provided performance, and customers weigh both as- system software such as control programs, pects when making purchase decisions. In language processors, and utilities free to hard- some cases this may entail buying softwareand ware purchasers. Now, separatecharges are hardware from different v endors and assem- the rule. For example, IBM currently sells tiiing a unique system. The spread of distrib- about Si billion worth of software per year, ac- uted intelligence, new applications of cqmput- counting for a little over 5 percent of the firm's ers in liouws and offices, shopfloor automation, total revenues; in newer systems such as the computer-aided engineering analysis and de- IBM_ 4300 series, nearly half the price of a typ- signall depend more heavily on versatile, reli- ical installation is for software.22 Similarly, able. user-friendly software than on hardware. more than half of the R&D commitmentof a typicil computer firmmeasured either in Since the begins tugs of large-scale commer- terms of total expenditures or in terms of man- cial production, computer hardware has be- powergoes toward software.23 come steadily cheaper relative tosoftware, Costs for hardware have decreased by a fac- Cost trends for the deyelopmt,nt of 'imE.ware tor of at least 1.000, holding processing power for dedicated applicationse.g., the logic for constant, over the past 25 years.20 In marked an embedded microprocessoraresirnir. contrast, software costs have not decreased ap- Even the simplest such application will requr e preciably, and may even have risen in real debugging and testing of the program to veri terms. A single line of programing, as a . :le- that it functions as desired. Software develop- a f-thu mb, costs in the range of SIO to $50in ter ment for a microprocessor application may inflation, about the same now as in 1955.21 As 31Missing Computer Software," op. cit. ""Computer Technology Shifts Emphasis to Software; A Spe- computt.r Business tVee.k Sept. 1, 1980, cial Report," Electronics. May 8, 1680, p. 142. p 41, ,Lf: iMprGvement depends on the type ur-e'-asioned. pry-inc.:I.:it:: in programii g--;is measured in lines of Figure 14.Relative Hardware and SoftwareCosts le per inof timehas pc,,babiy increased, the rata of in. Faced by Users of Larger Computer Systems ease has barn orders of magfAtUde slower than for hardware isfa uanan«,. fur instance, points out that cchile pro- 100 gre-rn r pro,luctivity has increased by about a factor of 3 since - y,,-emp,Tforinaiva, to-cost ratios have gone up by roughly 2r the -,arne brut; ind--J. S. Birbaum, -Computers: A 8 !i,ir%;'. :rends and Limitations,"Science, Feb. 12, 1382. p. 760. Hard,', are In s, i!ie uniru tasks. productivity has probably remained 60 rather stable -perhaps CI,V11 tit'Utt!atied. Applications program- may now be sornowhat more efficient because of improve- news in higher level languages. Productivity in systems pro- AO gra Ming, or developing software for dedicated microprocessors. tnicrocomputers, and minicomputers, has probably not improved Software as rapidly; when ,,sterns become more complicatcAl, many of Iht in program developmentfrom conceptual design to 20 (101,411..;golg---!)Icomo man! arduous. Even a relatively simple pro- };I:t111 MO' have of the orde r of 1020 different execution paths, 0 depending on the number e, r loops, branches, and subroutines. 2000 Cost,: per line cuts escalate program size and complexity in- 1960 1980 ; ease. Another common . ale-ofahumb is that a man-month of effort is required to demonstrate that 100 lines of coy .1 is. for Yam practical purposes, errur"Hsr and functionally correct, SOURCE- Office of T1 ,,notogy Assessment.

9 ,1 Ch. 3L-7f:ctronics Technology 87

perhaps Si billion in ,ff-the-shelfprograms per . year, and twice that amount in custom programing.25 GroWth of Computing Power 14 gave one picture of the rapidity of cl in computer technology, and in the co. !ter industry in general. But changeex- tends far beyond the relative costs of hardware and software, the rapid growth in themicro- computer market (now about 50 percentper year), or continued increase in performance/ cost ratios for computer systems. And, while distributed intelligence ma eventually have broader and deeper effectson the way people live and work than big machines, the absolute rise in computing power delineated in table4 illustrates simply :Jut dramatically how rapidly the capabilities of the most powerful digital Photo crecla TIktronlx, rric computers have increasednine orders of Microprocessor development system magnitude since the close of the Second World War, six orders of magnit-Ide in the 30years since the introduction of uae first commercial cost several hundred million dollars, with machine, the Univac I. All the computers listed estimates for 1985 running to S3 million or in table 4 would be classed as mainframes, And those of recent years as supercomputersrep-; The rising relative costs of software have resenting the maximum in computingpower been one factor in the rapid growth of inde- available at a given time.* pendent firms that develop and marketpro- While the biggest computers have beengrow- grams of 111 types. Many conioter manufac- ing in speed, smaller machineslike allcom- turers have iraditionally been re.= ter liz:rdware- putershave been growing in performanceper oriented, leaving an attractive market forven- dollar. Table 5 comparesan 8-bit single-board dors who concentrate. on software (seeapp. C, microcomputer r,Tresentative of 1970's tech- -Computers: A Machine for Smaller Busi- nology to the.IBM 650a first-generationvt,c- nesses,- on the role of "systems houses" in the uum tube processor of the mid-1950's. The two development of the minicomputer market). machines are roughly comparable incomput- Even IBMwhich has built its market domi- ing power, but the modern microcomputc:.is nance in larger machines on software as much. orders of magnitude Smaller and cheaper, as hardwarehas turned to independent software firms to supply programs for its personal computer. Independent software vendors sell z,ty. 0. Gardner, "The Key to Greater Productivity."Dun's Review, August 1980, p. 74. The total value ofcomputer soil- ware in Use worldwide probably exceeds $200 billion, ,""Missing Computer Software," op. cit. One supplier of rnicro- 'Arithmetic operations per socond, themeasure used in the procasso, s and microcomputers has estimated that a typical table for comparing computi71gpower, is not a perfect yardstick rnich1970"!'. applumm III carried a software developmentcost of because many data-processing about 520.000 ($2t) per line of code), but by 1980 the programs are limited by opera- cost was tions other than arithmetic.a.g., invertingmatrices. More so- 5100,000 to half a dollars ($35 per line, of code, phisticated comparisons employ -benchmark" but also many :non, lines). Meanwhile the b "rdware programs based costs have on repraso..tative tasks. Arithmetic operationsas used in the remained r.bout the sans, -in the vicinity of $100per unit. See table have the advantage of beim; J. G. rasa, "Inb.; Takes Aim at the '80s." Electronics, Feb. easy to understand and ap- 28, plicable to early Model computers.some of which could not ex- 1980. p. ecute modern benchmarking pi.ograms, 6'3 `^ti;-,,,-,,,,Vional Competrt, i.±_-,ness in Electra-rics

Tab -1.Increase in Comp Ming Power Over Time in time by only two decades. Figurc 15 gives an alternatic9 p:-.:ture ofgrowth in perfor- 2omputational speed (arithmetic operations manceperdollar. The plot shows the decline Year Model per second) in price for a minicomputer family after1965 1944 Ha^:ard Mark I the pioneering PDP-8, although the first several ielectro 0.4 years apply to an earlier modeltherapid fall 1946 45 stemming in part from learning curve phenom- 1951 Univac I 270 1953 IBM 701 615 .-na as for semict:1ductordevic=:.!s. Drops in 19t1,1 IBM 7074 33.700 prices for the semiconductors a machine.con- 1963 Ci:/C 360C 156,000 tainsfigure 16also lead to cost reductions. 1965 IBM 360/75 1,440,000 CDC Cyber 176 9.100,000 Digital Equipment Corp. introduced the PDF-8 1971972 Cray 1 80.000.000 at $18,000; by the early 1970's some versions 1981 CDC Cyber 205 800.000.000 were price.. as low as $2,500.20 fi;,!.Technologv Forecasting Literature Emergenne arid Im- ;r Inrcvatton."P. Kelly and M. I:ranz.trri; 'ids I. inlo.sfol. A Critical Review of Current ftii,ow,,:lgo "Generations" of computers can be distin- -ar.c.sco SanFrancisco Preis. 1978i. p. 300; "Tne CI:g.tal Age." guished based on advances in the technology. Frecricanics,-Apr 17, 1980. p 382. P. J Schuyiten. "Me. Ba rite in Super- Krea Yor-ii Times. July 22, 1980. p. Dl For example, the IBM 650 (table5) represents a first - generation machine, the F-8microcom- andat least as significantvastly rnore puter third generation. Zeroth - generation sys- Note that these computers are s1parated tems were similar to the650 in usingvacuum circuits 'l;. exhibit reliabilitiesmeasured tubes, but early computers such as Er.iac

as mean timer., beRveen fa. , Nof the order of 10'' hours! lacked the ability to execute stored programs, gate. Thus. a type al ink. .,wiessor containing 10,000 gates the hallmark of the modern digital computer. might have a mean time between failures of about 10 million cr ;moo years, In cont :.ast,meantimes between failures To change &program in Eniac mean: altering fur discr,!to tran,INtorsare :moot 108 hours, for vacuum tubes, leasthan I0" huufs. Sue, S. Midde1hoek, B. Angell, and D. 1. "%fin:oprocessors Get Integrated Sensors." IEEE "C. Lewis. "Small Computers," Ei;rtronic News (an. 25, 1902, .covtrinn. February 19B0. p. 42. sec. II, p.70.

Table 5.Comparison of IBM 650 (1955) and Fairchild F-8 Microcomputer (19/0's)

IBM 650 F-8 Remarks Physical volume (ft') ..... 270 0.01 F-8 about 30,000 times smaller' Weight (pounds) 5,650 Power consumption (watts) 17,700 2.5 F-8 consumes 7,000 times less power Memory (bits) 3K main, 16K ROM, EK RAM 101.)K secondary CPU 2,000 vacuum tubes 20,000 transistors 650 alsn needed many disgrE,te resistors a.no capacitors Time for adding two numbers (microseconds).. 750 150 Reliability (iitean time between failures) Hours Years (3 million to 10 million hours F-8 at least 10,001) times is a typical mean time between more reliable failures for a current microprocessor--More than 300 years--but the srbsystems with which the microprocesscr communizates e.g., terminals, printersmay be mucl, less reliable) Cost $200,00. (1955 dollars)ti!ider $1,000 with terminal 18M 650 mlormation horn... "1978 First Quarter and Shareholders Meeting Report," Texaa Instruments,Inc ,trcnild F-8 infoTrha.tIonfron1, G. Linvnll and souncr Mar, 18, 1977, p. O. CL iogan,-Intelkctua1 andEconomic Fuel for the Electronics Revolution,"Science, Ch. 5Competitiveness in the International Electronics Industry 187 development programs, and now have chosen ing other products as well. Because U.S. man- to sell Japanese VCRs under their own brand ufacturers cannot expect cost advantages, they names. may be left with only their distribution systems and brand recognition as prime competitive The point is that, once having lost product weapons. To a considerable extent, Japanese leadershipas has occurred with VCRs firms have already countered these advantages: American firms find it increasingly dif- thus, the long-term prospects for American ficult to compete in new technologies, and may firms in consumer electronics do not appear eventually find themselves importing or adapt bright. Semiconductors Technological forces have dictated the mar- circuits needed for consumer products. As the keting strategies of semiconductor companies technological level of Japan's semiconductor in all parts of the world since the inception of industry caught up with that of the United the industry. The case study on 4K RAMs in States, many of these imports were replaced appendix C points to the importance of engi- by indigenous production. The phenomenon, neering capability for U.S. merchant firms termed import displacement, has been charac- such as Mostek or Intel. Technology is no less teristic of Japan's computer industry as well. important now than a decade ago, when the Displaced items quickly become potential ex- 4K RAM was being developedbut as late as ports for Japanese firms. the mid-1970's the business strategies of foreign During the 1970's, awareness of the possible semiconductor manufacturers were of little in- consequences of foreign competition grew terest to Americans. As the 4K RAM case dem- within U.S. industry and Government, al- onstrates, U.S. firms appeared to have little to fear from producers in Japancertainly not though the production and trade data showed from those, in Western Europe. But from a little cause for concern. The. Federal Trade minor position in 4K chips, Japanese firms Commission, reporting on interviews conducted in 1976, stated: "... anumber of com- went on to cllim about 40 percent of the world pany executives expressed the opinion that market for r.ollowing generation of 16K competition from foreign companies would be RAMs. By 1982, the perception was wide- much tougher to handle than competition from spread that U.S. firms had "lost" the market other U.S. companies in the next 5 or 10 years. for dynamic RAMs. Certainly this is an over- In contrast, some other executives felt that U.S. dramatization, and the RAM market can by no companies would not have a difficult time means stand for the industry in microcosm; but maintaining their technological lead over the picture has changed. How did it change so foreign companies."27 Hindsight shows those fast? of the first persuasion closer to the mark. During the 1970's, Japanese companies ex- One sign that patterns of international com- ported considerable numbers of electronic petition would change came from subsidies, componentsincluding transistorstothis and promotional efforts adopted by foreign country, but the major growth segment, ICs, governments with the aim of fostering in- was dominated by 'American suppliers. Even digenous production. Japan, France, West Ger- though Japan's Government protected the local many, the United Kingdomall in one way or industry, U.S. shipments took a substantial part another marked the semiconductor industry as of the expanding Japanese IC market. Custom- ers in Japan depended on American firms for critical to continued economic vitality, an in- devices that domestic manufacturers could not 27Staff Report on the Semiconductor Industry: A Survey of providehigh performance or large-scale Structure. Conduct and Performance(Washington, D.C.: Federal chips, custom-parts, even some types of linear Trade Commission, Bureau of Economics, January 1977), p. 130.

193 188 International Competitiveness in Electronics dustry not to be given over to foreign interests. Figure 34.Distribution of U.S. Semiconductor Since the United States was far ahead in both Sales by End Market

technological expertise and production, vol- M !nary Consumer . Computer and ume. the implicit targets were American com- industrial panies, not excluding those that had invested in local production facilities. These government-led attempts to build competitive semiconductor industries have had Mixed result.s. Joint projects involving public and private sectors in Japan were quite successfulin semiconductors as in earlier Japanese industrial policy initiatives. European attempts have been far less fruitful, for reasons that may have as much to do with the characteristics of the industry and marketplace on the continent as with the policies pursued.

1960 1968 1980 United States Year Applications of semiconductors reflect ongo- SOURCES 1960. 1968: "innorafion. Corrpctitu;. !.16 t,,...,:entilental Policy in the Semiconductor Industry," l..;,,ertArt Associates, inc., ing synergistic relationships among merchant final report for Experimental Tei.hr..,:l!'iT, eAl^oires Program, Department of Commerce. Mi,rch.1130, suppliers and their customers. Purchasers out- 1980: Status '80. A Report o fthe Intck.vted :nOustry side electronics have lately presented growing iScottsda:e, Ariz: Integrated Circuit En;.,e market opportunitiese.g., in automobiles. p. 34. NonethelesS, from a technological viewpoint, firms building computer- or microprocessor- where the needs of customers are nr ire div.arse, based systems remain the most influential cus- and nontechnical dimensions like price.0 de- tomers (fig. 34). Manufacturers of consumer livery schedules more important. electronics, communications systems, instruments and controls, and office equipment have Factori in Strategic Decisions considerable impact as well. While most of the attention below goes to merchant firms,/cap- Competitive strategies adopted by merchant live operations have played a vital role in the semiconductor firms revolve around factors technological development of the U.S. ,indus- such as size, market power and technological try. Furthermore, production decisions' by the capability, internal need for devices (if any), larger integrated manufacturers sometimes and stage of development relative to others in have major consequences for the merchant the industry. A company's technical strengths market. shape its product line. Process technology whether a manufactu-re,f is strong in bipolar or Figure 34 shows that the phenomenal expan- MOS (metal-oxide semiconductor), which va- sion in semiconductor outputduring the 1970's rieties of MOS a-firm knows bestis one as- was accompanied by amajor shift from de- pect, design capability another. Some com- fense purchases to consumer andindustrial ap- panies are known for innovative circuit de- plications; competitive success inthe most signs, others for prowess at mass produc-, rapidly growing market segments depended on tionsome for both. Smaller entrants tend to the ability to make the transition fromspecial- specialize; only a few merchant suppliers have ized military requirements to thedemands of broad product lines (the world's semiconduc- private sector customers. Somecompanies that tor manufacturers supply perhaps 50 billion "fared quite well in the militarymarket could devices a yearof 100,000 different typesto not compete effectively forcommercial sales, several hundred thousand customers). Ch. 5Ccrruc,..1.tine!...:-..7 Pille.-iilational Electronics Industry 189

-%,67K at the close of the .1970's, c'or.:*,; -fuy's Q,::;-le.i.:pected entry contributing .,..;AoriF.,gPs of ih,''se devices. Capacity short- by firmsprimarily,._ stemming firoe:n invest in additional production is:,.pities .tn the of the recession of 1974-75 an open c"...)or for Japanese IC suppliers

_ , in. this c.:;:ultr,. Captive maotUat- Curers contribute in a major way to the Otrec-'1:, f-ti-ength of the U.S. position .n tnicroelectr.-i'cs through their R&D activities;: particu'ici-..1.BM and AT&T's Bell Lab- orazolies have !.,7;en.respons:ble for much of the Pho,;. Electric Co research ,..-ncieriying_tne semiconductor Technician loading r.i'i.taii:6,,i7:;::;,c) ik-,.clustry in thi-7, country, indeed around the semiconduclf;:r M'4?rcharit firmsbecause of the pace .f.nteristy.or.:pr '1:tuct. development, the con- cycles tr, improvement in design and I characterize succeeding A number of U.S. mi:.1-..::,!:.ani..::..e.7.;ufacturers '.Cs---must set different have integrated toSGic:l_:7 i'ato systems. -t .,,YIT also hava more limited resources A few, such as Texas IL-it.ri,,,7,.t,sAtfi. have always been diversified. Others htM1 put chased by larger enterprises but sZ;;Ii the great bulk Ir.aes.rt(A11:-.1.! rs of their production on the open ffiarhe't fch. 4, a new device must table 24). As merchant supplf...ers expand, so UF44/.;111 their design is at the does the range of their product offerings. thr. of the artit will be Smaller companies with limited resources aim supercedt.:1 sooner. Timing is crit- at niche markets. Newer entrants set out to de- Tt.i: . sometimes open, velop specialized or custom devices of less in- voviding opportilnities for leapfrogging the terest to larger corporations; tho 1981 startup cGaipetition. Companies .that quickly mastered Linear Technology - -a spinoff from National production of dynamic RAMs, or concentrated Semiconductorspecializes in linear. ICs, in on microprocessor design architectures in at- which the founders have expertise. tempts to tie up large portions of that market, Captive semiconductor producers have dif- were aiming at such advantages. Needless to ferent strategic aims. While most of the larger say, some firms have better records at ex- computer firms make some of their own logic ploiting these opportunities than others. A chips, IBM has traditionally produced most of number of companies that had been strong in its memory circuits as well. The company con- bipolar technologyincluding Fairchild and sumes so many that, for products such as Texas Instrumentsdid not move as rapidly RAMs, on occasions when it chooses to pur- into MOS as the competition; Texas Instru- chase from outside vendors it can account for ments staged a quick recovery,, while Fairchild a sizable proportion of total demand (the com- 'has continued to lag. Mostek, as its name con- pany is probably the largest single purchaser notes, was founded with the intention of spe- in the merchant market, as well as the world's cializing in MOS; the company has emphasized largest producer of semiconductors). This then memories, designing their own RAMsthe de affects the business decisions of merchant sup- facto industry standard 4K RAM was a Mostek pliers: II3M's external purchases were a power- design (see app. C)while serving as an alter- ful and rather unpredictable force on the mar- nate source for microprocessors. Electronic 19J 190 internacional Competitiveness in Electronics

Arrays, now ol.,:med by Nippon Electric, had is likewise rather easy to predict. Under such specialized in read-only memories (ROMs). circumstances, Japanese suppliers quite nat- Other firms seeking to exploit particular tech- urally emphasize memory products. nological paths have had less success: Amer- ican Micipsystems' work on v-MOS ICs is one In contrast, market acceptance of logic cir- example, RCA's pursuit of silicon-on-sapphire cuits or microprocessors is less predictable. In- c -MOS another. Internationally, Japanese firms vesting in a new microprocessor design=the moved into MOS ICs much more rapidly than 32-bit Intel iAPX 432 cost more than 5100 mil- the Europeanmost of whom are still well lion to developis risky, but the potential re- behind their competitors in the United States wards are great; designs with an edge over the or Japan and relying on technology imports to competitionin performance, ease of program- try to catch up; this has been one of the objec- ing, adaptability to a wide range of applica- tives of Le Plan des Composantsa major in- tionssell for premium prices. -'3 Furthermore, dustrial policy effort by the French Govern- microprocessorsbest thought of as families ment (ch. 10). In the United States, some firms of related ICs rather than unique deviceshave Signetics, Monolithic Memoriescontinue to longer product cycles, extending the period specialize in bipolar devices. IBM likewise re- over which investments can be recouped. mains relatively stronger in bipolar than MOS; Memory circuits are manufactured as long as demand holds up, but sales tend to peak and the speed advantages of bipolar circuits have decline more rapidly than for other device `!d many computer manufacturers to continue types. Five or six years elapsed between the emphasizing the older technology. onset of high-volume production for 8-bit mi- Quality and reliability comprise another com- croprocessors and mass production of thesuc- petitive realm where strategies depend both on ceeding generation of 16-bit parts, while life- circuit design and manufacturing practices (ch. cycles for RAM chipsthough slowly length- 6). While Japanese firms have zealously pub- eninghave been perhaps 3 years, onmetimes licized the quality and reliability of their ICs less.29 in much the same way that Japanese consumer electronics firms used reliability as a Abbreviated product cycles dictate strategies wedge into the American TV marketdomes- aimed at profitability within a narrow time tic producers like. Adva need Micro Degices window, along with continuous efforts to develop new or differentiated offerings. The lat- have also pursued an'age of quality and ter can be original designs but need not; sec- reliability as a marketing tool. ond-sourcing has been widespread for many years, in part because customers often insist Products and Prices on more than one supplier before they will One of the attractions of memory circuits design an IC into their end products. Thus, sec- in addition to the vast marketis the relative- ond-sourcing can accelerate market expansion ly orderly aid predictable progress of the tech- for everyone. Semiconductor firms choose to nology itself; circuit design is vitalalong with become alternate sources for chips developed excellent process capabilitybut more straightforward than for logic or microprocessors. za0n the costs of microprocessor design, see R. N. Noyce and Everyone in the industry knows that the next M. E. Ho:f, Jr., "A History of Microprocessor Development at Intel," IEEE MICRO. February 1981. p. 8. Intel's first micro- generations of dynamic RAMs will be 256K processor, a 4-bit device, was designed in 9 months by a single chips, followed by 1 megabit; circuits offered engineer; 100 man-years went into the iAPX 432. by various firms are much more similar than "Inters 8080 familyintroduced in 1974 and the largest selling 8-bit microprocessorwill no doubt remain in production the designs of competing 16-bit microproc- for many more years. Worldwide, more than 10 companies still essors. One result is the fierce price competi- produce 8080 chips. MOStekan alternate source for another tion that has often seemed the dominant char- popular 8-bit processor, the Z-80for a number of years produced more of thedevices than Zilog, the originator. See The acteristic of the memory market. Progress in Antenna." Electronic News, Mar. 12, 1979, p. 8; "Eight-Bit static RAMs, and in the various types of ROMs, Level." Electronic News, July 5, 1982, p. 12.

19u Ch. 5Cornpetitiverriss in the International Electronics Industry 191 by other companies to beef up their own prod- below those of its rivals, margins above. Com- uct lines, perhaps by complementing circuits panies with proprietary designs often license they already build, as well as to reduce market alternate sources, but at least at the outset risks and save on R&D expenses. From the second-source suppliers will be at a cost disad- viewpoint oi the originator, it may be more sen- vantage. If the initiator decides to follow a pric- sible to settle for a smaller piece of a rapidly ing strategy keyed to anticipated cost improve- expanding market than to try to keep others ments, follo-vv-on firms may find it difficult to from duplicating an IC design. Attempts to pre- make a prosily. Texas Instruments, for instance, vent duplication are virtually impossible if a has had-the reputation of practicing advance circuit finds an enthusiastic reception in the pricing whenever possible. In a very real sense, marketplace. As one consequenc-e, formalized then, later entrants can be at the mercy of in- alternate sourcing agreements have largely re- novatori thc latter choose to cut prices placed the copying that was once common- and exercise the cost advantages of being far- place. Sometimes alternate source manufac- ther down th.f: learning curve. On the other lurers acquire the originator's technologye.g., hand, an innovating firm might choose to in- mask sets for lithography. Other times only crease margins by holding price levels high. drawings or specifications are provided. The Under such circumstances, an alternate source recent agreement between National Semicon- may itself be able to carve out a place through ductor and Fairchild, the latter acquiring the price. One facet of Intel's corporate strategy right to build National's model 16000 micro- has been to choose products where it could processor in exchange for developing a com- enter the market first., reap high profits, then plementary line of peripheral chips, is an in- move onleaving later sales, at lower margins, creasingly popular route. to others. Nonetheless, in many cases, especially where the innovating company is small, lin- Cost reductions via the learning curve (ch. ing up an established supplier as a second 3; fig. 5) help shape competitive strategies. As production volumes increase, yields rise and source may be a preret,:lisite to sales in any vol- manufacturing costs drop. Pricing decisions ume. . have often been based on projections of ex- A further strategic choice, increasingly crit- pected cost reductions into the future. For a ical for American firms, is whether to design firm early to market with a new design, cost and produce commoe!ty-like chips or to con- advantages over potential rivals can build rap- centrate on custom or semicustom devices. The idly, increasing with leadtime. Firms that are first option entails high-volume production of. late to market face a dilemma; they may have ICs, that are, or may become, shelf items to choose between foregoing participation or standard circuits serving the needs of cus- pressing on with their own design in the hope tomers who design them into end products. that it too will win acceptance. In early 1982, The alternative, customizing, can be ac- with six Japanese entrants mass-producing 64K complished in a variety of ways; semicustom RAMs, versus only two American manufactur- chips such as gate arrays or programmable ers; a number of U.S. firms were confronted logic arrays are specialized' only at the last with such decisions; Advanced Micro Devices, stage of processing. Regardless of the techno- for one, decided not to build a 64K chip. logical approach, firms in the custom or semi- In different circumstances,then,firms custom business create specialized ICs meeting assume different strategic postures. Companies the needs of one or a few, rather than many, entering the market with a new device, par- purchasers. Because circuit design is expen- ticularly one incorporating proprietary technol- sive, prospective order quantities must be large ogyproduct or procesSmay have several ad- enough to cover engineering costs; alternative- vantages over competitors, that follow. An early ly, the buyer must be willing to pay a higher entrant will:normally try to remain ahead on price. Custom chips for automotive applica- the learning curve, keeping production costs tions are an example of the high-volt, me case, 197 792 fntemationaY Competitiveness in Electronics military systems of custom chip markets where cant monopolistic profits Yri in new technolo- production volumes tend to be small. Occa- gies), and 2) just as inevitably the pace of tech- sionally. end users design their own ICs and nical advance will render new product offer- contract out production. ings obsolete within a few years at most. Tht-, price history of the 64K RAM illustrates the At the center of competitive strategies in first point: offered in sample quantities at 5100 semiconductorsas for many industriesthen each in 1-59, and $2n tr, during 1980, lies the choice of products. Firms with broad prices dropped to the :5 range in early product lines may offer devices based on a 1981 and $5 to $7 a year later; during 1981, 16K variety of technologies while attempting to stay RAM prices, were driven down from S4 to at the technical frontier with only some of about Sl, largely as a result of price declines these. Otherssuch as Mostek in the mid- for 64K parts.3° Such pricing trends have meant 1970 'soperate within narrower boundaries that all firms, U.S. and foreign, have had to where they attempt to be leaders. Some en- work continuously at cost reduction. In (-entrants are content to-follow the obvious trends, trast to numerous other industries, passive: or offering unique designs infrequerely while reactive pricing policies are hardly possible; relying on other strengthsperhaps low prices although product differentiation is a viable or a reputation for qualityto attract custom- alternative under some circumstances, price ers. In its early years, Advanced Micro Devices competition in semiconductors is a constant took such an approach (see app. C). forceenough by itself to set this industry off from many others. The second market charac- Beyond 'these common themes, mostly hing- teristiE, rapid technological change, has forced ing on aspects of the technology, companies managers and technical personnel alike to plan their strategies according to the strengths adapt constantly; firms that have r and weaknesses the.; perceive in their own po- wedded to older technologies have faltered or sitions compared to those of their rivals. No disappeared from the marketplace. single company has the resources to manufacture and sellall the tens of. thousands of International Dimensions semiconductor products now marketed in the United Statesone of the reasons for the From an international perspective, the larger periodic emergence of startups. Extensive U.S. merchant firms have shared three major product lines can confer advantages where strategic thrusts: 1) offshore manufacturing to customers prefer to deal with only a few ven- reduce labor costs, 2) foreign investment to dors; broad-line manufacturers may also be serve overseas markets, and 3) attempts to do able to achieve economies by spreading mar- business in Japan. This last efforton which keting costs over many items. Nonetheless, a number of companies are just embarking. or such factors are sev.ondaly compared to choice reembarking after past rebii fr -it- of product and process technology, along with ical to.the continuing abili a variety of ingredients that could be labeled firms to compete with lai , Akir ,Ikt:U Japa- "entrepreneurship." Successful new compa- nese manufacturers, particulariy in commodity nies hale frequently been esta&ished to exploit products like memory. a particular product, often one that larger com- Offshore manufacturing investments have panies liave failed to pursueperhaps because been concentrated in developing Asian na- of limited resources, or simply a judgment that tions. Generally, the more labor-intenSive the pi.kaatial market was too small. assembly operationse.g., wire bonding and encapsulationh il.,lbeen moved. In the semi- No rnA,..,tter the decisions they themselves take, of semiconductor firins can be cer- ,,A dr '..1;i'erm 64K Pacts Until 4.11 of Jr,,L H-,ilkires of their market: 1) competi- U.S Fire; ':. I:; FterfroaP: P.Ti.u.s. Feb. 8, 1982. m. p. C. f I7 ;II ape (114) Studied." New Yu::: limes, :'on .P.,511,;ventually drive prices downward (it Mar. p ; for 64F. chips eventually fell ;;r, lows bopn extremely difficult to capture signift- of aboutS3 duriag the1982 slump.

1 96 Ch. 5--;a2mpe.titeness rn the interrtalionatElectronics industry T93 conductor industry, the :stimulus for these cii, MITI, and other agencies, controlled in- transfers has no bee-n. import competition, as ward investment flows and for the most part in consumer electronics, but domestic rivalry. prevented the establishment cf manufacturing Transfers offshore began in the early 1960's, facilities under foreign ownership:32 Joint -aen- long before Japan appeared a significant threat tures in which a Japanese. company held the in semiconductor production. By 1970, Ameri- maiority interest met a more favorable re- can companies operated more than 39 subsidi- sponse. The purpose was obvious: to pro-aide aries in such locales as Hong Kong.. Singapore, :shelter for 'Japanese companies which a;. the Malaysia, the Philippines, and Mexico." Re- time were well behind in semiconductor tech- locating labor-intensive production operations nology. MITI believedwith good reason, if has been especially attractive because transpor- the European case is indicativethat allowing tation costs are low; chips are often .,hipped American firms to r.roduce in Japan would sti- by air. A cost comparison illustrating the ad- fle the domestic industry, particularly when it vantages of offshore assembly is included in .carne to more advanced device types. In acting appendix B (table B-2). this way, the Japanese Government was behav- A second international involvement of U.S. ing much like others that haVe sought to pro- semiconductor companies has been point-of- tect infant industf.es, but ,Japan has often been sale production to serve developed country accused of maintaining rizotectionist mea;ures markets. Investments in point -of -salt plants long past the point at,Which her industries have began about the same time as offshore been able to fend .fir themselves. but from the standpoint of industry strat- In any event, as a consequence of protec- egy the motives were quite different. These tionism in Japan, American suppliers were have- been twofold. First, foreign governments forced to adopt business tactics differ ii. from have ofte:, taken steps to make local produc- those Pursued on the continent. Most re- tion attractive, or, conversely, to make export- sponded to MITI's entreaties and entered into ing from the United Slates onerous. European licensing agreements with Japanese produc- countries, in particular, have relied on incen- ers.33 Such steps were entirely rational, pro- tives combined with tariff and nontariff bar- vided the U.S. firm could be reasonably cer- rierstoattractU.S.high-technologyin- tain the technologies transferred would not vestments. Second, point-of-sale production find their way into products they would face can become a competitive necessity to the ex- at home or in third-country markets. With this tent that other firms have already made such proviso, it would pay to sell technical knowl- moves. edge, the proceeds from which could then at Efforts to establish sales, production, and/or least partially offset the costs of generatine :hat R&D 'at:jiffies in Japannow a bigger market knowledge. The outton.of licensing the,,ill of Europecomprise the most recent melds with Japanese firms have led to many overseas thrust by American firms. While second thoughts within the American industry. Texas Instruments had been able to establish Nonetheless, clear-cut, cases in which U.S. a wholly owned subsidiary in Japan, other technology was an irreplaceable ingredient in American firms were kept out until recently. the growing ,capability of Japanese semicon- Semiconductor manufacturers attempting to ductor manufacturers are ripe, particularly ill invest in Japan suffered much the same fate later year:,ble etxr,'optioLbeing perhaps de- other U.S. companies; the Japanese Govern- velopments ilowinp: from Bell Laboratories. ment, through the Foreign Investment Coun- "M. Y. Yeshino, "Japanese Foreign .'Direc. Investment,"The Japanese Economy in lnternationall'etrspective, I.Frank (ed.) (Baltimore; Johns Hopkins University Press, 1975), p. 248. 31W. F Finan, The !mei-national Transfer of Semiconductor "See, in particular, W. F. Finan, "The Exchange of Semicon- Technology Through U.S.-Based Firms, National Bureau of ductor Technology Between Japan and the United States,"First Economic Research Working Paper No. 118. December 1975, U.S.-Japan Technological Exchange Symposium,Washington. p..57. This figure excludes point-of-sale facilities in industrial D.C.. Oct. 21, 1981. Finan points out that American firm); general- nations. ly did not transfer proprietary information to licensees (p. 9). 199 194 b-i!ernational Competitiveness i,n il-i'ectranics

Texas Instruments became the single excap- necessity. The usual explanations for the exits tion to MITI's licensing ruleri reality only of a number of large corporations during the a partial exception. Because it held a series of earlier years of the industry revolve around fundamental patents covering ICs, Texas in- rigid decisionmaking styles. struments.was in stronger position than other More recently, the character of the market American companies. As a condition for li- has been shifting; American companies have censing its patents to Japanese firms, Texas In- been forced to alter their thinking. In some struments demanded that it be allowed to estab- respects the changes are- a continuation of- lish manufacturing operations there. When familiar patterns: more complex ICslarge and MITI refusedconsistent with its decisions re- very large-scale integration (VLSI)make still garding other electronics firmsthe stage was more applications co ;t- effective, creating new -et for prolonged negotiations." Texas In- and different puzzles for chip-makers. More struments and .MITI eventually compromised fundamentally, VLSI has altered the cost struc- in a 1968 agreement permitting a joint semi- ture of the industry in at least two ways. First, conductor manufacturing venture with Sony. production is growing more capital-intensive; Four years later, Sony sold its share to the new sources of financing are needed to pur- American company. chase more expensive manufacturing equip- Thus, Texas Instruments, although reported- ment (ch.7).Some of the capital has come via ly subject to a production ceiling, became the mergers, which have changed the industry's only U.S. semiconductor firm to both manufac- structure. The second way in which VLSI is ture and sell its devices in Japan, just 6s IBM affecting the structure of the industry stems hada few years earlierbecome the only from shifts in product design. What had been American company to build computers there. a hardware-oriented business is now systems- (IBM was also able to gain entry by taking ad- and software-based as well. ICs are becoming vantage of its patent position.) Only recently, more than components. To tap the vast poten- as the Japanese have gained confidence in their tial markets made possible by micr.i.processors own technical abilities, has MITI softened its coupled with cheap memory, semiconductor attitude toward foreign investment;; a growing manufacturers must commit substantial re- number of U.S. electrcnics manufacturers are sources to computer-aided design and software now contemplating wholly owned subsidiaries development. This comes at a time or in'!n in Japan. While the longer term consequences Eying internatio 011--) . or of these decisions are not yet clear, investment v. 'I ..out the Japanese in the picture, the prob- within Japan couldgiven the examples of lem facing U.S. merchant firms is one of lo- othci industriesprove a vital support for cating sources of new capital in substantial American firms seeking to compete'With Japa- amounts while battling to ---'serve even their nese rivals in third-country markets as well as c:isting pr-Ait mar 1:s,A companies devise in japan. t eir respt,..ises, se, ,ids are emerging. Greater vertical integration will probably Current Trends have the farthest reaching consequences. Larg- The competitive strategies of American semi- er merchant companiese.g., Texas Instru- conductor firms have been aimed first and ments, which has entered a variety of con- foremost at survival in a highly competitive, sumer markets, including that for personal rapidly changing market. Companies big and computers- -are taking advantage of broad- small have had to stay abreast of and adapt to based positions in microelectronics to integrate technological change. Flexible management ;lownstream into the manufacture and sale of and organizational structures have been a final products. The reasoning behind such decisions is straightfo ward. If much of the tech- 341. E. 'Mon, In!ornational Diffusion of Ton.iinology: Thu Cast, of Semiconductors (Washington, P.(;., Brookings Institu- nology of data procw..,',og and other electronic tion, 1971), pp. 141147. systems is incorporated in ICs, why not make

20(1 Ch. 5Competitivene;s in the International Electronics Ind-..i.stry 195 end products too, increasing value-added and where it is becoming common for such agree- profitability? To this strategyessentially an of- ments to spell out in considerable detail the fensive onecould be added a defensive ele- R&D and/or circuit design obligations of each ment. For semiconductor manufacturers with partner.35 the resources to contemplate entry into systems Arrangements in which two or more com- markets, greater vertical integration reduces panies independently develop different mem- vulnerability in the event that customers begin bers of a family of chips fall at one end of the to inte.z.-4-rate backward-into lee pi iuilLiiT::. The fact is that backward integration is on the R&D spectrumcomplementary product devel- upswing. as manufacturers of computers, of- opment. At the other end, closer to basic re- fice equipment, consumer durables, and a host search, industry groups are moving toward of.other products sense the need to develop in- cooperative rather than independent but com- house capability in microelectronics. One path plementary projects. The Semiconductor In- is purchase or merger with a semiconductor dustry Association and the American Elec- company. Merger activity in the industry has tronics Association have each established pro- been high since the latter part of the 1970's; by Grams that channel contributions from 1983 only a few of the larger, broad-line mer- member firms to university projects. The Semi- conductor Research Cooperative is funding re- chant suppliers remained independent. search directly, while the Electronics Educa- Mergers have been of several types: some tion Foundation aims to improve training in semiconductor firms have been absorbed into electrical and computer engineering, primari- conglomeratesone example is United Tech- ly through fellowships and faculty support." nologies' purchase of Mostek. Other acquisi- Another effort, Microelectronics & Computer tions have been more directly motivated by in- TechnolrCorp., will be an independent prof- ternal needs, as in General Electric's acquisi- it-seeking iacEi organization capitalized by the tion of Intersil. Foreign t L .13 been participating firms.37 At least six universities prominentSchlumberger cchase of Fair- are also establishing centers for R&D in semi- child. Sometimes the appare.motive on the conductor technology and/or systems applica- part of the semiconductor company is the need tions of inicroelectronics.38 Whether all these, for new financinc.; this no doubt a factor efforts will survive and flourish remains to be with Mostek, explicitly so with IBM'' seen. of a substir..!tial interest in Intel. The motives The emerging strategic picture in the United of foreign i-vestors have varied: buying an States, therefore, is fluid and uncertain. Semi- -nerican firm can be a quick way to get tech- conductor manufacture; along with other por- nology as well as a convenient entrance into tions of electronics, is undergoing far-reaching the U.S. market. restructuring, with outcomes that are hardly In a related development, many American obvious. Given settlements in the IBM and semiconductor companies are seeking alter- AT&T antitrust cases, the way also seems clear natives to "going it alone" in the development "For examples, see S. RuLisell and S. Zipper. "Intel. Motorola of new technology largely because of rising Tighten Fold on General-Purpose MPUs: See Peripherals Key costs. New variations on accepted technotfogy Market for Niche Suppliers,"Electronic News.Mar. 8, 1982. p. 1. U.S. merchant firms are also negotiating such agreements shanig arrangements have been devised. Some with Japanese menufacturers-5'. Russell. "Zilog. Toshiba to semi :inductor manufacturers have prevailed Swap MPU, CMOS Technology."ElectronicNews. Apr. 19, on customers for assistanceindeveloping spe- 1982, p. 53; "National Semiconductor Sets Venture With lap. nose Firm,"Wall Street loth-nalJan. 23. 1983. p. 22 (the Japanese cialized chips and software. Both General participant is Oki). Motors and Ford have supported such efforts. "S. Russell, "SIA Eyes $5M Funding for Research Coopera- Semiconductor firms have also .sought new tive."Electronic News,Dec. 21. 1981. p. 6. "C. Barney, "R&D Co-op Gets S6t To Open Up Shop."Elec- ways to share product development costs, tronics,Mar. 24. 1ry83. p. 89. among themselves, sometimes through exten- '8C. Norman. "Electronics Firms Plug Into the Universities," sions of past practices in second-sourcing; Science,Aug. 6; 1982. p. 511. 2 01 1.96 international Competitiveness in Electronics for continued expansion by these two giant ly half the American market for 16K RANIs. semiconductoricommunicationicomputer com- Two years later, Japan's manufacturers seemed panies. AT&T's manufacturing arm, Western well on their way to comparable levels of Electric. plans to be the first American firm to penetration in the next-generation 64K RAMs; deliver 256K RAMsa rather spectacular en- indeed, as sales began to build, the Japinese trance into the merchant market. Other firms share soared toward 70 percent. While there Xerox is oneare also contemplating broad. is considerable feeling that ultimately they will systems-pi ienteu strategies. Nicctim n e,Sind ltui be able to 1101 11,311 dbuut 11(1111 tilt, er companies continue to seek specialized U.S. market for 64K chips, any temptation to product niches that will prove lucrative while underestimate the capabilities of Japan's not attracting large and powerful competitors. semiconductor manufacturers has long since And in the background are the Japanese, add- passed. Seemingly countless studies recount ing another dimension that will continue to in- the strategic attack, tracing the targeting prac- fluence the strategies of American firms both tices of government and industry. domestically and internationally. From Linear to Digital Circuits Japan Firms in Japan had long since become major Until half-a-dozen years ago, few in the U.S. producers of linear semiconductors, a main- semiconductor industry gave much thought to stay in consumer electronics. By the early Japan as a serious competitive threat. Japanese 1970's, some American companies began to manufacturersalmost exclusively divisions of abandon this part of the market, especially as large corporationsmostly produced devices domestic sales seemed to be drying up. Leading for consumer products; even today, nearly half U.S. producers put their resources into rapid- of Japan's semiconductor output goes ,:o con- ly expanding digital IC technologies. Mean- sumer appiications.39 During the1970's, while, for tine Japanese, strength in consumer Japan's budding computer manufacturers de- devices was both a blessing and a curse. While pended on American suppliers for advanced giving their engineering staffs experience in ICs. While Japanese companies were clearly on circuit design andmore important in high- the way to the skill levels needed for mace ad- volume production, the concentration on linear vanced devices, the prevailing belief in the circuits did little to raise overall levels-of com United States was that they could not really petence. At the time, the primary customers for hope to catch up. The primary concern was the digital ICscomputer manufacturerswere a closed Japanese market. American companies relatively minor component of Japanese de- had been prevented from establishing a pres- mand. enCe remotely comparable to that which they This was the situation when, in line with its had achieved in Europe; customers in Japan longstanding policy of fostering internationally bought only those devices that were not pro- competitive industries, MITI acted to break the duced locally. impasse created by the focus on consumer Today the situation seems quite different. products. The agency helped fashion an R&D Japanese firms have emerged as viable global program intended to increase Japan's capabili- competitors in VLSI devices. Although their ties in large-scale digital ICs, particularly MOS prowess has centered on memory chips, they devices, and accelerate movement toward have made up a great deal of ground in logic VLSI. The organization of the program, which circuits and other device types as well. By 1980, began in iate 1975, is described in chapter 10; the gravity of the threat had become obvious; five companies and three separate laboratories quite suddenly, Japanese firms captured near- were involved in the 4-year effort. Funding totaling about $300 millionwas provided part- "Japan Electronics Almanac1982 (Tokyo: Dempa Publications, ly by the participants and partly by govern- Inc.. 1982). pp. 142. 143. ment. The program has had far-reaching im- Ch. 5Competitiveness in the International Electronics l:Idustry 197 pactsas much through diffusing technology rather than product technologies; one-third of and training people as through the technology the funds were spent in the United States sim- developed. A parallel government-sponsored ply on purchases of state-of-the-art manufac- V LSI program---Cc.is one focused on telecom- turing equipment. This suggests that the major municationswas carried out in the labora- thrust was to develop skills in low-cost produc- tories of Nippon Telegraph and Telephone tion of commodity-like devices such as RAM (NTT). which had the most capable microelec- chips. fz:rtpi,-,-; Rk-n !Ppan Two aspects of MITI's approach deserve par- MITI's objective was not only to aid Japan's ticular emphasis. First, subsidization of micro- semiconductor manufacturers: the VLSI pro- electronics-R&D was only the opening move gram was part of a much more extensive ef- in a broader strategy for building a competitive fort to move the nation toward knowledge- computer and telecommunications sector. intensive products with high export potential. Hindsight provides ample corroboration of Like its counterparts within the governments what was in fact an explicit goal: MITI's subse- of carer industrialized countries, MITI recog- quent support of computer and software devel- nized that semiconductors would be funda- opment, as well as the Japanese Government's mental building blocks for many sectors of the reluctance to allow open competition for NTT Japanese economy. Supporting the computer procurements. NTT was a principalthough and information industries was the first step. independentactor in the VLSI program; the MITI was fully aware that technical compe- government evidently hoped to restrict its high- tence in digital ICs would be essential, and that volume purchases to domestic manufacturers without some form of stimulus private compa- (the company does not produce its own semi- nies would find it difficult to shift rapidly from conductors), helping generate the economies linear to digital devices. From MITI's perspec- of scale so necessary for international correti- tive, support for "cooperative" R&D was a nat- tiveness. ural extension of past efforts in other industries; the VLSI program itself has been followed The second pointsuggested by MITI's level by related work in computers and robotics, as of support, generous for a government R&D well as further microelectronics projects. program but certainly not enough by itself to. boost the Japanese industry past American Still, by American standards, MITI's sub- firmsis that the VLSI project was never con- sidies were not large. Individual U.S. firms like ceived purely as an exercise in technology de- Texas instruments had R&D budgets that came velopment. Consistent the usual Japanese cl,-Ise to matching tho yearly outlays of the VLSI approach to govern:: '-ant- supported R&D, it project; IBM's corporate R&D spending was an was intended to focus industry efforts, help order of magnitude larger. Of course, partici- train engineers from private firms in state-of- pating Japanese companies continued their in- the-art technologies, diffuse these technologies ternally funded R&D programs; MITI spending within the Japanese industryin other words, thus gave a substantial incremental boost to to overcome weaknesses in Japan's technologi- Japanese semiconductor research, reducing .cal infrastructure created in part by the lack risks and supporting longer term work. Even of personnel mobility (ch. 8). so, total expenditures in Japan remained well below those here. Nor did the VLSI project re- This makes it doubly difficult to assess the sult in large and direct benefits to _Japanese contribution of the VLSI project. While sep- firms, at least in terms of product offerings. A arating what might have happened from what great deal of attention in the United States has did occur is impossible, pieces of evidence do centered on the thousand orSGpatents asso- exist: for instance, MITI excluded Oki Electric ciated with the program, but it is not clear what from participation and subsidies, yet Oki man- value these have. There are no signs of major aged with NTT's help to develop a 64K RAM innovations. Primary attention went to process that the company now exports in considerable 193 International Corn;;etitiveness in Electronics volume. Many of the events of the past few Domestic firms have been effectively protected yearsthe ups-urge in Japanese production and from import competition as well as from pro- exports of RAMswould probably have oc- duction within japan by foreign-owned con- curred in any event. albeit at a slower pace; cerns. Protection of growing industries through memory chips were obvious targets for Jap- government action is hardly unique, but can anese firms confident of their :abilities to only be judged to have succeeded if the pro- mass-produce relatively straightforward de- tected companies eventually emerge as viable :, sl:Indards.Thcy wan; also corn-petitors. In mi,-r-caectronics, the- "infant needed for the computers that these same firms industry" approach has been attempted else- were determined to make in greater numbers. where, most notably in several European nations, but only Japan has achieved success. Jap- In the United States, the impact of the VLSI anese industrial policy is discussed in detail in project has been exaggerated. It has come to chapter 10; here the point is that none of the symbolize not only direct subsidization of com- policy measures adopted by Japan's Govern- mercially oriented R&D but also interfirm co- ment, taken separately, appear to have been operation that might he illegal under American major forces in the ultimate growth and ma- antitrust. law. In fact, as pointed out in chapter turation of the semiconductor industry. Taken 10, cooperation diming Japanese companies together, they paint a different picture--one in has been rather limitedevidence of the which industrial policy provided vital guidance strength of the harriers within the Japanese in- and support for the development of an inde- dustry that MITI was trying to overcome; this pendent capability in semiconductor design aspect has been overplayed by American man- and manufacture. Cumulatively, the policies of ufacturers understandably distressed at in- Japan's Government have had a major impact. roads by iverseas competitors. While the VLSI project makes a convenient target, by itself it Strategy and Structure is a far-from-adequate explanation for penetra- tions of what had been traditional American Despite MITI's pervasiva influence, the com- markets. Indeed, government policies in sup- petitive strategies of individual Japanese semi- port of Japan's information industries have conductor manufacturers are governed first by ranged far beyond R&D subsidies. Among the the basic structure of the industry, which is other poliCy tools have been: populated by companies for whom microelectronics comprises a relatively small part of preferential -governineA procurement; their business. Most of these companiesOki favorable credit allocations, especially dur- and Nippon Electric being partial exceptions' ing the formative years: are large, integrated firms whose sales consist special depreciation and other tax meas- predominately of final products such as com- ures; and puters, consumer electronics, and telecommu- grants for training and education. nications systems. Table 45 shows that only for

Table 45.Proportion of Sales Accounted for by Semiconductor Products°

Japi.nese firms (1981) U.S. firms (1979) Nippon Electric Co. (NEC) 19.8°4 Mostek 93% Fujitsu 12.9 Advanced Micro Devices 89 Oki Electric 9.8 Intel 75 Toshiba 8.1 Fairchild 69 Hitachi 5.3 Texas Instruments 36 Mitsubishi 5.3 Motorola 31 5Including internal consumption. SOURCES: Japanese FirmsTable 29 (ch. 4). U.S. Firm "U.S. and Japanese Semiconductor IniluLt,les: A Financial Comparison," Chase Financial Policy to, the SernicOnductor Industry Association, Juno 9, 1980, p. 1.5. Ch. 5Competitiveness in the International Electronics Industry 199

NEG and Fujitsuthe latter Japan's largest competition in products representing a sub- 'computer manufacturer, with heavy internal stantial part of their total business leaves few consumptionhave semiconductors contrib- options for counterattacks. ye(' a proportion of total sales even half as great as for those U.S. merchant firms that are The Japanese strategyprotecting domestic least dependent on their semiconductor divi- semiconductor manufacturers from overseas sionsMotorola and Texas Instruments. Semi- rivals while providing modest R&D subsidies conductors account for less than one-tenth the acid at the same time fostering domestic com- sales of the other Japanese manufacturers. In petitionparallels that in television. It has th.s,, they are closer to American companies yielded 'equally impressive results: deep pen- like Rockwell or RCA, which nonetheless dif- etration in targeted markets based on low fer in being primarily suppliers of specialty prices and quality levels above previous norms. rather than mass:.market circuits. There has been a fortuitous element as well; unexpected demand swamped U.S. suppliers The fact that the major semiconductor sup- during 1979 and 1980. As a result of continued plitrs in Japan build end products creates capacity expansions during the preceding mar- potential intracorporate synergisms absent in ket slump, Japan's producers were ready to fill companies that are primarily chipmakers. the void. While some U.S. managers view integration as dyrtfuncticinal, likely to sap entrepreneurial Sbme spokesmen for the American industry drive and retard innovation, it has been an ad- find other familiar features: claims have been vantage for Japanese companieswhich have repeatedly voiced that the Japanese practice di rerent sets of strengths and weaknesses than price discrimination, maintaining high mar- American firms. To begin with, the same half- gins in protected home markets while slashing do: ,,en corporations ,that produce most of Ja-. prices in the United States and Europe. Such pair's ICs account for perhaps tvvo-thirds of de- tactics would imply either explicit or implicit inE;nd; given the focus on vertical integration monopolistic agreement among Japanese man- and internal production, it should be no sur- ufacturerse.g., tacit acceptance of existing prise that U.S. suppliers have had difficulty market positions at home, with price cutting selling in Japan. Second, as the next chapter confined to foreign markets. Even so, questions points out, the quality of Japanese ICs has been of dumping are problematic for integrated highagain, thi might be foreseen. given that firms; companies making ICs for both internal fit-ins producing for internal consumption will consumption and open-market sales have a find themselves bearing high downstream costs good deal of latitude in allocating costs and set- vt:'hen quality lags. A further synergism asso-o ting prices. Dumping, as defined under GATT Cratedith size and diversification stems from rules and the laws of most countries, would be (le ability to tap cash flows generated in other difficult to prove; nor would the usual,4.ation. lines of business; these funds can be channeled ales for prohibiting dumping necessarily be to R&D or added production capacity, matters very relevant. simplified on in chapter 7. Further, diversified MITI's push toward ICs for computers and t."::ompanies can more easily tolerate short-term communications has contributed to Japan's losses resulting from price-led penetration of strength in world markets for memory chips. hew markets. Diversified Japanese companies At the same timeone legacy of the industry's have combined such tactics with an emphasis roots in devices for consumer applications on qualityboth image' and realityto drive Japanese product lines remain more narrow- boldly into markets once the province of ly based than those of the leading American American firms. Indeed, few other strategies suppliers. Microprocessors are a case in point; could have worked. Unfortunately, from the the major Japanese firms all continue to pro- standpoint Of smaller and less diversified U.S. duce American designs. NEC, Toshiba, Mit- merchant manufacturers, unrelenting price subishi, and Oki sell members of the Intel 8080

205 200 International Competitiveness in Electronics family; at least three Japanese firms are build- ly view as critical; U.S. firms heavily depending Intel's 16-bit 8086.4° Although several Japa- ent on memory products have been severely af- nese manufacturers have designed microproc- fected. In other product categories, Japanese essors For internal use, these have not found competition is also stiffening; a major effect other markets. And if made-in-Japan dynamic was to further depress prices and profits dur- RAMS now claim a major share of worldwide ing 1981 and 1982, when domestic firms were sales, the overall Japanese presence in the troubled by a deep recession. United States remains modest. In 1982, imports In the longer term, American semiconduc- from Japan accounted for about 51/2 percent by tor manufacturers have every reason to be value of total U.S. integrated circuit consump- wary of continued pressure from powerful tion; although increasing rapidly (fig. 26, ch. multinationals with headquarters in Japan 4), Japanese imports remain small in absolute firms that have already demonstrated their abil- terms. Still, the inroads have come in a market ity to compete successfully in major world mar- segment that American manufacturers right- kets for other technically demanding products.. The U.S. merchant manufacturers have their "R. I I. Sil in, The Japanese Semiconductor Industry: An Oyer- viet.s' (Hong Kong: Bank of America Asia, Ltd., January 1979), own advantagesthey do well some things that p. 148 "Background of VI,SI War With United States Reviewed," Japanese firms do poorlybut they cannot ex- japan Report, Joint Publications Research Service JPRS L/10662, pect an easy time of it in the future. July 16, 1982, p. 43.

Computers If American manufacturers were for many suppliers in Europe and the United States. It years unchallenged in world markets for semi- is too early to predict the extent to which the conductors. the United States has been still Japanese strategy may succeed, but structural more preeminent in computers. Even in Japan, changes in the world computer industry are American-owned firms continue to account for creating new opportunities for firms every- over 40 percent of mainframe sales; the U.S. where. The Japanese will probably be able to share of the Japanese market for small systems exploit at least some of these, certainly better is lower, but such firms as Data General and than European producers. Hewlett-Packard have recently established production facilities there. In Europe, U.S. com- The Environment for U.S. Suppliers panies are far out front except in'the United Kingdom, where the government has actively By virtually any standard, the United Slates supported ICL through procurements and sub- has far and away the most computer-intensive sidies. American-owned enterprises account economy in the world, a position it can expect for nearly three-quarters of all computer sales to maintain indefinitely. From the early days in Europe. of the industry, the number of computers in -. stalled in the United States mounted at a pace This section again concentrates on the that kept the total about an order of magnitude United States and Japan. While Japanese com- greater than for all of Western Europe.'" By puter manufacturers have not yet proven no- 1981 there was a computer terminal for every tably effective competitors outside their home 48 people employed in the United States; by market, they are at present uniquely situated to launch a campaign aimed at the U.S. posi- tionin part because of their newly acquired strength in microelectronics, in part because "Gaps in Technology: Electronic Computers (Paris: Organiza- of their active. pursuit of joint venture ties with tion for Economic Cooperation and Development, 1969), p. 16.

200 1 Ch. 5Competitiveness in the International Electronics Industry/ 201

1986 there should be 1 for every 10.42 American product lines that match up reasonably, well leadership in design, production, and sales while at the same time staking out their own as tell as utilizationis reflected both in trade territory.e.g., Control Data in high-perform- data, where the computer industry continues ance scientific machines, Burroughs in small to be a prodigious net exporter, and in the business systems (on the latter, see the case prominence of U.S.-owned subsidiaries in study in app. C). In these efforts, American other parts of the world. computer firms have been aided by the technological lead of the U.S. semiconductor indus- Strategic Patterns try. Although IBM has relied heavily on internal semiconductor design and manufacture, For many years the story of American su- other firmswhether or not maintaining cap- premacy in the gtobal computer industry was tive production facilities--have been able to the story of one companyInternational Busi- take advantage of components available on the ness Machines. Although IBM trailed Reming- merchant market that were often superior by ton Rand, builders of the Univac, in marketing conventional yardsticks to IBM's devices. This early computer models, by the late 1950's IBM is one of the reasons IBM has itself begun to had gained the huge lead it still enjoys. For the purchase ICs on the outside. A major element other old-line firmsincluding Burroughs, in the strategies of other mainframe suppliers NCR, Honeywellcompetition has mostly (excluding those making plug-compatible ma- meant jockeying for places in the residual mar- chines) has been to expand into new applica- ket left by IBM; these companies have found tions while tying their installed base to propri- it difficult to reach the scale needed to offer etary softwarethus keeping old customers; a full product line and to support a sales/serv- None have had more than limited success; ice organization competitive with IBM's. In other mainframe-oriented firms have general =, general-purpose mainframes, IBM has ac- ly been a good deal less profitable than IBM, counted for 60 to 70 percent of the world and have made little headway in eroding IBM'S market over the years, with lower figures in market share. Several have done better abroad; such countries as Japan and the United King- Honeywell's joint venture in France has been dom balanced by even higher percentages else- a greater force in the European market than where. To 1- .; sure, numerous entrantsmostly Americanhave attempted to carve out com- the parent has been in the United States. petitive positions against IBM, with much more success in rapidly growing markets for The Impacts of Microelectronics and small systems than in mainframes. Companies Reliability Improvement ranging from Digital Equipment Corp. or Con- Because the market has enlarged and changed trol Data in the United States to Fujitsu in so radically, focusing on the older mainframe Japan and Nixdorf in West Germany have es- companies hardly gives a fair sampling of cur- tablished themselves solidly in some portions rent strategies. As figure 35 indicates, market of the market. But none has come close to growth for general-purpose mainframesthe IBM's overall sales, despite the rapid shifts in mainstay of the industry just a few years ago= overall market structure described in the preis now much slower than for other types of sys- vious chapter. tems. As sales of minicomputers, small business installations, and desktop and personal Most of IBM's American competitors have machines exploded, competitive dynamics taken a straightforward approach to their situa- altered fundamentally. Newer entrants have tion: following IBM's lead in the development staked out major shares of markets for products of faster and larger systems, trying to maintain like word processors. These structural shifts are continuingindeed accelerating. 421- M. Branscomb, "Computer Communications in the Eight- iesTime To Put It All Together," Computer Networks, vol. 5, What are the implications for international 1981, p. 3. competitiveness? As in semiconductors, prod-

99-111 0 - 83 - 14 207 ?02 International Competitiveness in Electronics

Figure 35.Market Segmentation of U.S. Computer Sales by Value

.1 . '

.. '&p,i,ocessqrs

. .

1980 1975 "'

Word processors 10%

business Minicomputers. systerris... 21% . 13%

1985 (projected)

SOURCE: "Moving Away From Mainframes: The Large Computer Makers' Strategy for Survival," BusinessWeek, Feb. 15, 1982, p. 78. 208' Ch. 5 Competitiveness in the International Electronics Industry 203

uct. planning decisions in the computer in- banking or distributed word processing. Start- dustry are shaped by technical possibilities. ups of earlier yearsData General, Prime, Tan- Specialized machines of all sizesave long had demhave become substantial multinationals their place, but the turn toward small and ver- in their own right. Independent software ven- satile computers is a direct consequence of dors are creating a whole new industry. twin driving forces: advances in software for networking and distributedrocessing, plus Better reliabilityin addition to broadening advances in microelectronics. Computer sys- the applications of computershas had a sec- tems need no longer be structured around a ond, equally important, impact. As in con- single processor. A central unit can be sur- sumer electronics,it has allowed ,manufac- rounded by a number of satellites, or the en- turers to skirt traditional distribution channels tire processing load can be shared throughout and reach customers through new outlets. This a network. Given cheap microprocessors Find trendwhich began as early as the 1960's with single-chip microcomputers, designers can put systems houses that purchased minicomputers "intelligence" where they want it. Computer and peripherals in quantity, assembling them, firms that fall behind in such developments together with software, to supply turnkey in- more broadly, manufacturers of systems incor- stallationsalso promises to continue and per- porating machine intelligence, not all of whom haps accelerate. Greater reliability has reduced think of themselves as part of the computer in- the need for onsite maintenance and repair; dustrywill be poorly placed to compete in fu- where field service was once a vital element. ture markets. in any marketing strategy, smaller manufacturers are now less constrained by the need to fi- Improvements in system reliability--flowing nance service networks. Moreoverwhile hob- partially but not wholly from the growing byists, engineers and scientists, and many bus- reliability of microelectronic devicesyield inesses could be reached through specialized another powerful driving force. Mean times be- distribution channelsselling personal or desk- tween failure for computer systems have been top computers to the general public requires increasing steadily despite greater complexi- retail distribution. This, in turn, is realistic only ty. Today's computers are orders ofmagnitude if the need for aftersales service is modest. Cur- more dependable than those of even a decade rently, the personal computer market is mov- ago; this not only cuts operating and maining through'a transition paralleling the earlier tenance costs but helps expand applications. shift in color TVdesktop machines are be- Computers can now be used in a host of ap- coming off-the-shelf items rather than products plications where the dangers of a failure were sold and serviced by specialty outlets. The per- formerly too greatreal-time air traffic control, sonal computer is a product in which IBM had electric, utility load management, critical in- no great advantages beyond name-recognition dustrial processes. and abundant internal resources for product development. While these are far from trivial Ever-greater reliability combined with ever- assets, IBM will probably not be abletO dupli- greater computing power per dollar has eaten cate its position in mainframes in the far more away at IBM's traditional strengths-7-cuatomer diverse and competitive desktop market, just suppOrt and service, plus the ability to lock in as the company has not been able to do as well customers via a product line broad enough to in small business systems, supercomputers, or satisfy virtually every need. Now, so many ap- word processors. The general point is: to be in plications have opened up that no one com- the computer business no longer necessarily pany can hope to cover them all; in many of means confrontation with IBM; it need not en- these, IBM's mcrket powerso valuable in sell- tail attempting to cut into the installed base of ing general-purpose mainframeshas been, if any mainframe manufacturer, much less try- not irrelevant, al least a far smaller advantage. ing to match IBM's hardware or software New entrants can specialiie in systems for across a broad spectrum of products. 204 International Competitiveness i-

.Even in small systems, only a few companies puter goes in large part to the specialized ap- have been able to span a major portion of the plications programs that were available. Even market. Somee.g., Hewlett-Packardhave more, as hardware costs fall, specialized soft- specialized in powerful machines for sophis- ware becomes the pacing factor in applications ticated customers. Others, like Wang and Data- ranging from office automationwhere much point, have aimed at business applications. of the competition for word processor sales Digital Equipment Corp. (DEC)which started revolves around softwareto industrial robots. as an OEM suppliernow has a relatively Limited growth in software productivity and broad product line, including personal com- high associated costs (ch. 3) are problems that puters. As in the semiconductor industry, man- now confront all firms in the industry, here and agers have had to search for market opportuni- overseas; among the possible solutions are iiiul- ties that match their organizations' strengths. tinationalized software generation. In the Their choices and decisions, constrained by re- future, the importance of software compared source lir stations and conditioned by govern- to hardware can only increase; the exceptions ment Ellilicies, will determine the future com- are perhaps at the very high and very low ends, petitive posture of the U.S. industry. Mistakes where supercomputers remain hardware-inten- will be made, and weaker companiesmost sive and small machines selling for less than likely those making peripherals, office automa- a thousand dollars compete on the basis of tion equipment,fand desktop computerswill price. find their..,esibeing absorbed or merged with From a slightly different perspective, soft- competitors. ware can become a constraint: switching to new software, particularly system software, is Product Strategies: Hardware and Software time-consuming and expensive. Customers The mainframe-oriented companies do retain with extensive data processing installations advantages in structuring complex and far- and large software inventories become locked flung inforthation systems. Designs for such in because of the high costs of transferring. system's are often shaped by existing software This is a constraint on the system manufactur- inventories. The original supplier has an easier er as well, who may be burdened with obsoles- time achieVing compatibility; indeed, computer cent software that cuts into potential perform- firmshave had a good deal to gain by making ance. The picture is somewhat different for it difficult for competitors.to reverse-engineer computers sold to purchasers who are techno- their software and develop compatible systems. logically adept---e.g., OEMs who integrate com- Some have-gone so far as to replace portions puters into their own products, or those with of the system software with "firmware" stored needs in engineering or science. Such custom- in ROM chips which can be changed from time ers commonly have the internal resources for tntime. Generally, such efforts have been in- solving their own software problems, and find tended to thwart plug-compatible manufac- shifting to new systems, though a difficult task, tvi ers. not an insurmountable one. Still, given their software investments, virtually all customers iniportanc'e of software extends far behave strong motives for replacing or augment- yond the system level. Machines capable of ing their equipment with new models from the data processing for business needs are now same manufacturerand manufacturers within the financial reach of even the smallest strong motives for ensuring software compati- firmsand, of equal significance, can be.used tlbility within their product lines. Therefore, by people with little special training. Software once markets begin to mature, a manufactur- that is user-friendly, as well as reliable, is a key er's share of the installed base becomes a good element in selling to those without previous indicator of future prospects; competitors need data processing experience. As the case study hardware that is substantially better or cheaper in appendix' C points out, credit for the suc- to stand much chance of convincing customers cess of IBM's System/32 small business corn- to switch allegiance. Brand loyalty has been

21 u Ch. 5Cornpetitiveness in the International Electronics Industry 205

high in general-purpose data processing mar- the technologyalthough occasionally market kets, largely for such reasons. demand outstrips what the industry can sup- Plug-compatibility is aimed at breaking this ply, as happened with word processors and cycle. Originally referring to peripherals such it is difficult to predict where it May lead. Some as disk drives, plug-compatible manufacturers observers ly.11ieve that IBM's market power will (PCMs) later moved into mainframes that can continue to deteriorate, even in areas where .operate on IBM software; now some build the firm's position has heretofore seemed un- equipment compatible with DEC minicom- assailable. Others think the future lies with puters or IBM Personal Computers. Basically, large and powerful companies able to combine the PCM strategy has been to make equipment far-flung communications and information net- that can be used interchangeably with IBM's, works into vast integrated systems. In fact, both views are probably correct, given the fragmen- while undercutting the latter's price/perform- tation and specialization brought by cheap ance ratio5.43 hardware. The forces outlined above shape the strat- egids of companies striving to keep up in the American computer manufacturers, living marketplace. New approaches to product de: nervously with rapid technical change at home, veloprnent are appearing throughout the face another series of choices in foreign mar- dustry; even IBM has begun to purchase more kets. Governments in industrialized nations hardware outside, as well as software. In where American subsidiaries have long been another new departure for the company, IBM dominant continue to follow policies trans-, has 'started selling disk drives on an OEM basis. parently intended to reduce that dominance. As in semiconductors, technology-sharug Such policies are nothing new: France's Plan agreements have become more common Calcul was set forth more than 15 years ago, cross-licensing of patents, direct purchases of and the Governrrients of Great Britain and Ja- technology, joint developmentas firms con- pan have, over the years, found many ways to serve resources through specialization. This is support local computer manufacturers. While the idea behind Microelectronics & Computer most such policies have had only limited ef- Technology Corp.spearheaded by Control fects in the past, certainly in Japan the tech- Data and presumably aimed not only at oncom- nological fervor is now intrmse. ing competition with the Japanese but the con- If competition from Japanese computer firms tinuing struggle of smaller entrants with 113M. is on the rise, American entrants are them. Movement toward technology purchases and selves fashioning new international strategies. technology sharing appears to have even more Already DEC operates six plants in Europe and momentum in Europe, despite earlier failures three more in the Far East. A partial list cif of joint efforts like .Unidata. Siemens, ICL, and other American minicomputer manufacturers Olivetti are among the companies now market- with foreign production facilities would In- ing Japanese-built mainframes in Europe. clude Hewlett-Packard, Wang, Data General, Datapoint, and Texas Instruments. U.S.-based International Aspects multinationals specializing in desktop ma- The picture that emerges in the U.S. com- chines include Apple, with plants in Ireland puter industry is one in which the long-dom- and Singapore, and Tandy. Manufacturers pro- inant leader is being challenged on all sides. ducing plug-compatible mainframeshave also Structural change has been driven largely by begun to expand abroad: Amdahl has opened an Irish facility intended in part to supply the Common Market, as has Trilogy Systems. ',The founder of Amdahl. the leading supplier of PCM mainframes. has said that surviving in competition '9M requires The rules of the competitive game are in par- Costs that are 15 percent lower or performance that is 20 percent better. See "Makeshift Marriage." The Economist, Aug. 11, ticular flux in lesser developed parts of the 1979. p. 78.' world. Developing countries are putting in-

21i 206 International Cornpetit,'Ynness in Electronics dustrial policies to work attracting technology eral industrializing nations are attempting to and foStering local computer manufacturing. improve the capabilities of their labor forces Mexico's appriach has been to restrict lInports, so that they can produce softwareneeded in limiting sales to companies that agree to estab- advanced economies. Countries like Singapore, lish production facilities. With an annual mar- Hong Kong, and Taiwan are seeking to create ke: now approaching $500 million; Mexico has "software centers" where Western computer been able to attract a pair of U.S. minicomputer manufacturers could establish subsidiaries that firms willing to live with these rules. Brazil's would transfer skills and provide training for Government has reserved the domestic mini- the local work force while also producing computer market for locally controlled enter-. much-needed software. Once the people were prises; transfers of technology have been en- available, locally owned companies could take couraged, but foreign investments are limited over at least some of the work. to minority interests. While American companies have generally chosen to stay out, several Japan European and Japanese firms have agreed to _Objectives announced by Japan's Govern- participateno doubt hoping for benefits sim- ment over the past few years herald 'a com- ilar to those now flowing to the Japanese,con- petitive onslaught directed at the U.S. com- SU mer electronics manufacturers that accepted puter industry. MITI is sponsoring a pair of such conditions in earlier years. long-term R&D projects dealing with computer How' will the onset of local production in systems, plus several related efforts." The fifth developing countries affect international com- generation computer projectthe origins and petition in computers? While any answer' reorganization of which are described in chapter mains conjectural, it would be foolish to dis- 10is software-intensive, directed at artificial miss the possibility that some of these nations intelligence, information organization and may evolve. into viable forces -in the market- manag mt, and natural language input and place. Although their ability to compete will t l :e ond project, MITI is helping probably be restricted to simpler products over lune one development of a supercomputer in- ;'.';: foreseeable future, developing economies tended to surpass the most powerful offerings will begin by building equipment such as ter-- of American companies like Control Data and minals, printers, and disk drives, where labor Cray. A related 10-year project will support is a major cost element. It is not a big step from development of the high-speed microelectron- making TV, receivers to producing the simpler ic devices needed to implement the software types of computer terminalsindeed a step concepts of both fifth-generation machines and that countries like Korea and Taiwan have al- supercomputers. The goal is nothing less than ready taken. With the experience. gained in to thrust Japanese companies into the forefront_ such products,. and with protected '.markets of world computer technology, to leapfrog the contributing to scale economies, a number of United States in the design and marketing of the newly industrializing countries could move both hardware and software. The objectives of fairly quiCkly into world markets. As a final point, again consider .software de- "Outline of Research and Development Plans for Fifth Genera- tion Computer Systems (Tokyo: Japan Information Processing velOpment. By its nature, programing hOF', been Development Center. Institute for New Gederation Computer labor-intensivetherefore increasingly costly echnology. May 1982): Computer White Paper: 1981 Edition in high-wage nations. Software generation de- (Tokyo; Japan Information Processing Development Center. 1982). pp. 59-75: "Machinery. Information Industries '81 Pro- pends on people with ability and experience grams Outlined." Japan Report, Joint PubliCations Research Serv- including an understanding of the problems ice JPRS L/10086. Nov. 2. 1981. p. 21: "Archety" e of Fifth Genera- faced by users. Such faCtors have prevented the tion Computer Described." Japan Report. Jo:,, Publications Re- search SerVice JPRS L111007 Dec. 14. 1982. p. "MITI Proj- transfer of this work to developing countries, ect To Develop Supercomputer Starts in January." Japan Report. even those like. India where the raw program - Joint Publications Research Service JPRS L110348. F h. 23. 1982 ing skills might be available. Nonetheless, sev- p. 34. Ch. 5Competitiveness in the international ElectrclnIcs Industry 207 \ these programs are by no means unique to their weaknesses, 4nd have made plans accord- Japanese manufacturers; they are squarely in ingly. the mainstream of the evolution of computing. It is the strength of Japan's commitment the Technology backing by MITI and other government agencies, the I0 -year schedulesthat differentiate Carefully targeted R&D isle central strand in them from efforts in other countries. the Japanese computer strategy,as in earlier ventures into other industries. Japanese pro- As the rhetoric associated'ith such producers and their government realize, justas grams makes clear, Japanese !I ,ms, with the they did in microelectronics, that international" help of their government. hope within 10 or 15 competitiveness in computers cannot be at- years to lead the world in computer technol- tained so long as they rely\\on technology from ogy. Despite Japan's 'alatively successful ex- the United States. The reasons are twofold. perience with previous government-sponsored First, American firms are faress likely to R&D effortsthe Pattern Information Process- license technologies than in past. The Jap- ing System (PIPS) project, the VLSI project---: anese know that computer Jnanufacturers in this is a tall order. At present, the market pcsi- the United States, unlike at leastsome of their tion of Japanese manufacturers is modest; as predecessors in other sectors,are acutely of 1981, American-owned companies held aware that technical leadership is a primary more than three-quarters of the world com- source of competitivij strength, and that to puter market in value. terms, Japanese-owned make 11,eir !nclinology too easily available companies only about 7 percont;Still, Japan is would weaken their own poSition. The second now the second largest supplier of gener,a12 reason is even more fundamental. In the basic purpose computers lo the world market, will building blocks of computer hardware, semi- ,a very high rate. of export growth (ch. 4, fig. conductors, Japanese firms are near parity with 30). The country is also second only to the American companies; in some areas they may United States in intensity.of compuxer utiliza- be ahead. Japan can hardly dependon im- tion. After the experiences of consumer elec ported technology; rapid progress towardan ironies, automobiles, and semiconductor mem- eventual goal of leadership in information / ory chips, few in the American industry would processing requires extensive indigenous capa-/ take Japan's goals lightly. bility of the sort that Japanese firmsnow have Nevertheless, because of ,lie role that factors in high-density memory chips. such as installed base and s )ft ware inventories The Japanese also recognize that their short- play in the marketing of computer systems, comings in the marketplace are not so much Japanese manufacturers must begin with the matters of hardware as of software and related knowledge thatno matter how good their applications-based constraints. Several japa- \ technologythey cannot hOpe to come close nese firms now offer computer hardware as to the United States for many years. In this powerful as any. However, IBM's Mige in- Sense, the computer market is not at alilike that stalled base and vast catalog of applacations fCir semiconductors, where purchasers quick- programs have forced Japanese competitors, ly switch suppliers to take advantage of low as those elsewhere striving to break` into the priCes, quick delivery. or new device types. mainframe market, to build plug-compatible Success in niche markets. for computer systems machines that run on IBM softv4re. To get is quite possible, indeed a necessary first step, around this impediment is perhapS the major but breadth in an industry expanding in as reason for the fifth-generation project. While many different directions as information proc- companies like Amdahl have demonstrated essing can only be a long-term undertaking. that a comfortable business can be builtsup- The U.S. position, both in technology and plying PCM mainframes, markets tied to market share, is simply too strong. Leaders in another manufacturer's software are inevitably Japanese\Government and industry recognize limited. Japan's gamble is that it can jump 21J 208 International Competitiveness in Electronics ahead of American entrants with families of $100 million over roughly the same period.** computer systems having performance capa- A parallel miercelectronics projectwhich bilities that will render present-day software goes' by names such as "R&D on New Func- inventories obsolete. This goal has shaped.the tion Elements"has a budget ef about $150 hardware and software R&D planned for the million and is scheduled to run from 1980 to fifth; - generation project: to take full advantage 1990. Money v,-;t1 go to three major develop- of emerging microelectronics technologies in ment efforts:45 more closely linking the needs and abilities of Three-dimensionalcircuit elements people with the capabilities of the system. If which can be visualized as more-or-less these objectives are met, individuals--even conventional ICs stacked atop one those with little trainingwill be able to com- another, increasing the density. municate withfifth-generationmachines Highclef trc rimobility transistors through ordinary language in spoken or writ- !HE'N,`Tsi, obe \,..iriety of which consists of ten form, as well as through graphical (); very thin layers of semiconducting mate- tonal bnages, Such systetn: would not only rials such as gallium arsenide or gallium be user-friendly, but might,ultimately display aluminum arsenide; HEMTs offer poten- something of the independent decisionmaking tially higher switching speeds, hence faster capability associated with human reasoning.* computers. If the technical objectives of the fifth-gen- Radiation-hardened devices suitable for eration projectand similar efforts in -other use in extreme envirumntsrits such as nu- countriesare achieved, even novice users clear powerplants or outer space (resist- would be able to harness enormous computing ance to heat and vibration is a related ob- power. The commercial potential is immense. jective). Government Assistance The first two especially will support both supercomputer and fifth-generation projects. The Japanese Goverriment has supported Among related government-sponsored pro- R&D activities in information. processing over grams, another of major significance for the Many years. MITI has been selective in finan- corporate strategies of Japan's computer man- cial aid, directing funds to potential bottle- ufacturers has aimed at the development of necks, exemplified by the VLSI program's sup- software and peripheral devices with Japanese port for digital ICs, or to R&D that could help language input-output capability. Scheduled Japan's industries leapfrog the competition, the over the period 1979-83, nearly $200 million intent of the fifth-generation computer project. was allocated to this effort." Funding for the latter is projected at about.$500 million over a 10-year span (1.981-91); the super- As in microelectronics, R&D is but one of computer project is expected to get another many ways in which Japan's Government as- sists the computer industry. The Japan Devel- opment Bank loans money to the Japan Elec- tronic Computer Corp. (JECC), a jointly held firm which purchases computers from partic- An example from the field of artificial intelligencethe area ipating manufacturers and leases them to known as expert systemswill illustrate. Research in expert systems aims at computer programs that mimic attributes of people who are "experts" in some realm of knowledgee.g.. medi- cine. where such programs might help automate diagnosis. The **Planning for the fifth-generation program began several objective would generally be to augment or complement rather years earlier, as outlined in ch. 10. A variety of funding levels than supplant human skills: an expert system would not have for both projects have been reported: spending plans and sched- the judgement of a physician, but could offei, for example. per- ules will no doubt shift as they progress. fect recall of vast amounts of information. Expert systems typ- ""FY82 Government Projects in Electronics Listed." Japan ically depend on complex software and large data bases; thus. Report, Joint Publications Research Service JPRS 1110676, July advances in hardware as well as software may be needed if they 22. 1982. p. 55. are to be widely implemented. "Computer White Paper: 1981 Edition. op. cit., pp. 4ff.

21 Ch. 5.Competitiveness in the International Electronics Industry 209

Table 46.Resoarch and Development Expenditures by Several U.S. Computer Manufacturers, 1981

R&D spending (millions of dollars) Burroughs $220 Control Data Corp. 202 Digital Equipment Corp. 251 Hewlett-Packard 347 Honeywell 369 IBM 1,600 SOURCES: Annual reports

dustries in Japanincluding indirect support

Photo credit: IBM Corp through tax preferencescould only be estimated by making a large number of essential- Memory cells in experimental Josephson junction integrated circuit chip ly arbitrary assumptions; see ch. 10.) MITI's R&D subsidies are also modest in comparison to the research budgets of Japanese companies. Fujitsu spent $260 million on R&D in 1981, users. Manufacturers can set up tax-free re- while Hitachi and NEC spent $610 million and serves to offset losses incurred when lease con- $230 million, respectively.'" The government tracts with ) ECC are canceled and equipment money does have an important function: help- must be repurchased. Since 1979, tax-free re- ing with the kinds of long-term R&D that -in- serves have been permitted for up to half the dividual companies might otherwise have dif- income associated with some categories of soft- ficulty in justifying. In addition, MITI-spon- ware. Purchasers of certain types of computers sored projectsthough not cooperative in the can write off 13 percent of tliFf value, beyond usual senseattempt to stimulate creative normal depreciatiOn, in the first year. The gov- thinking, technology interchange, and the com- ernment has also established special deprecia- plex of synergies so vital to engineering re- tion schedules for high-performance remote search. The Japanese electronics industry prob- data processing equipment. ably benefits more from these factorswhich A panopoly of support measuresof which tend to be lacking within the laboratories of inmany mor- examples could be citedhas thus dividual corporationsthan a strict compari- been desigaed to help Japanese companies son of funding levels would suggest. achieve technological superiority and commer- Of course, ether governments also provide cial success in the 1990's. At first glance, the assistance to their computer industries, not ex- sums of money involved may seem large; in cluding the United StateS. European nations fact, when viewed in the context of the world routinely channel direct financial aid to local computer industry, they are modest; as chapter companies, along withindirect subsidies 10 stresses, it is the consistent support provided through procurement and tax benefits. Hand- by many individual measures acting in concert some incentives designedtoattractin- that gives Japanese industrial policy its impact. vestments and technology have been dangled To place the expenditures of the Japanese before the European subsidiaries of U.S. and Government in perspective, table 46 lists R&D Japanese companies. In the United States, spending by a number of U.S.-based computer funding by the Department of Defense through firms. On an annualized basis, subsidies pro- the Very High-Speed Integrated Circuit' provided by Japan's Government for R&D in in- gram and this country's own supercomputer formation processing come to less than the ex- projectstill in the planning stageswill have penditures of any one of these American companies. (Total subsidies for the information in- "Annualreports.

2 1 J 210 International Competitiveness in Electronics commercial spillovers. The lesson is that no international business; experience compared to t ustrial ized nation has l,,,3en content to accept American firms of even suite modest size. a !au:011(1m.y position in technologies and mar - Until recently, most (but not all) Japanese com- ker m considered essential to future ec;onoinic panies have preferred to manufacture at home deve!opment..The concern,is that the Japanese for export; this could prove a major weakness may bo more successful in implementing their 111 computers. While a few Japanese electronics policies than other countries. companiesToshibais anotherhave expanded aggressively via overseas investment, Marketing Strategies and most of Japan's past international successes Multinational Operations have come in products where integrated man- Individual computer manufacturers in Japan ufacturing and marketing in foreign countries have begun to formulate product strategies has not been essential. The examples include based on technologies expected to flow from steel, automobiles, and semiconductors; con- government-supported R&D projects,, as well sumer electronics is only a partial exception. In each of these cases, Japanese companies, at as their internal activities. Marketing c.om- leaSt in the beginning, concentrated on export puters presents special difficulties because the sales. Generally able to take advantage of es- chief competitors are already well entrenched; tablished distribution systems, they invested only in peripheral equipment such -as ter- overseas only when import restrictions com- :ainals, printers, and disk drives have Japanese pelled local manufacturing. The competitive manufacturers made a significant impact out- pressures that led U.S. semiconductor or com- side their home market. In Europe and the puter firms to invest in Europe and elsewhere United States, where nearly two-thirds of the have on!.y recently begun to impinge on the ,world's computer systems have been installed, Japanese. By now, American computer firms Japanese companies are inconsequential as in- not only operate wholly owned sales and serv- dependent suppliers. ice networks in many parts of the world, they In an industry where salas depend on a thor- have established internationally dispersed and ough grasp of user needs at a technical level integrated manufacturing operationspartly in software as well as hardwarelate entry is a response to governmental demands and part- major handicap. American suppliersinclud- ly due to the nature of the market. Japanese ing newer participants like DEC and Hewlett- firms, on the other hand, have been largely un- Packardhave built networks of sales and serv- willing or unable to make the enormous in- ice centers staffed by engineers and techni- vestments required to participate in the world cians who now have longstanding ties with marketplace for computers. customers; IBM has such a network within Ja- Managers of Japanese firms, along with bu- pan. Even those.Japanese firms with strong in- reaucrats within the government, recognize ternational positions in microelectronics or their lack of background and experience, and telecommunications cannot match the distribu- are seeking remedies. The international (as option systems of U.S. computer manufacturers. posed to R&D) strategy appears to be an in- To make much progress, Japanese entrants will cremental one, geared to minimizing the re- have to invest substantial sums over many sourct, at risk and taking advantage of existing years without the expectation of immediate strengths. As part of this strategy, Japanese returns. The history of fields like consumer electronics firms, with the.encouragement of electronics indicates that at least some lap- MITI, are beginning to establish manufactur- anese companies will be willing to make this ing plants in other industrialized countries. commitment. Following investments by Japanese consumer Perhaps surprisingly, given the global per- electronics suppliers in the United States and spectives of consumer-oriented firms like Mat- elsewhere, tentative steps have been taken in sushita, the Japanese computer industry as a semiconductors, a market in which the Japa- whole suffers from a pronounced lack of in- nese have already become well entrenched Ch. 5Competitiveness in the International Electronics Industry 211

through exports. These same semiconductor minicomputers or small systems. Moreover, manufacturers are of course the major com- the Japanese participants remain a critical step puter firms. Experience gained from invest- removed from the customers whose applica- ments in semiconductor production will help tions their equipment is intended to serve in structuring multinational computer opera- joint ventures will provide limited help at best tions. in remedying past weaknesses of Japanese As a parallel step, Japanese manufacturers firms in software or customer support and have established marketing links with a num- service. To become viable international com- ber of foreign firms. Fujitsu now furnishes petitors, Japan's computer companies will need Siemens (West Germany) and ICL (Britain) with to accumulate experience in dealing directly large mainframes, while Hitachi has similar ar- with the requirements of customers inparkets rangements with BASF (West Germany) and Where they hope to sell. Olivetti (Italy). Fujitsu has taken a minority in- Computer manufacturers in Japan do not terest in SECOINSA of Spain, while agreeing share these problems in equal measure; the in- to a technology transfer tie with a company dustry is far from monolithic. Fujitsu, at the partially owned by the Brazilian Government. moment in a clear leadership position (ch. 4), In the United States, Fujitsu holds a minority has, along with Hitachi, chosen to stake its in- interest in Anidah:, the PCM pioneer to which ternational position on supplying IBM-com- it exports large machines; for several years, Fu- patible equipmentdecisions that will limit jitsu distributed its smaller systems within the both companies' options for many years to United States through a joint venture with come. NEC has taken a different route, devel- TRW. National Advanced Systems, a subSidi- oping its own system software (although deriv- lay of National Semiconductor, sells Hitachi ative of U.S. technology), Nor has. NEC yet computers here. entered into marketing arrangements with for- Such arrangements build from the fact of eign concerns. Instead, the company's manage- Japanese parity in hardware for large com- ment appears to be shaping a strategy intended, puters, parity which does not extend to soft- to take advantage of the overlap and merger war,3; both Fujitsu's and HitaChi's systems are of computer and communications technolo- IBM compatible. European firms have been gies, areas where the company is already prom- unable to attain the economies of scale that inent. Despite its relatively small size compared '.Japanese manufacturers get in their home mar- to other Japanese electronics firms, much less ket, and have chosen to compete with Ameri- IBM or AT&T, Nippon Electric's managers are can producers by importing from Japan. From attempting to position their organization for the collective viewpoint of Japanese firms, what they see as an eventual competitive 'strug- these ventureseven where the equipment is gle with these two American giants for domi- labeled with some other brand-nameincrease nance of the international information indus- market exposure and add to production scale. try.

. For'some time, such relationships will continue At several points above, the entry barriers to be essential elements in the marketing strat- created by the well-established sales and serv- egies of at least several of Japan's computer ice networks of American firms have been de- manufacturers. Even so, they link companies scribed. This aspect of the market for com- none of which has more than a minor share puters effectively turns one of the supposed ad- of the global computer market. Siemens, ICL, vantages of the Japanese system on its head. and the other partners of Japanese firms to- Barriers erected by government to keep out gether do not account for even 5 percent of foreign firms have given Japan's manufacturers world computer sales. Wth the possible excep- advantages in a number of industries, partly tion of ICL, none has a ;Gale of operation and through scale economies. Closed markets cre- distribution approaching that of the competing ated by import restrictions and foreign invest- local subsidiary of IBM. None is strong in ment controls have been reinforced by corn-

2 1 212 . International Competitiveness- in Electronics

plea distribution structures and a deeply in- tion and marketing experience will be able to grained "Buy Japanese" attitude. In computers, exploit the new technologies most effectively. however, the longstanding customer ties main- In this context, present efforts would not be tained b'I ;.S. firms combine with technologi- so important in themselves; rather they would cal strengths to create formidable entry barriers be preparatory steps for rapid growth in the- for Japanese companiesindeed, new entrants 1990's. Another path, one that some firms will from any part of the world. Windows do open certainly pursue, is to concentrate on selling because of technological advance; through smaller systems and personal machines. Here these vindows newcomers have moved into the now-traditional Japanese entry strategy is markets for microcomputers, small business feasible because distribution networks are open systems, and other specialized products. Thus to all comers. Thus far, attempts to challenge far, most -of these entrants have been American American companies like Apple or Tandy in firms -in Fart because the U.S. market is so personal computers, or the many U.S. entrants large, but also bectiuse American companies in the market for small business systems, have control the distribution apparatus in most parts not been notably successfulin the United of the world. The going will be difficult for States or elsewhere. Still, if and when such Japanese manufacturers, although they are be- products become more nearly -standardized ginning to find niche productsdesktop corn-. and interchangeable, Japanese companies paters may be onesuited to their strengths. could expect an easier time. But even if companies based in Japan were to expand into In medium and large systems, Japanese corn-. these .ma;:kets, it is not at all obvious that this panies can choose from a number of alternative would help them in ether types of systems. (or complementary) cotOses of action.One is to continue to build joint relationships with Japanese producers of computers are thus foreign enterprises. As noted above, suth a taking what seem the only paths available in strategy will require, first, deeper involvemen':: their attempts to break into the world market: with -end users by the Japanese participants, independent technology development coupled .and, second, movement into markets in more with joint marketing relationships. That the. parts of the world. If firms such as Burroughs, marketing ties involve firms that are them- Control Data, or Honeywell were to be en- selves weak-and in need of partners is hardly ticedeach has a relatively small but well- surprising, but makes the establishment of a established market sharethe prospects for viable international presence that much more Japanese firms would look a good deal better. difficult. At this point, the Japanese have had The constant pressure of trying to achieve costs only marginal impacts on global markets; at .comparable to IBM's could well force one or home, IBM-Japan remains a formidable com- more American companies to accept such ties. petitor. Whether or not technical developments As an adjunct to joint marketing-ventures, in ,microe.l.ectronics and software will thrust Japanese manufacturers will probably seek Japan into a position nearer the forefront re- other ways of incrementally expanding sales, mains to be seen. If Japan's computer manufac- while awaiting the fruits of the fifth-generation turers do begin to increase their market shares computer project. If Japan succeeds in pioneer- significantly, the most likely victims will be ing a new generation of hardware and soft- smaller competitors first in Europe, then per- ware, companies with multinational produc- haps in the United States. _, Summary and Conclusions

While international competitivenessin any are central. Costs of labor and capital, techno- industrydepends on many factors, the busi- logical resources, government policies, human ness strategies pursued by private corporations resource endowmentsall can, at least in prin-

218 Ch. 5Competitiveness in the Internaii:nal Electronics Industry 213

ciple, be looked on as forces impingingon man- products. But they were in a good positionto agement decisions, as features of the landscape catch up. The first signs of successcame early, for business tactics and strategies. when Japanese manufacturers like Sonycre- While a useful perspective, the strategic view ated new families of transistor radios smaller of competitiveness is nonethelessan imperfect and lighter than those offered by American substitute for more quantitative indicators.Un- firms. The transistor was invented in the fortunately, the swiftness of technical change United States, the first transistor radios also in electronics precludes useful quantitative made here, but Japanese firms pushed their measures. Productivity trends mean little product development efforts vigorously and where the standard products of todaywheth- outstripped their U.S. rivals withina few years. er semiconductors or mainframe computers Now that Japan is in the lead withnew genera- have capabilities that may be orders ofmagni- tions of consumer products it will be difficult tude beyond those of a decade past. Compara- for American or European manufacturersto tive manufacturing costs carry weight insome regain the lost ground. cases, but not where one company can build In computers and semiconductors, Western products exceeding the reach of competitors. Europe came out of the war well ahead of Little meaning attaches to patent statisticsas Japan. The Europeans had good fundamental surrogates for technical ability when incentives technology, but were stymied by small and for acquiring patents vary widelyamong coun- fragmented markets, as wellas by manage- tries and nowhere correlate very closely with ments that had neither the resources nor the qualitative aspects of technology. vision of their counterparts here. Subsidiaries If shifts in international competitivenesscan- of U.S. corporations became the backbone of not be extracted from statistical series, careful the European computer industrythey still examination of business activities can yield in- areand took the lead in microelectronics. In sights into future prespectsas well as past the Japanese market, American Prrn5 could not trends. In semiconductors and computers, not match the'ir accomplishments in Europe be to mention consumer electronics, American cause of the protective policies of Japan's Gov- firmsonce undisputed leaders in technology: ernment. Still; if not dominant, the United as in sales in their home markets and virtlially States wasand remainsa major force in Jap- around the worldface much stronger compet- anese computer sales, particularly for large itive pressures. Foreign enterprises, most13, machines, as well as in some types of semicon- Japanese but also entrants with headquarter; ductor devices. Continued efforts by American in other Far Eastern countries,are selling firmsbacked if necessary by the U.S. Govern- larger volumes of electronics products within mentto participate on equitable terms within the, United States; American corporationsare Japan, whether by exporting or by direct in vest- having a more diificult time in foreign markets. ment, appear vital for maintaining U.S.com- The sources of these shifts are many. By-and- petitiveness in electronics. The Japanese elec- large, they are not due to mistakesor faulty tronics market is large and still expanding strategies by American firms or by the 'U.S. rapidly; it is now more important than Europe. Government. First and foremost, rising foreign Japanese industrial policy has been a more competition flows from continued rebuilding significant source of support in semiconduc- of the electronics industries of Europe and tors and computers than in consumer elec- Japan in the aftermath of World War II. It is tronics. The MITI-sponsored VLSI research not a new phenomenon. By the mid--1050's, programWhile notas important as some when much of the basic reconstruction ofover- Americans have claimeddid help Japanese seas economies was complete, companies in firms master process technokigies forvery Japan and much of Europe found themselves large-scale digital ICs. In standard device still well behind the United States in their abili- families like memory chipswhere the path of ty to design, develop, and produce electronics technological evolution is clear for all to see, .214 International Competitiveness in Electronics and technological success a function more of hard-pressed to keep up. even survive; their painstaking development and detail design managers will face hard choices in allocating than highly creative engineeringthe Japanese limited resources. Few companieseven in- have excelled. While they cannot as yet match cluding the largest, here or in Japanwill be the breadth of American product lines, they able to cover more than a small fraction of will certainly continue to improve their capa- product markets. bilities in circuit design as well as processing. While no one can foretell competitive out- If U.S. semiconductor manufacturers can ex- comes in the world computer industry, it is ob- pect intense competition, they too have their vious for all to see that Japan has made a series advantagesa different set than those of Japa- of explicit decisionsgoing back as far as the nese firms. If American companies continue 1960's and involving both government and in- to capitalize c.ithese strengthsthe best dustryaimed at claiming a major share of trained engineers in the world, quick recogni- sales and applications. Based on past perform- tion and response to market needs, innovations ance in other sectors of electronic.:(he prob- in circuit design, applications of computer- ability of continued expansion by the major aided techniques, specialized products pursued Japanese computer manufacturers is high. Be- with entrepreneurial zealthe United States cause the characteristics of themarket for data should be able to maintain a leadership posi- processing equipment differ from those for tion. Still, American companies will not be able semiconductors, the United States remains in to monopolize world sales as they did a decade a stronger relative position, There is no reason ago. why the United States cannot continue to hold Competitive pressures, evolving technology, an overall lead in both technology andsales. and growing capital intensityalong with the As events. in all three portions of the elec- continuing expansion of captive production by tronics industry demonstrate, competitive posi- integrated firmsare changing the structure of tions in global markets have shifted, more-or- the U.S. semiconductor industry. New struc- less continually over time. Some firms in some tures bring new strategies. Structure is chang- parts of the world rise, others decline. Of those ing in the computer industry as well, driven that decline, a few may eventually revive, by the technology of computing, itself depend- others disappear. No country can expect all its ing heavily on microelectronics. As computing industries to thrive in international competi- power becomes ever cheaper, more and more tion; any nation that trades will be more com- applications become cost effective. These at- petitive in some industries than others, the tract new firms, designing and developing not leaders in competitiveness shifting over time. only peripherals and software, but specialized That U.S. competitiveness has slackened in processorsminicomputers, personal and consumer electronics does not imply that sim- desktop units, business systems. While the ilar events will follow in other sectors. This mainframe is hardly a dinosaur, a "computer" could happen, but there is no reason to expect can now be a great many thingsmany never declines in microelectronics or computers par- envisioned by the designers of the general- alleling those in color TVparticularly so long purpose machines sold two decades ago by a as the technology continues: its rapidevolution small number of companies such as Univac and markets expand at high rates. These are and IBM. Computing power is now cheap and conditions under which American firms have widely dispersed, often invisible to users. As traditirmally prospered. When the pace of distributed processing and data communica- events slows, other sectors of the Nation's tions continue to spread, new firms will try to economy might begin to find themselves far- establish themselves, entering through win- ing better than electronics in interactional dows of technological or market opportunity. a Some of the older firms will find themselves competition. CHAPTER 6 Manufacturing: Quality, Reliability, and Automation

221 Contents

Page Ot.ervicr.s. 217 Qua lit:. and Reliability 219 , and N1easortinienl. 219 ;Ind \lonaging le,. :)uality 221 in:i..0:.fanc,.! of Design 222 Appri.ycli 294 E,.]..dbility of Intgraf,,!,..i 926 . i:eilability of Color TVs 231 Automation 233 anti Flexible :Automation 234 Llectronics NI.anufacture 235 239 :summary and Conclusions 246 Appendix GAQuality and Reliability Comparisons for Integrated Circuits 247

List of Tables

T-Ildt. NO. Page 17. t:auses of Field Service Failures in Color TV Receivers 223 48. Typical Costs of Detecting and Repairing Faulty Components in an Electronic System 229 49. RankingS by Repair Shops of TV Receivers for Quality and Reliability 232 i0. Reasons Given by Japanese Electronics Firms for Automating 236 6A-1. Hewlett-Packard Data on 16K Random Access Memory Circuits 248 6A-2. Quality Levels of Japanese and U.S. Random Access Memory Circuits 249 6A-3. Reliability Levels of Japanese and U.S. Random Access Memory Circuits . 249

List of Figures Figure Pago 36. Data for MOS RAMS Showing Constant Failure Rates per Circuit as Integration Levels Increase, Decreasing Failure Rates per Bit 220 37. Reliability Improvement With Cumulative Production Experience for a Microprocessor 221 38. "Typical Trends in Failures Attributed to Design and Production 223 39. Typical Failure History for Semiconductor Devices 227 40. Testing Sequence for Point-of-Sale Terminals 229 41. Reliability Trend for Analog Integrated Circuits 230 42. AutoMatic Installation of Integrated Circuits Onto Ceramic Substrate 237 43. Average Labor Hours for Assembly of 21-Inch Color. TV Receivers - 238 44. Two Approaches to the Design of Industrial Robots 239 45. Manufacturing Costs for Robots, Hard Automation, and Human Operators as a Function of Production Volume 241 46: Cost Comparison for Human Operator and RcbotAssuming One-for-One Replacement and Two-Shift Operation yi-the. Automobile Industry 241 47. WOrldwide Annual Sales of Robots,-Past and Projected 242 48. Robot Sales in Japan by Application 243 49. Estimated Numbers of..Robots in Use, 1980 244

222 CHAPTER 6 Manufacturing: Quality, Reliability, and Automation

Overview

Assuming comparable productsanda lack sign, and development, as well as marketing of subsidies or other strongexogenous influ- and distribution. While accounting procedures encescosts are a primary determinant of in- vary, such costs are generally treated as international competitiveness, in electronicsas direct expensesi.e., a percentage is added to in any industry. Comparable productsmay not the direct manufacturing cast of each itempro- be identical, and small differences in perform- duced, as for other overhead. Depreciation of ance or specifications can override small dif- plant and equipment is handled thesame way. ferences in costs in the eyes of customers.But Direct manufacturing cost then consists even for military systems, manufacturing costs primarily of parts, materials, and labor. Re- which depend on both the design of the prod- search, design, and development costsare uct and the design of the production system much higher for products suchas computers are almost always a major consideration. or large-scale ICs than for consumer items In electronics, costs are muchmore critical where technical change is slow and incremen- to the successful marketing of some types of tal, major redesigns infrequent. In the produc- products than others. Intense price competition of semiconductor devices and computers, tion in consumer electronicstelevisions, vid- research and development tends to account for eo cassette recorders, stereo equipment "'a considerably greater percentage of costs; de- makes low manufacturing costsa vital com- preciation charges are also likely to be greater petitive weapon. Much the same is true for because new production equipment must be standardized semiconductor products ranging purchased as the technology advances. from discrete transistors to randomaccess But costs are not the only way in whichman- memory chips; price cutting is the rule, costs ufacturing operations affect competitiveness. highly sensitive to the yields of the production Beyond production costswhich dependon process (ch. 3). For other types of semiconduc- wage rates, prices of materials, supplies, and tor devices, low manufacturing costs and low components, capital charges, and related fac- prices are less vital; if only a single firm makes torslie dimensions such as the quality and the a particular integrated circuit (IC)perhaps reliability of the goods produced. Whilemore one that meets unusual or demanding perform- sophisticated purchasers are most interested ance requirements such as a high-speed, high in the quantifiable dimensions of quality and resolution analog-digital converterit will reliability, in markets for consumer products probably set prices to maximize profits, given perceptionswhether or not well foundedin- the lack of competition. Leading-edgecomput- fluence the decisions of prospective customers. er hardware and software falls in much the Along with other qualitative aspects, suchas same category. Even so, electronics firms are appearance, purchasers base their assessments seldom able to establish and maintain techno- of value for money ,on perceptions of quality logical advantages so large that manufacturing and reliability. costs are of little relevance. These attributes bath the reality and the per- -In addition to direct and indirect manufac- ceptiondepend on factors such as engineer- turing costs, prices charged to purchasers ing design, how the pebple in the work force reflect expenses associated with research, de- are trained, organized, and managed, and on

217 99-111 0.- 83 - 15 9 0 0 218 International Competitiveness in Electronics the capabilities, indeed the quality, of the only did this hurt the company's sale: , RCA manufacturing equipment. Automation can itn- also lost some of its dealer base. Automated proi e quality by reducing theprobability of production was at the heart of the company's human error or simply improving the con- effort to improve the quality and reliability of siste ._y of the production process. In other in- its TV line. stances, there will be no effect. In some cases, Despite the importance of direct costs of pro- quality may be degraded; people are better at duction for competitive success in electronics, some jobs than machines, and viceversahu- OTA has not attempted to estimate or compare man skills far exceed those ofmachines for manufacturing costs. Companies guard cost tasks involving pattern recognition, or where data closely. More important, the dynamics of judgments must be made based on partial or shifting cost structures, rather than costs at a imperfect information. Regardless of specifics, given point in time, are central to ch-mging the quality of a firm's products will ultimately competitive fortunes. To some extent these depend on the stress top management places dynamics can be inferred without the need for on quality as a goal of theproduction process. proprietary data. This chapter then focuses on By the end of the 1970's, issues ofproduct aspects of manufacturing such as quality and quality were in the public eye for industries as reliability, plus automated production technol- disparate as nuclear power, automobiles, and ogies. semiconductors. Perceptions were widespread Product quality is treated primarily from a that the quality of American goods had de- hardware perspective: What are the relative clined compared to those from foreign coun- levels of quality in the United States and Japan? tries.1 Some observers speculated that Ameri- (The comparison is limited to these two coun- can firms and Americanlabor had slipped, tries.) How do product design and the applica- others that consumers had become more de- tion of production engineering and quality as- manding and were no longer satisfied byqual- surance techniques affect quality?How do ity standards that had once been acceptablein products fail?3 Less tangible but equally impor- the U.S. market. Either way, a "qualitygap" tant matters of the human element in man- with respect to imports, extending even to com- ufacturing and quality controlincluding ques- modity items such as steel, has frequentlybeen tions of management and organization, as well advanced as a contributing factor in the declin- as the education and trainingof the work force ing international competitiveness of American are also discussed in chapter 8. firms and industries. To take an example from electronics, the reputation of RCA's color TVs had slipped badly by the end of the 1960'5.2 Not 'Much of the material on quality and reliability assembled below is drawn from "Quality and Reliability of Semiconductors and CTVs.: United States v. japan," op. cit. This report is based 'OTA's contractor on quality and reliability notes that about in part on a series of questionnaires and surveys-200 covering 80 percent of the attendees at an April 1980 American Manage- both manufacturers and purchasers of ICs, 60 covering ment Association seminar on Japanese techniques forquality independent TV service shopsplus 42 visits to facilities of U.S. control and productivity improvement felt that the products of and Japanese firms that make ICs or semiconductor manufac- their own firms were surpassed by products from Japan. Those turing equipment. surveyed were at the seminar to learn from the experience of While comparisons between products of American and Japa- Japanese companiesnot a random sample. See J. MihalaskY nese firms were of primary interest, some of those surveyed also and A. B. Mundel, "Quality and Reliability of Semiconductors commented on the West European electronics industry. In gen- and CTVs: United States v. Japan," report No. C972, prepared eral, the feeling was that European firms had been behind both for OTA by Consultant Services Institute, Inc.. under contract American and Japanese manufacturers in the quality and relia- No. 033-1170.0, p. 6. bility of their ICs and TV receivers. While European producers 'R. A, Joseph, "Automation Helps RCA and Zenith Keep Color- may recently have caught up to the United States in the quality TV Leadership in Face Of Imports," Wall Street journal. MaY and reliability of certain types of semiconductor devices, overall 5, 1981, p. 56. they probably still lag both the United States and. Japan,

2.2.1 Ch. 6Manufacturing: Quality, Reliability, and Automation 219

Quality and Reliability

Meanings and Measurement written by the purchaser. Nevertheless, the perceptions and subjective judgments of cus- Quality, meaning fitness for functiontheex- tent to which a product meets the specifica- tomers sell many computers, and ICs are in- tions of its designers and manufacturers, the spected to be sure that logotypes and part expectations of users can be treated subjec- numbers are properly printed andconvey the tively or objectively. Consumers typically make desired image. subjective judgments concerning the quality of Reliability is a measure of continuing fitness competing products. Manufacturers attempt to for function once a product is placed inserv- define quality in terms of parameters thatcan ice. While quality is determined at a single be measured quantitativelye.g., the ability of point in timegenerally the end of the manu- a TV set to receive weak signals. In addition, facturing process or the beginning of service they may adopt rating scales which trained in- lifereliability is measured over time,as a spectors apply to characteristics that are not failure rate or similar parameter. inherently quantitative. Often the indicesare based on comparisons with samples or stand- The most common indicators of reliability ards; an example would be the appearance of are mean time to failure or mean time between the cabinet for a TV setwhether the trim fit failuresthe interval between disabling fail- properly and colors matched, whether the ures averaged over a large number of items, panels were free of waviness, the number of usually in terms of actual hours of operation: visible flaws and blemishes. Failures that average one per million hours can Through measures like these, manufacturers be expressed as a mean time between failures try to satisfy consumers' perceptions of quali- of 106 hours or as a failure rate of 10-6 per ty as well as ensure that their products func- hour. The graphical presentation in figure 36 tion properlyall at a reasonable cost. For shows the number of ICs (from a much larger products like ICs or computers, the quality im- group) expected to fail in 109 (1 billion) hours age that a firm establishes is likely to be more of operation. A billion hours is 114 centuries; nearly consistent with quantitativemeasures such plots are constructed from short-term data than for consumer goods; indeed,some elec- using statistical techniques. A failure rate of tronic components are sold to specifications one per billion hours means that the expected or most likely lifetime for a single item chosen at random is 109 hours. Definitions of reliability based onlyon failures that prevent the product from functioning are straightforward. Measurementcan nonetheless be time-consuming, as wellas pre- senting difficult statistical problems. Still,a light bulb works until it burns outtestinga I large enough sample. of nominally, identical bulbs will yield a statistically validmean time to failure. Partial failure, or gradual degrada- tion in performance, is more difficult toquan- tify. A 10-year-old TV setmay still function, but not' as well as,when new; there are no simple Photo credit: Bell Laboratories measures of "reliability" that apply to such Probes for testing integrated circuit chips phenomena. 223 220 International Competitiveness in Electronics

_Figure 36.Data for MOS RAMs Showing Constant the reliability of an entire computer system Failure Rates per Circuit as Integration Levels depends on both hardware reliability and soft- Increase, Decreasing Failure Rates per Bit ware reliabilityfailures of the first type having physical causes (although ultimately depending on design and manufacturing prac- 100 Failure rates per circuit tices), failures of the second type depending wholly on the design of the software.

Exhaustive engineering efforts are directed 10 at ensuring the reliability of new and complex systems of all types, particularly where failures can be costly or dangerouse.g., airplanes or 1 nuclear powerplants. To improve reliability, designers apply techniques such as failure mode analysisestimating the probabilities of different typeS of failures and attempting to Failure rates per bit minimize the more serious. A common practice is to add redundancy to the system, providing functional alternatives so that the failure 10-2 of one part does not compromise the whole. A wire rope has a great deql of redundancy because one strand, or many strands in a large to-3 enough cable, can break without impairing the 1970 1972 1975 1977 strength significantly. A chain, in contrast, has (256) .(1K) 14K) (16K) Year no redundancy. Complex computer systems (bits) often include redundant processors and other SOURCE T. Goto and N. Manabe.low Japanese Manufacturers Achieve hardware components, as well as fault-tolerant High IC Reliability." Electronics, Mar. 13, 1980, p. 140. software that can reconfigure the system follow:ng hardware failures. In any type of sys- The reliability of computer software presents tem, degraded performance will normally be another type of problem. Software does not preferable to sudden and total failure. For ex- "fail" or "wear out" from physical causes as ample,elertr nniccot 'rel systems for auto- does hardwarealthough the media that store mobile en ,Ines are designed so that component the software may suffer such failures. But soft- failuresperhaps of a sensor or a memory chip ware still needs continual maintenance; com- will not cause the engine to suddenly stop plex programs are altered andupdated,period- running. Instead, the engineers aim for "soft" icallysometimes to correct errors that are failure modes, or "limp-home" capability. caught only after the software has been placed in service, other times to improve performance. While quality and reliability are related, they The reliability of a piece of software then de- are by no means synonomous. Reliability cle7 pends on the frequency of modifications re- pends more heavily on design engineering, quired to correct programing errors that could qualityon control of the manufacturing proc--. cause the system to malfunction," As a result, ess. In general, as experience in making a prod- -J. D. Musa, "The Measurement and Management of Software uct accumulates, levels of quality and reliability Reliability," Proceedings of the IEEE, vol. 68, 1980, p. 1131. More both increase, Note the similarity with yield in- generally, see R. Dunn and R. Ullman, Quality t:ssurance for Computer Software (New York: McGraw-Hill, 1982). New com- creases for semiconductors, as discussed in puter programs tend to have of'the order of one mistake per hun- chapter 3. Figure 37 illustrates the reliability dred lines; some but not all of these will be found before the improvement over time of the Motorola 6800 program is placed in service M. Lipow, "Number of Faults per Line of Code," IEEE Transactions on Software Engineering, vol. microprocessor, a popular 8-bit circuit that has

SE-8, July 1982, p. 437. been in production since 1974. _ 22 Ch. 6Manufacturing: Quality, Reliability, and Automation 221

Figure 37.Reliability Improvement With Separation of responsibility for design, pro- Cumulative Production Experience for a duction, and quality control characterize man- Microprocessor (Motorola 6800) ufacturing enterprises all over the world, but perhaps more so in the United States than elsewhere (e.g., in Japan). One reason for the prevalence here appears to be the heritage of scientific management, an approach to job methods and the organization of production originating in the work of an American engineer, Freder- ick Taylor, during the early part of the century.6

10 Production engineering includes all the technical aspects of the manufacturing process: plant layout, process design, work methods, se- 1974 :1976 1976 1977 1978 1979 lection of equipment, quality assu-ance. In Year larger firms some of these functions may not SOUPCE D Projeciing %%SU s Impact on Microprocessors, IEEE only fall in different departments, but be fur- .47 1919 p 38 ther subdivided. Still, regardless of organization charts that isolate the design, manufactur- Statistical methods can be applied to quali- ing, and quality functions from one another, ty and reliability problems during both the these activities are closely related functionally.? design and manufacturing stages, but the spe- Product design affects the choice of manufac- cialized discipline of statistical quality control turing technology. The equipment that a firm is largely a tool of the production process.5 As has on hand, together with the costs of invest- an example, defects in ICs can be monitored ing in new equipment, can severely constrain over time to give insight into the effects of proc- the design of its products. Inspection and test- essing variables. In contrast, reliability analysis ing, quality and reliability, depend not only on techniques are applied, not to process variables the choice of manufacturing technologies, but but to tests conducted on finished products and on the overall control of the process. Applica- to field service experience. Steps take: to im- tion of statistical quality control techniques to prove quality sometimes but not always im- individual steps in manufacturing may be prove reliability. straightforward, but overall integration and control of a complex production process is Organizing and Managing for Quality much more than the sum of control cf the individual steps. Managing the interface between design en- Although design and production are inher- gineering and manufacturing engineering pre- ently interdependent, in some cases even sim- sents a classic set of problems that affect pro- ple communication is lost. Stories of design duction costs, as well as quality and reliability. and production supervisors who are not on Designers specify the characteristics of prod- speaking termsor the commonplace of the de- ucts in great detail, while manufacturing en- sign group "tossing the drawings over the gineers must determine how lo make the product so that it will have those characteristics. "See. in particular. Quality Control Handbook op. cit., sec. Sometimes the same people are involved in 48 on "Quality Control and the National Culture," which points out that the sharp divisions of responsibility typical of larger both functions, but more commonly the re- organizations in the United Statese.g., separate departments sponsibilities fall on different parts of an or- for quality control or inspactioncreate reservoirs of specialized ganization. expertise, but at the same time may hinder the widespread application of this expertise. Scientific management is discussed in more detail in ch. 8. 'See. for example. J. M. Juran, F. M. Gryna, Jr., and R. S. Bing- 'J. A. Alic, "Manufacturing Management: Effects on Produc- ham, Jr. (eds.), Quality Control Handbook 3d ed. (New York: tivity and Quality," Efficiency of Manufacturing Systems (New McGraw-Hill. 1974), especially secs. 22-27. York: Plenum Publishing Corp., 1983), p. 281.

227 222 International Competitiveness in Electronics

wall" to the manufacturing departmentare cussed in more detail belowwhere prevention rife. There is at least anecdotal evidence that of defects is emphasized more strongly than de- foreign firms may handle, not only the prob- tection through inspectiun. lems of training design and manufacturing per- One reason for the lo.. status of manufactur- sonnel, but of managing the interface between ing in the United States is simply the low priori- design and production, better than many ty that industry places on it, as indicated by American companies. One approach is to make low pay scales in manufacturing relative to the same individuals or groups responsible for other parts of the firm; engineers employed in both design and production, or at least extend manufacturing and quality control get salaries management responsibility for integrating de- near the bottom of the range for their age and sign and manufacturing farther down into the experience groups at all points during their organizatior.al structure.8 In Japan, for exam- careers; engineers doing administrative work ple, companies often rotate design engineers earn 50 percent more than those involved in through production departments early in their production." careers.8 Not only do Japanese electronics firms tend to stress integration of product and Another indication of lack of attention to process design within their organizations, but manufacturing is that only 4 percent of grad- they frequently involVe vendors, distributors, uates of engineering technology programs in and customers in the wc r;s of their manufactur- the United States specialize in the "manufac- ing er gineers. turing, quality control, industrial" category." Engineering technology is a relatively new field V..ie some American firms have grappled intended to provide practically oriented train- with such problems more successfully than ing meeting the needs of industry (see ch. 8 for others, companies here begin with a fundamen- further discussion of technology education), tal handicap: low, prestige and low pay tend to this it is particularly surprising that such a be associated with white-collar jobs in manu- small fraction of graduates are oriented toward facturing relative to other categories of engi- careers in manufacturing. In engineering pro- neering or management; the best people are sel- grams, so few U.S. graduates receive degrees dom attracted to such jobs. Manufacturing car- in manufacturing that they are not separately ries higher status in European or Japanese cor- thulated. Although students who have studied porations. And, on the manufacturing side of mechanical or industrial engineering often find an American firm, quality control tends to be manufacturing jobs, many programs in these althe bottom of the pecking order. Too often fields have dropped the once common required it seems that manufacturing managers see qual- courses in such topics as manufacturing pipc- ity control only as an obstacle to production. esses and plant layout. American management has -been criticized for overemphasizing the costs of quality, The Importance of Design whereas some quality control professionals argue that a comprehensive program for de- Figure 38 contrasts schematically the effects , signing andbuilding quality (and reliability) of design and manufacturing on reliability. Re- into a product at all stages can save money. liability tends to improve with production ex- Again, there seems to be a contrast with the perience, but failures stemming from design typical attitude in Japanese companiesdis- weaknesses sometimes show up only after long periods in service, hence may even increase °R. E. Cole, "The Japanese Lesson in Quality," Technology over time. Such behavior is typical of many Review, July 1981, p. 29. See also "Sources of Japan's Interna- tional Competitiveness in the Consumer Electronics Industry: An Examination of Selected Issues," report prepared for OTA 10R. Connolly, "Career Outlook," Electronics, June 16, 1981, by Developing World Industry and Technology, Inc., under con- p. 266. tract No. 033-1010.0, pp. 103-104. 1,P. J. Sheridan, "Eugineen es and Technology Degrees, 1982," °J. M. Juran, "Japanese and Western QualityA Contrast," Engineering Education, April 1983, p. 715. The percentage is Quality Progress, December 1978, p. 10. the total for associate and bachelor levels.

2 2 h Ch. 6Manufacturing: Quality, Reliability, and Automation 223

Figure 38.Typical Trends in Failures Attributed as saying, "At Warwick, much of the design to Design and Production work happened after the product was introduced. We relied on field failure information to tell us where we had a problem."12 According to the estimates in table 47, about half the failures in TVs are due to defective Attlbuted components. Some of these may be purchased, to design others manufactured internally---some components fail because they themselves suffer from design problems. Many of the components in a television receiver are transistors or ICs. As illustrated by figure 36, failure rates per chip Attributed tend to remain about the same as circuit den- to manufacturing sity increases. If so, going to higher levels of integration and increasing the number of circuit functions per chip will have two impor- Time from product introduction tant consequences.First, it willcut assembly SOURCE Quat,%' and Re/43bil,ty. of Semiconductors and CTVs. United States Ja;,an ,epc1 No C972. prepared for ON= by Consultant Sentces In. costs because the total number of components 51,tute, Inc, under contract No 03711700. p. 18 will decrease. Second, there will he fewer components to fail, hence reliability should improve. The cost and quality/reliability advan- types of manufactured products, not just elec- tages of chassis designs based on fewer but tronics. more complex ICs have led to rapid reductions Table 47 indicates the extent to which the re- in the numbers of components in TV receivers. liability of color TVs depends on design as In 1977, Zenith's 25-inch color TV contained compared to production. According to the 685 components. Less than 2 years later, the table, .a greater percentage of service failures number had been reduced to 441.'3 have their sources in the design and develop- As part of their strategic thrust into the U.S. ment process than in assembly. One of the rea- market, Japanese consumer electronics firms sons that Japanese TVs achieved better relia- set nut to create an image of high-quality, re- bility than American-made sets during the liable products (ch. 5). They needed trouble-free 197 ',Af appears to have been more conservative products in reality as well as appearance in des gn practice. For example, Japanese sets order to exploit the distribution channels avail- were designed to draw less power. Picture able to them. Reductions in parts counts were tubes operated at lower voltages, with some one of the techniques adopted. Likewise, by the sacrifice to picture quality but lower internal end of the 1970's Japanese semiconductor tempe..atures and less stress on components. products had attained enviable reputations for In some contrast, the vice president for engineering of an American TV manufacturer, now taken over by a Japanese firm, has been quoted 12-American Manufacturers Strive for QualityJapanese Style."Business WeekMar. 12, 1979. p. 32B. "Ibid. Over roughly the same time period, the Japanese-owned Quasar firm reduced its parts count from 516 to 406. while Table 47.Causes of Field Service Failuresin Toshiba claims a 60-percent deCrease in parts count between Color TV Deceivers 1971 and 1979. Other Japanese firms have reported similar re-_ ductions. typically coming earlier than for AmericanTVman, Percentage of ufacturers. For example, the number of parts in a particular Pan- AttributiOn field failures asonic color TV model went from 1023 in 1972 to 488 in 1976-L- Design (and development) 20-40% see "Quality and Reliability of Semiconductors and CTVs: United Quality of components 40.65% States v. Japan." op. cit., p. 47. Japanese TV manufacturers often Final assembly 15-20% pursued simpler chassis designs in parallel with the develop- SOURCE: J. M Juran, "Japanese and Western QualityA Contrast,- Quality ment of automated production facilities, as discussed later in Progress. December 1978, p. 10. the chapter;

22ad 224 International Competitiveness in Electronics quality and reliability. But manufacturers in Scientists and Engineers (JUSE), which helped Japan have not relied on design improvements to locate foreign expertise in quality and reli- alone; emplorees of the large, integrated Japa- ability, and diffused this knowiedge through nese electronics companies tend tohave con- publications, training courses, and confer- siderably more training in quality control and ences. As many as 10 million workers may now production technologies than their counter- have passed through JUSE training courses." parts in the United States. During the 1950's, well-known Americans such as W. E. Deming and J. M. Juran traveled The Japanese Approach and lectured extensively in japan; Juran, in par- Managements of Japanese electronics firms Acular, is credited with much of the visibility profess to believe that improvements in quali- that quality control now enjoys at upper man- ty and reliability will automatically cut costs agement levels in Japanese companies. In and increase productivity, as well as aiding many respects, the quality control movement their marketing strategies. Tliv rhetoric ema- in Japan began at the top and r-pread down- nating from top managers in japan emphasizes wardin considerable contrast to the situation quality to a greater extent than statements by in the United States, where the principal ad- American executives. More concretely, Japa- vocates of quality assurance have often been nese manufacturing companies rely much lower level technical specialists. The well- more heavily on line managers for quality as- known Deming Prizesestablished in 1951, surance, rather than the staff specialists com- and given to both companies and individuals mon in American firms. for achievements in quality controlillustrate the prestige of such activities; they are among Despite this and other organizational differ- the most coveted industrial awards in Japan." ences, most of the methods that Japanese manufacturers use in pursuit of quality and reliabili- Japanese executives like Hajime Karatsu, ty have been borrowed from the United States, Managing Director of Matsushita Communica- just as for product technologies. Japanese in- tion Industries, have been quality control ad- dustrialists have been noted for their study mis- vocates for years; the Reliability Center for sions to visit foreign companies and research Electronic Components of Japan was formed laboratories. Engineers and managers from Ja- in the early 1970's at the urging of industry pan have become skilled at picking out useful leaders, including Karatsu. Financed private- ideas from such visitswhether related to ly by more than 200 electronics firms, the cen- product technologies or to aspects of manufac- ter conducts tests on components and systems, turing such as quality controland improving establishes procedures for determining relia- on them. The theory and, practice of quality as- bility, drafts specifications, and diffuses infor- surance may have diverged more in the United mation on quality improvement within the in- States than in Japan. dustry." The Japanese emphasis on line responsibili- Origins of Qualify Consciousness ty has led to extensive training programs for-- Stress on quality and reliability within Japa-assembly workers and foremen. Efforts/to nese manufacturing firms goes back at least toreach the latter have included radio and TV the period of postwar reconstruction." Man- agers realized that Japan's exports were wide- is"American Manufacturers Strive for QualityJapanese ly viewed as cheap and shoddy. Much of the Style," op. cit. leQuality Control Handbook op.cit.. sec. 48, p. 48.9. On the early effort toward improving Japanese prod- prominence of the Deming Prizes, see U. C. Lehner, "Japanese ucts was orchestrated by the Union of Japanese Firms' Stress on Quality Control Is Reflected in Dogged Vying for Award,"Wall Street Journal,Sept. 24. 1980. p. 52. There is 14"Quality and Reliability of Semiconductors and CTVs: United even a widely publicized "Quality Month" in japan. States v. Japan," op. cit., pp. 38-40. The historical material that 11"Guide to RE]," Reliability Center for Electronic Compo- follows is drawn Largely from this report. nents. Ch. 6Manufacturing: Quality, Reliability, and Automation 225 broadcasts; about 100,000 of the accompany- to the United States. Companies in Japan have ing textbooks were sold in the first year (1956) often dispensed with some fraction of in-proc- of the radio series alone. A monthly magazine ess inspectors, making each worker responsi- Gemba -to -QC (QC for the Foreman), was estab- ble for accepting or rejecting the parts passing lished about the same time and evidently through his or her station. This is but one ex- served as a breeding ground for quality cir- ample of the diffusion of responsibility through clesa technique that has recently receiveda the organization. It is effective in part be- great deal of attention in the United States (see causeat least in the larger companiesem- ch. 8). The first quality circle was registered ployees are carefully selected even for un- with JUSE in 1962; within 15 years, member- skilled, entry-level jobs. Transfers of blue-collar ships in registered quality circles had grown employees within the firm are commona beyond 800,000. JUSE reports that about practice facilitated by unions organized on a 100,000 circles are now in operation, with companywide rather than craft basis, and new- about 80 percent of the nation's blue-collar ly hired workers, or those transferred toan un- ork force involved.18 familiar job, typically pass through training programs that are lengthy compared to those Standards in the United States. At Matsushita, for in- In the United States, product standards tend stance, new assembly line workers are given to be voluntary, but Japan's Industrial Stand- a month of trainingwith a week devoted to ardization Law, passed in 1949, places the re- quality controlbefore they begin to work on sponsibility with government. The law deals the line." In the United States, new assembly explicitly with quality and provides that all line workers would typically get a few minutes Japanese exports must carry the approval of the informal instruction by a foreman, who would Japanese Institute of Standards (NS)." Con- then monitor their performance as they learned sumers in Japan are also said to look for the by doing. Both approaches have their advan- JIS mark. In 1957, the Japanese Government tages. took a further step aimed at- upgrading-the An apparent paradox has developed in the quality image of the country's products, pass- wake of the 30-year history of quality control ing the Export Inspection Law. This regulation activities in Japan outlined above. Many of the created set of standards aimed original techniques imported from the United mostly at smaller companies, and also provided States were concerned with statistical quality for the establishment of testing laboratories. controla subject in which Deming and Juran were authorities. Yet there is little evidence that Organizing for Quality the application of statistical teglniqUes to quali- Despite the visibility of quality circles, they ty or reliability has advanced, any further in are only one tool among the many that Japa- Japan than elsewhere. In fact, applications of nese electronics firms have adopted. Because statistics are seldom mentioned in descriptions training in quality is widespread, and respon- of the quality control procedures of Japanese sibility for quality assurance diffused within electronics firms. Rather, the Japanese appear the organization, quality control departments to have focused their efforts on making individ- in Japanese firms tend to be small compared ual employees aware ofand committed to the achievement of quality. Statistical quality la"Quality and Reliability of Semiconductors and CTVs: United States v. Japan," op. cit., p. 60. Circles also enroll clerical and control is no more than a small part of the management personnel. It has been claimed that the average quality circle in ;apan saves an employer about $100,000 per ""Quality and Reliability of Semiconductors and CTVs: United year. States v. Japan," op, clt., p. 52. While three-quarters of the work- p. 66. A numher of other Asian nations have followed ers in Japanese electronics firms were classed as unskilled at the Japanese example in trying to improve the quality image of the end of the 1970's, the proportion of skilled as compared to their exports. In Taiwan, a small tax is levied to cover the cost unskilled employees is expected to rise rapidly as automation of inspection: the tax drops as quality levels go up. See "Amer- proceeds. Presumably this is an important motivation for the ican Manufacturers Strive for QualityJapanese Style," op. cit. training programs found in many companies.

2q 226 International Competitiveness in Electronics quality programs of typical Japanese electronics manufacturers, which the companies themselves often refer to as "companywide quality control." Intangibles and consciousness-raising are at least as important.

Quality and Reliability of 1111 -7-1'..J- gin Integrated Circuits Manufacturing and Testing Chapter 3 outlined the steps in making ICs. Most of the larger American merchant firms perform some but not all of these in domestic plants, with labor-intensive operations carried out offshore. A typical pattern might be as follows: 2.11._ Operations performed inthe United States: 1. Silicon crystals, generally purchased from outside vendors, are sliced into Photo credit: GenRed. Inc wafers and prepared for lithographic processing. Test equipment fir integrated circuits 2. Wafer fabrication processes such as lithographic patterning, oxidation, etching, 8. The packaged ICs are subjected tofunc- diffusion of dopants, metallization, and tional tests. annealing are carried out; some of these may be highly automated. Often the circuits are shipped back to the 3. Each of the hundreds of ICs (chips) on United States for the final testing in step 8, par- a wafer is tested; thosethat fail are ticularly if destined for American rather than marked, typically with an ink drop. third-country markets. (Economic aspects of Operations often performed in offshore fa- offshore assembly are outlined in app. B.) cilities: Outcomes at all these stages in processing 4. Individual circuits are separated from purity of the silicon crystal, wafer flatness, lith- the wafer, and the defective chips de- ographic precision, integrity of wire bonds, tected in step 3 discarded. hermetic sealing can affect quality and relia- 5. Each good chip is mounted on a sub- bility. Some are more important than others; strate (chip carrier). patterning flaws and poor wire bonds are 6. Lead wires are bonded to pads on the among the most common causesof failures.22 chip (the lead wires connect to external During the manufacturing process, inspection pins, which plug into sockets installed "For a discussion of failure modes in semiconductordevices, on circuit boards). see E. A. Doyle, Jr., "How Pails Fail," IEEESpectres; October 7. The chip is encapsulated in a metal, 1981, p. 36. An important technique, particularlyfor, 'Suring plastic, or ceramic package that pro- reliability, is the analysis of failures. ICs that fail during ting or in service can be examined by a varietyof methodse., vides mechanical and environmental rect observation in a scanning electron microscopeand protection (metal and ceramic packages causes of failure diagnosed. Corrective action,which might rave are normally heTmeticallysealed),21 from a modified circuit resign to simple adjustmenxs In process parameters such as temperature, can then be taken. A compre- "For a more complete description of packaging and assembly, hensive treatment of reliability, emphasizing t.si importaine of see A. B. Glaser and G. E. Subak-Sharpe, Integrated Circuit En- the design of the circuit, is C. G. Poattie, et al., "Elemensof gineerirg: Design, Fabrication, and Applications (Reading, Semiconductor-Device Reliability," Proceedings of the IE??E,vol. Mars.: Addison-Wesley. 1977). ch. 10. 62. 1974. p. 149. Ch. 6Manufacturing: Quality, Reliability, and Automation 227

and testing are possible atsome points but not production process sufficiently well controlled others; in the absence of good methods for di- that testing becomes simply a verification of rect testing following-a-particular processing that control. step, the engineers must rely on control, of process parameters based on downstream test Because of these varying customer demands, results. the electronic components industry has, since well before the semiconductorera, supplied Customer Requirements products to a range of quality and reliability specifications; as many as half a dozen levels Differing customer demands lead toa range developed from the initial distinction between of standards for the quality and reliability of military and commercial pmts. The lowest semiconductor devices. Purchase agreements level has been for inexpensiveconsumer prod- often specify the testing procedures to be fol- ucts such as toys and games, ,the'-,highest for lowed. Military circuits must meet especially applications such as communications satellites. demanding .'specifications for resistanceto Failure rates for the most reliable devicescan- shock, vibration, and severe environments (in- be more than a factor of a hundred below those cluding radiatirm); reliability is emphasized for for the least reliable. satellite applications. Limited-volume markets for partsintended for military orspace applica- Failuresin Semiconductors tions are often served by small firms specializing in,ultrahighleliability components. While The time history of failures fora large pop- semiconductors for commercial markets have ulation of semiconductor devicesas forman- ,seldom faced specificationsas demanding as ufactured products of many kindswillnor- for military and space applications, the actual mally follow a pattern like that in figure 39. functional requirementsparticularly for Ion- ,Early in life, the failure rate tends to be high, sgevitymay be at least as severe. For instance, with most of the failures caused by random 'some computers operate virtually continuously manufacturing defects. The distinctions be- for years, albeit in a service environment that tween quality and reliability become rather r:r- is well controlled and benign; semiconductor bitrary during these early stages. A strict quali- devices for automobiles must function relia- ty- standard, for example, might weed ots'4: parts blyalso over many yearsin an environment that would otherwise fail during the infant Characterized by vibration and extreme temperatures, as well as eXposure,to gasoline, oil Figure 39.Typical Failure History and grease, rain, road salt, and do-it-yourself for Semiconductor Devices repair efforts. Reliability estimatione.g., by accelerated life testingis costly, thus life testing of devices intended for consumer products is minimal. Considerably more reliability testing is carried out on parts destined for computers or communications systems. Because the service record of their products is critical for future sales, and because the costs of locating and replacing faulty parts are high, particularly after the system has gone into service, manufacturers of complex electronic systems demand reliable components. This is one of the reasons firms like IBM or Western Electric chose to build many of their own ICs. Regardless of applica- Time tion, however, the chip Manufacturer seeks a SOURCE: Ottice of Technology Assessment.

233 228 International Competitiveness in Electronics mortality perioci. "Burn-in" tests help detect where other parts, are less likely to fail than the infant mortality failures; during burn-in, ICs ICs. are cycled to high temperatures and exercised Semiconductor products are normally by computerized tasting equipment.23 screened and purchased to an acceptable quali- Afte: the high failure rates early in life, fail- ty level (AQL), a procedure much less expen- ure frequency usually declines to a nominally sive than 100 percent testing. From the stand- , constant valuethe middle portion of the curve point of the purchaser, the AQL is the permissi- in figure 39. For Semiconductor devices, this ble fraction of delivered parts that can be de- period typically spans hundreds of thousands, fectivei.e., that escape detection during in- even millions of hours, during which the prob- spection and screening. A 1 percent AQL ability of failure is extremely low. Eventually, means that no more than one defective circuit the curve may turn up again as devices "wear out of every 100 is allowed, on the average, in outor otherwise deteriorate with age. an acceptable lot; statistical samplingmethods are tailored to this requirement. While semiconductor products do-not wear mechanically, they are susceptible to degrada- Figure 40 outlines the testing program tion from environmental exposure, thermal cy- adopted by a manufacturer of point-of-sale ter- cling, and a variety of physical processes. Com- minals for purchased ICs. Tests are conducted mon causes of long-term failures in ICs in- at many points prior to shipment because clude: loss of hermetic seal, with consequent downstream failures cost much more to find damage from moisture or other environmen- and fix. Costs are even higher for field fail- tal agents; thermal fatigue of the bond between uresboth the direct expenses of warranty the chip and its substrate or of the lead wire repairs and the possible.costs in terms of dam- bonds; gradual thinning and cracking of me- age to the firm's reputation. Table 48 illustrates tallized layer due to electromigration associ- the growth in costs of locating and repairing ated with high current densities (even though faulty components at successively later stages. the currents in ICs are low, the small conduct- The indirect and intangible costs can be much ing paths result in high values of current den- greater than the direct expenses. sity). Failure probabilities associated with particular mechanisms can be reduced by conservative design at both device and system levels, Testing and Screening in Japan a common tactic in applications such assatel- lites or submarine cables. When first qualifying a new vendor, Japanese"- purchasers normally test all incoming parts. Testing With satisfactory experience, statistical sampling replaces 100 percent testing. If the defect Testing costs for ICs increase with levels of fraction remains below 0.01 percent (100 de- integration. Although 100 percent testing is fects per million parts) and downstream fail- common during the early steps in fabrication, ures are rare, the purchaser may stop screen- manufacturers normally screen their final out- ing. Even when purchaser, and supplier are put by random sampling; that is, only a small unrelated firms, customers prefer to depend on fraction of the outgoing product is subjected their suppliers to guarantee quality levels. Jap- to a full battery of tests. Many customers do anese manufacturers do tend to rely rather ex- their own screening of incoming parts. On the tensively on in-process testing, aging, and burn- other hand, a toy manufacturer may not test inin part to minimize infant mortality fail- incoming chips at all, cutting costs by relying ures. on returns and complaints from thefield to locate problems. Such an approach is favored Such practices differ from the arms-length relationships common in the United States. "Eleven percent of nearly 20,000 ICs tested for the 1977 Pio- neer mission to Venus were rejected, many of these tests involv- Perhaps because the major Japanese manufac- ing burn-in periods of 100 to 200 hours. The very high reject turers of semiconductors are also the major rate reflects the severity of the application. See "Quality and Re- users, they often appear to take the attitude that liability of Semiconductors and CTVs: United States v. Japan," op. cit., p 14. the objective of quality control is to deliver 2-3 On. 6Manufacturing: Quality, Reliability, and Automation 229.

Figure 40.Testing Sequence for Table 48.Typical Costs of Detecting and Repairing Point-ofSale Terminals Faulty Components in an Electronic System 100% or selective, Point of detection Direct cost Intangible cost depending Device level Cost of device, ifMinimized If more on the part not refunded by devices than needed manufacturer. are purchased Initially. Circuit board level Manufacturing process Selective dislocated. System level $50 Shipment may be delayed, disrupted. In the field $500 Customer upset. SOURCE: "Calls Volume Key to Testing Decision," Electronic News, Feb. 18, polit;iioldprivIsuiC 1980, supplement p. 20. (after Ootitttriiigd".-100% inspection '!-a890041Y).;!...... Quality and Reliability Comparisons

1000/.. Although respective quality levels of ICs made in the United States and Japan have been debated for several/years, there is little concrete 100% - ATE or data bearing directly on this matter. For a valid special test - equV- rent depending on comparison, circuits from U.S. and Japanese cost trade-off firms should be tested under the following conditions: 1. The devices should be the same type and 100% of similar designe.g., 4K dynamic RAMs, 8080 microprocessors. 2. Test procedures should be identical, the tests conducted at about the same time. (It 100% is not possible to compare quality or reliability at the present moment. Quality comparisons always refer to some point in the past. And, while the most recent re- 130% sults are always desirable, quality and reliability are dynamic characteristics; they fluctuate with the vagaries of the manufacturing process.) 100% 3. The ICs should be produced to the same purChase specifications in terms of AQL or other quality requirements, ideally for 100% delivery to the same customer. While it is no surprise that little of this kind of data has been made public, the unfortunate RATEAutomatic Test Equipment. consequence was a series of public relation SOURCE: Adapted from R. Fleishman, R. J. Lever, and R. N. Parente. "Total Testing," Circuits Manufacturing, November 1979, p. 32. ploys obscuring the technical questions: Were there real differences in quality? If so; what parts that meet their own in-house standards. were the reasons? A more common attitude in the United States By any measure, semiconductor quality and has been that parts need only meet the cus- reliability have improved immensely over the tomer's requirements; customers that demand years, regardless of whether the devices have high quality may get it, others receive less been produced in the United States, Japan, or attention. Europe. As an example; figure41shows de-

2 3 5 230 International Competitiveness in Electronics

Figure 41.Reliability Trend for been made public on quality levels attained by Analog Integrated Circuits captive producers such as Western Electric or IBM. Captives account for about 40 percent of

"..=, 1,000 1974 all ICs made in the United States (ch. 4); the 1975 1976 1977 quality and reliability attained by captive pro- 1978 ducers would, if available, be a useful indicator of the relative technological capability of the American industry. 100 1979 The ranges in quality level included in ap- 1980 pendix 6A, particularly table 6A-2, show a re- markcble lack of consistency on the part of all vendors. Even the top 16K RAM suppliers ex- 10 hibited a factor of five difference between best Cumulative units produced and worst lots. Much larger spreads wc. re the SOUNCL. G. Peattie. "Quality Control tor. ICs." IEEE Spectrum, October 1981, rule, particularly for the American firms: This p 93 illustrates the danger in generalizing from limited samples of IC quality data. It also indicates creases in failure rates for analog (linear) ICs the importance of a consistent and well-con- as used in consumer electronic products. Other trolled manufacturing process, and the diffi: types of ICs show similar trends. Nonetheless, culty of maintaining that control. sources in the American electronics industry both manufacturers and purchasers of semi- Spokesmen for the U.S. semiconductor in- conductorsagree that, during the mid to late dustry have sometimes claimed that Japanese 1970's, quality levels delivered by Japanese firms create a false image of higher quality by firms were superior to those delivered by U.S. sorting ICs and sending only the best to impor- firms. There is also broad agreement that quali- tant customers like Hewlett-Packarda prac- ty levels delivered by American semiconduc- tice that has been termed "quality targeting" tor firms have greatly improved since the pub- or "quality dumping." The claim is further licity given the Japanese "quality advantage" made that this is a high-cost strategy, intended during 1980.24 The available data is summar- to "buy" U.S. market shareand that after their ized in appendix 6A.Butat the same time that American competitors have been forced out, U.S. semiconductor firms have made rapid the Japanese will raise their prices and ship strides, Japanese manufacturers have also their normal product, which will be found to proved. While the gap has certainly narrowed, be poorer in quality.25 Indeed, manufacturers Japanese firms on the average may remain in many industries and in many countries ahead in quality. sometimes attempt such strategies. American semiconductor firms will sort ICs and ship It is also important to recall that discussions higher quality lots to purchasers who demand and data on IC quality have centered on prod- them. However, as a widespread and general ucts sold in the merchant market. No data have approach to marketing in the United States, quality dumping by, the Japanese seems im- 24Much of this publicity stemmed from a seminar entitled plausible. In order to ship higher quality lots "Quality Control: Japan's Key to High Productivity," organized by the Electronic Industries Association of Japan and held in to the United States, they would have to ship Washington, D.C., on Mar. 25. 1980. Data first released at that lower quality products to other customersin seminar appear in appendix table 6A-1, pt. A. A perspective com- either export or domestic marketsthus run- mon i,i much of the American merchant semiconductor induStry ..at that time can be found in T. D. Hinkelman, "The Economics ning the risk of jeopardizing those markets. It of Quality: U.S vs. Japan." An American Response to the Foreign is difficult to believe that Japanese IC manufac- industrial Challenge1' /IHigh Technology Industries, Proceedings of the Semiconductor Industry Association Government Policy 2,T. D. Hinkelman, op. cit. See also "The Quality Goes On Conference, Monterey, Calif.. June 18-19, 1980, M. Hodgson (ed.) Before the (Japanese) Name Goes On," Rosen Electronics Let- (Palo Alto. Calif.: Worden Fraser Publisher, 1980), p. 85. ter Mar. 31, 1980, p. 1. Ch.6Manufacturing: Quality, Reliability, and Automation 231" turers would do so in any concerted way, par- no reason why U.S. firms should not keep pace ticularly at home. with, or surpass, their overseas rivals on these dimensions of IC technology. It is clear from the data in appendix 6A that, at least until the recent past, Japanese large- scale ICs have had, on the average, both bet- Quality and Reliability of Color TVs ter quality and better reliability than compar- That Japanese TV manufacturers have able American parts. This does not mean that achieved excellent quality and reliability, and some products from some U.S. companies were largely succeeded in their marketing strategies, not as good as or better than products from is self-evident. In order to bypass the fran- Japan. As the tables in appendix 6A indicate, chised dealer networks that American manu- the range in quality and reliability delivered by facturers relied on, they had to forgo extensive any firm is likely to be wide; this is intrinsic service organizations. Failure by Japanese im- to the technology of semiconductor manufac- porters to maintain both the image and the re- turing. But as a generalization, the United ality of a reliable, trouble-free product would States had fallen behind in both quality and re- risk the largest market in the World. Most sur- liability. It is also clear that the performance veys continue to show the reliability of TVs of American firms on these dimensions has produced by Japanese firms to be better, greatly improvedin part because of the com- though differences in quality appear small. petitive pressures generated by the publicity given this issue. According to recent reports, Many of the TVs sold in the United States the quality levels of 16K RAMs supplied by a by Japanese firmsare now assembled here. number of American firms are now, on the Quality levels achieved in the U.S. operations average, about the same as those of Japanese' of both Sony and Quasarthe firm that Mat- devices.28 sushita bought from Motorola in 1974have received a good deal of publicity.28 Such plants While this is a positive sign for the future, tend to combine features typical of Japanese it does appear that Japanese firms devote more and American manufacturing operations; see resources to analyzing field failures so as to chapter 8 for a discussion of management find and eliminate their causes. In Japan, elec- styles and their effects. At least as important, tronics firms have normally maintai ,Jd cap- TVs assembled in the United States by Japa tive service organizations which gather and nese-owned firms contain large proportions of analyze field service results, and feed them imported components. Based on the findings back to design and manufacturing depart- for semiconductor devices outlined in the pre- ments. One American purchaser of Japanese vious section, imported components might be semiconductor. devices was reportedly quite expected to exhibit somewhat higher levels of surprised to find a team of engineers dis- quality -and/or reliability than similar parts patched to explore the reasons for a batch of from American suppliers. circuits with a defect rate of only 0.25 percent.27 MoSt of the information bearing on quality and reliability for TVs comes from sources like In the future, if American managers devote Consumer Reports. Several years ago this pub- as much attentionand as many resources to the quality and reliability of their products as do the managers of Japanese firms, there is 280n Sony, see "Statement of Sadami 'Chris' Wada, Assistant Vice President, Sony Corp. of America," Quality of Production and Improvement in the Workplace, hearing, Subcommittee on "E. R. Hnatek, "Semiconductor Memory Upda, ., DRAMs," Trade, Committee on Ways and Means, House of Represents= Computer Design, January 1982, p. 109; "Faults Show Up in fives, Oct. 14, 1980, p. 62. Japanese RAMs," Electmnics, Jan. 13, 1982, p. 33; "In Semicon- At Quasar, quality levels improved rapidly after the-Matsushita . ductors, Perfection Is the Goal," Business Week Nov. 1, 1982, purchase; however, the baseline is deceptive idthat Motorola p. 72. devoted few resources to its TV operations for a nbrither of years ""Quality and Reliability of Semiconductors and CTVs: U.S. prior to the sale. This case is discussed in more detail in the ap- v. Japan," op. cit., p. 57. pendix to ch. 8. 237, 232 International Competitiveness in Electronics

lication surveyed nearly 200,000 owners of Table 49.Rankings by Repair Shops of TV Receivers 19-inch color TVs, the most popular size, sold for Quality and Reliability during the period 1975-79. Nine of the fifteen Rankings in terms of picture quality and brands for which the origins are knownall other performance features: the Japanese makes but no otherswftre given 1. Zenith 2. RCA reliability ratings of "better than average" 3. Sony based on the average cost of repairs during the 4. Sylvania 1979 calendar year. The remaining six brand 5. Quasar 6. Magnavox names-/--for practical purposes, all the Amer- 7. Panasonic ican brandswere rated "average" (one brand) Ranking", in terms of reliability: or "worse than average" (five brands).29 The 1. Zenith Consumer Reports survey reflects reliabilities 2. S,)ny 3. F CA of sets sold" during the period 1975-79 only. 4. Quasar However, TV designs do not change rapidly; 5. Sylvania these trends should still be a reasonable indica- 6.Panasonic tion of comparative reliability levels. 7. Magnavox Rankings in terms of increasing costs of repair: Similar but not identical reliability rankings 1. Zenith 2. RCA come from a survey conducted by Trendex in 3. Sylvania the same year, 1979, but again, covering TVs 4. Quasar manufactured over a period of years.30 Of the 5. Magnavox 6. Sony 12 brands included in this survey, TVs made SOURCE: "Quality and Reliability of Semiconductors and CIVs:11.S, v. Japan," by Japanese-owned firms filled four of the top report No. C972, prepared for OTA by Consultant Services Institute, Inc., under contract No. 033.1170.0, p. 79. The survey, conducted dur. five places in terms of reliability. The reniain- Mg 1980, covered 60 repalr,shops In Chicago, Boston, and Northern ing Japanese brand ranked seventh, with the New Jersey. bottom five positions filled by American firms plus Magnavox. teriaperformance, reliability, and costs of Table 49 presents data from a survey of TV repair. In particular, the largest-selling U.S. . repairshops that ,:oint in a direction rather dif- TVsthose made by Zenith and RCA show up ferent from the L ..)nsumer surveys discussed very well, with Zenith top-ranked in, each category. In contrast, Zenith and RCA are rated above. ""7'";;able covers a smaller number of brands. .13e American (Zenith, RCA, and Syl- "worse than average" in reliability by Consum- vaniathe latter at that time U.S.-owned, er Reports. Because the Consumer Reports sur- though since purchased by Philips); three Japa- vey covered such large sample sizesmore nese (Sony, Quasar, and Panasonicthe latter than 40,000 owners of 19-inch Zenith sets, and two are Matsushita brand names); plus iviag- 35,000 made by RCAit must be given consid -yriavox. The repair shops rated the American erable weight. However, the data in table 49 brands generally superior on all three cri- are not restricted to any particular screen size, and might be more representative of each man- "19-Inch Color TVs," Consumer Reports, January 1981. p. ufacturer's overall product line. 34. The nine brands in the "better than average" reliatility category included TVs sold by Sears, most of which are made As is true for ICs, American manufacturers by Sanyosome imported, some assembled in the United States. Other private brand merchandiserse.g., Montgoniery Ward, of TVs have clearly made considerable strides J. C. Penney tend to purchase from both American and foreign in improving quality and reliability-7-spurred suppliers. Exchuding both the Wards and Penney TVs because by competition among themselves as well as of th,- Airlortain origins; 15'brands remain. Of the 16, 5 are ,; are Japanese, and the other Magnavox is owned with the Japanese. Consumer electronics firms by Philips. As Magnavox is much more nearly independent of now screen and burn-in components more its parent than the American subsidiaries of Japanese firrnS. it thoroughly; they also burn-in complete has been considered a U.S. brand for purposes of.this compar- out ison. boards and assembled sets to weed out early ""Quality and Reliability of Semiconductors and CTVs: U.S. failures. Automation has helped quality. Final- v. Japan," op. cit., p. 78. ly, American TV makers are using__ larger num- Ch.6Manufacturing: Quality, Reliability, and Automatio.' 233 hers of imported componentsmostly from Consumer perceptions created by and re- Japan and other Asian countries. Imported flected in surveys like those discussed above components often cost less, but in at least some can be extrapolated with some confidence into cases they have been chosen because of su- at least the near future. Furthermore, because perior quality and/or reliability. Even picture TVs have a design life of 7 to 10 years, the tubeswhich are bulkier and more difficult to surveys discussed above should do a good job ship than other componentsare being im- of predicting the reliability of sets presently in ported in increasing numbers; one ,U.S. man- use. The weight of the evidence points toward ufacturer stated that Japanese picture tubes had an advantage in reliability for Japanese TV one-third the in-process failure rate of Ameri- manufacturers during the 1970's. Even if can-made tubes.31 American manufacturers today are producing TVs as reliable as their Japanese competitors, "Ibid, p. 80. Japanese-owned firms that assemble and sell TVs the image of reliability that the Japanese have in the United States still import many components, but are grad - uallyjncreasing value added here. Mitsubishiwhich produces gained will persist for a number of years to sets in the United States for sale under the MGA brand name come. On the other hand, differences in quali- imports about 30 percent of their parts from a subsidiary in ty among TVs appear to be sma11.32 Singapore, and another 15 percent from Japan. Sony continues to bring in from Japan about 35 percent of the parts, for their American-made sets. In general, the more critical components and subassemblies from a performahpe and quality standpoint "For example, "Small-Screen Color TV's," Consumer Reports, are importede.g., circuit boards. Cabinets and nonelectronic January 1982, p. 17, where the distribution of brand ratings by parts are the first to be purchased domestically. See Quality of set performance and quality shows no systematic differences Production and Improvement in the Workplace, op. cit., p. 85. among U.S. and foreign brands.

Automation

Managers make decisions involving the auto- mechanical cams, to those that operate mation of production processes largely on the under computer control, to machining basis of costs. Automation typically involves centers; tradeoffs between labor cost and capital cost automated gaging, inspection, and testing; that depend on production volumes; mecha- examples include inspecting circuit boards niZed production facilities also tend to lack by means of video image processing to flexibility,. which raises the costs of adapting check for solder runs or othersible de- them to new product designs. Factors less di- fects, measuring the dimensions of ma- rectly related to costs include the impacts of chined parts, and determining the chemi- automation on quality, and the possibility of cal composition of steel; mechanizing unusually dangerous, dirty, or mechanized systems for materials han- onerous jobs. dling, ranging from computer-controlled Modern automated production systems usu- conveyors to fully automated warelmses; ally rely on electronics, although electrome- process control systems incorporating sen-, chanical control systems were common until sors and processors that implement con- recently. Examples of automated processing trol algorithms based on feedback, feedfor- ward, or some combination (see ch. 3, app. include:33 3C, on industrial process control); automatic machine tools, ranging from use of computers in management or sup- lathes and milling machines controlled by port functions such as scheduling of job flows, inventory control, or statistical quality control; and "See, in general, M.P. Groover, Automation, Production Sys- tems, and Computer-Aided Manufacturing (Englewood Cliffs. computer-aided design methods to aid in N.J.: Prentice-Hall, 1980). geometric modeling, in engineering anal-

99-111 0 - 83'- 16 233 234 Intern:Itional Competitiveness in Electronics

ysis, or in preparing design drawings or mechanize.34 In essence, the flexibility gained equivalent design information coded for through electronic controls makes automation automated production processes. cost effective in applications where production volumes are low. In the past, automation was The earliest numerically controlled (NC) ma- practical only in continuous process industries chine tools operated from instructions on a such as food preparation and packaging, or in paper tape or similar storage medium, analo- high-volume batch production industries like gous to the cams and other electromechanical automobile manufacture. In the automobile in- controls used for many years to automate man- dustry, simple assembly operations, as well as ufacturing. The NC tape, however, could be prepared with the aid of a computer, and easi- machining, have been carried out by transfer ly duplicated or modified. In the next stage, lines linking a series of machines for many rather than following 'a sequence of instruc- years; human operators have worked along the4 tions held in ayead-only memory such as a line performing tasks that were difficult or cost- paper tape,direct numerically controlled ly ly to mechanize. (DNC) and computer numerically controlled (CNC) machines were developed. These re- Fixed and Flexible Automation spond in real time to commands from the proc- Automated production in either continuous essor of a computer. As a result, control algo- process or batch industries can be thought of rithms based on gaging or sensing of machin- as spanning a range from "fixed" or "hard". ing parameters can, at least in principle, be im- automation to "flexible" or "programmable ". plemented. In a DNC system, one computer automation. Fixed automation is exemplified controls several machines; CNC machines by an automatic lathe in which the "instruc- operate under the control of a dedicated proc- tions" are encoded in the profiles of cams. To essor, typically a small minicomputer or a mi- set up the lathe for a different job, the cams crocomputer. must be changed. Designing and machining a Sophisticated control systems use informa- new set of cams .is a time-consuming job pertion from sensors for regulating the process, formed by skilled craftsmen. Conventional typically by adjustments that keep/measured transfer lines are examples of fixed automation parameters within predetermined bounds. For applied to a sequence of tasks. When an auto- a machining operation, dimenions can be mobile manufacturer designs a new engine or measured; for a. wafer fabrication line in a transmission, the entire transfer line might semiconductor plant, possible control param- have to be scrapped and replaced. Much the eters include temperatures, pH of reagents, and same is true if an electronics firm using such current flows during ion bombardment. In con- equipment wishes to introduce a new design trast to such "closed loop" systems, in which, fora printed circuit board, TV chassis, or com- information flows from the proCess back to the puter termi-J. controller, systems in which there is no sens- Until recenti,automated production equip- ing and transmission of information, but which ment with the flexibility to accommodate sub- operate purely on preprogramed instructions, stantial variation in the design of the product are called "open loop." A skilled machinist was the exception rather than the rule.35 Ma- closes the loop just as does an automatic con- chines seldom adapt very well to perturbations trol system on a CNC lathe equipped for auto- 34j. A. Alic. "Government Attitudes Toward Programmable matic gaging. But in fact, most NC machines Automation," Proceedings of the Twenty-third International Ma. chine Tool Design and Research Conference, B. J. Davies (ed.) still run on an open loop basis. (London: Macmillan. 1983), p. 521. 3,Flexibility in the context of manufacturing systems carries Electronic control systems make possible the a number of possible meanings; see, for example, D. Gerwin. automation of many processes that in earlier "Do's and Don'ts of Computerized Manufacturing." Harvard years were too difficult or too expensive to Business Review. March-April 1982. p. 107. Ch. 6Manufacturing: Quality, Reliability, and Automation235 in the processe.g., a part that comes down As flexible automation technologies incorpo- a conveyor sidewaysmuch less to new prod- rating computer-based control systems im- uct designs. When flexibility has been needed, prove, an enormous pool of potential applica- manufacturing operations have depended on tions will open; the consequences will include, people. Engine lathes, which are operated en- not only cost reductions and productivity im- tirely under manual control,are the flexible provements, but shifts in the composition of counterpart of the automatic lathe. A skilled the factory work force. Skill mixes needed in machinist can make an amazing variety of dif- manufacturing industries will change, and the ferent parts on an engine lathe, but the cost per total number of employment opportunities in part will be high. the manufacturing sector of the U.S. economy One reason for the lack of flexibility charac- may shrink even as total output increases. (Em- teristic of fixed automation is that new hard- ployment levels and work force composition warefixtures, toolingis needed to accom: are discussed in chapter 9.) modate a new design. A second reason is that the controls must be reprogramed. A hard- wired electronic control systemWhether ana Automation in Electronics Manufacture. log or digitalrequires new circuitry every Reasons for Automating time the control logic is altered. This is costly and time-consuming, just as for an automatic Most applications of automation by U.S. elec -. lathe that requires a new set of cams. In recent tronics firms have been driven by costs; non- years, computer control has became cost effec- cost factors have perhaps weighed more heavi- tive for replacing many mechanical or electro- ly for Japanese manufacturers. The industry in mechanical control systems. Japan has at times faced labor shortages; in ad: dition, Japanese firms may sometimes have While the performance of a computer-based been motivated by potential quality improve- programmable controller as a control sys- ments to automate earlier than their American temwill generally bq superior to the alter- counterparts.37 natives, this is not neceSsarily the.case for hardware. Often, flexibility l'n hardware trades off Table 50 presents the results of a 1979 survey against performance, a d perhaps capital cost in which. Japanese electronics manufacturers as well. For example, a r bot can be programed were asked to list reasons for their decisions to weld together section of pipe, but might not to automate. The most common response was be as fast as,ta specially designed automatic to reduce costs, with quality improvements sec- welding machinewhich might also produce -: ond; in contrast, a 1975 survey found labor better quality welds. However, the robot could shortages ranked at the top. Comparison of be programed to do other. tasks. In general, 1975 and 19'9 results shows a rapid increase then, a fleXible facility may be less efficient for in automation by Japanese electronics manu- making any one prods 1 than a dedicated, facturers,38 hard-automated manufacturing system.38 Another example of flexible automation in assemblythis one already in useis a machine developed in Japan by Nippondenv, "Recent R&D work at Westinghouse illustrates a typical appli- that can put together 288 different versions of an automobile cation of flexible manufacturing Ihere assembly, one of the most dashboard indicator. The average lot size is 40, with 200 change- challenging tasks for automatic). Westinghouse makes more overs per day. See "British Government Finances Robotics De- than 450 different models of sma !electric motors. with :..n aver- velopment," West Europe Report: Science and Technology, No. age lot size of 600: model changbs average 13 per day. In such 70, Joint Publications Research Service JPRS 78820, Aug. 25, cases, labor-intensive manufacturing "Methods have generally 1981, p. 14. been preferred. Fixed automation using transfer lines has been "K. Ito, E. Taira, S. Yagi, K. lwamoto.and K. Tsukamoto, "The a real option only for long prodtiction runs of similar or identi- Progress of Automation and the Improvement of Reliability in cal products. For a descriptionf the flexible assembly system Production of Color TV Receivers," IEEE Transactions on Man-. under development, see R. G. Atraham. "APAS: Adaptable Pro- ufacturing Technology, vol. MFT-3, December 1974, p. 55. grammable Assembly System."' Computer Vision and Sensor- " "Quality, and Reliability of Semiconductors and CTVs: United Based Robots. G. G. Dodd and L. Rosso! (eds.) (New York: Plenum States v. Japan," op. cit., p. 47. -Japanese consumer electronics Press. 1979), p. 117. firms reportedly spend about a third of their R&D dollars on man-

241 236 International Competitiveness in Electronics

Table 50.Reast,is Given by Japanese Japanese continued to take the initiative in Electronics Fly ms for Automating automation, even though their labor costs re- Percent of firms mained lower. By 1976-77, 50 to 80 percent of surveyed' Reasons for automating all component insertion in Japanese TV fac- 84% Reduction in manufacturing cost tories had been mechanized." Computer-con- 69 Improvement in quality trolled testing and inspection of components, 43 Increase in production volume 43 Improvement in workplace:. subsystems such as circuit boards, and com- conditions plete TVs also spread rapidly. Concurrently, 32 Labor shortage chassis were redesigned to take advantage of °Multiple responses were common. the characteristics of automated production SOURCE:Nikkan Kogyo ShIrribun,May 1, 1979, and July 11, 1979, cited In "Quality and Reliability of Semiconductors and CTVs: United States v. Japan,' equipment. According to one study, labor pro- report No. C972. prepared for OTA by Consultant Services Institute, Inc.. under contract No. 033.1170.0, p. 50. ductivity in the assembly of color TV receivers was a little greater in the United States than in Japan in 1970, but by 1977 productivity in Consumer Electronics Japan was more than twice that here--figure 43.40 Manual assembly was at one time the rule for electronic products ranging from radios and TVs to computers. Components were first Semiconductors inserted and/or soldered into circuit boards-, the In the earlier years of the semiconductor in- boards then installed in a chassis, the assembly dustry, virtually all production operations finally tested and adjusted. Component inser- fabrication, assembly, inspertior, and testing tion was one of the first tasks to be mechanized. were labor-intensive. Among LI firms, the This is relatively. easy for discrete components spread of automation may have Wen retarded with axial leads, more difficult for ICs. Con- by the widely publicized difficulties of Philco sumer electronics manufacturers first moved Corp., which invested ho''' inautomated to automatic insertion of discrete parts; as ICs manufacturin, duri7 g th AP; 1950's only to were designed into TVs, they were at first still see rapid ulidages ill trabstor technology inserted manually. By the end of the 1970's render its equipment obsolt;te.41 Philco later much of this work had been automated as well, dropped out of the semiconductor. business. using' equipment roughly similar to that pictured in figure 42.. At present, semiconductor firms in all parts of the world -are automating rapidly, not only 41'he spread of automation in the U.S. con- in manufacturing but in computer-aided circuit sumer electronics industry has been incremen- design. A few companiesboth American and tal. Firms automated at different times and for foreignhave installed fully mechanized pro- different reasons, depending in part on strate- duction lines, from wafer fabrication through gic responses to foreign competition. In most inspection, testing, wire bonding, and packag- cases, the initial reaction, based on Japanese ing. The benefits include increased yields as advantages in labor costs, centered on moving labor-intensive operations offshore rather than "For a description of some of the technology used by Mitsu- automating. bishi. see T. R. Crossley. "Study Tour of Industrial Robots in Japan," European Office of Aerospace Research and Develop- When American TV manufacturers did re- ment report No. EOARD-TR-80-3. London, August 1979, pp. 32- spond to competitive pressures by automating, 33. At the time of this visit. Mitsubishi was using robots of their own design for assembling printed circuit boards for TVs. The cost was the driving force, quality and reliabili -. assembly line could be changed over for a different board con- ty improvements secondary. Meanwhile, the figuration in 2 hours. 01. C. Magaziner and T. M. Hout, Japanese Industrial Policy ufacturing developments: see Transfer of Technology in the Con- `(London: Policy Studies Institute. 1980), p. 22. The data comes sumer Electronic Industry, Sectoral Study No. 2 (Paris: Organi- from work performed by the Boston Consulting Group. zation for Economic Cooperation and Development. Sept. 14, 41J. E. Tilton. International Diffusion ofTechnology: The Case 1979), p. 41. This percentage is probably a good deal higher than of Semiconductors(Washington. D.C.: The Brookings Institu-- in the United States or Europe. tion. 1971). p. 83.

24 , ti- Figure 42.

a result of better pro example, claims that rication facility has 15 percent, and prods cent; the control sy: mainframe compute] ers and74microcor . Rates of Automatic the United Statesa While much of the ward automation in 42"Toshiba Completes Full Japan Report, Joint Publicatii Jan. 19, 1982. p. 85. 238 International Competitiveness in Electronics

Figure 43.Average Labor Hours for Assembly of goods, particularly those for export." Such ac- 21-Inch Color TV Receivers (1977) tions have probably affected the timing of in- 6 vestments more than the eventual extent of automation. 5 Although a few American semiconductor firms make some of their own manufacturing equipmentnotably the larger, more highly in- 3 tegrated companiesmost such equipment is designed and built by independent suppliers. 2 In Japan, it is more common for electronics firms to design and fabricate their own. Mat- sushita, for example, meets 30 to 40 percent of its equipment needs internally.4e Internal capability for equipment development can help Japan United West United States GermanyKingdom speed automation. SOURCE: I. C. Magazlner and T. M. Hout, Japanese Industrial Policy (London: Policy Studies Institute, 1980), p. 23. Data are from a client study per. As integration levels for ICs continue to in- formed by the Boston Consulting Group. crease. automation will become a necessity. Fine line widths and other requirements for electronics companies do seem to have adopted very large-scale integration (VLSI) demand robots more quickly than their American coun- levels of cleanliness that are much easier to terparts. Two to three! times more robots are achieve if human intervention is minimized. at work in Japanese factories than in the United More complex circuits can only be designed States; about 40 percert of these have been in- with computer-aided techniques, together with stalled in the factories of Japanese electronics computer-aided drafting and mask generation; and electrical equipment firms." But again, a once built, such circuits can only be tested with number of the larger American electronic com- computerized equipment. Better process con- panies are well known; both for their R&D in trol modelsnow limited by gaps in funda- robotics technologyandfor their applications mental understanding of the physics of elec- of robots in manufacturing. tron deviceswill be needed to ensure the qUal- ity and reliability of VLSI circuits, Continued Some American electronics firms may have automation may reduce pressures for offshore had difficulty in finding the capital needed for automation. The replacement of labor-intensive '6"Quality and Reliability of Semiconductors'and CTVs: United production operations by capital-intensive States v. Japan," op. cit., p. 25. equipment aggravates the problems of financ- 461J. C. Lehner., "Japan Strives To Move From Fine Imitations ing expansion (ch. 7);atransfer line for insert: to Its Own Inventions," Wall Street Journal, Dec. 1. 1981. p. 1. Japanese firms continue to purchase a good deal of automated ing components in a circuit board can easily production equipment from American suppliers; as pointed out cost halfa dollars. In contrast, to their in ch. 4, about half the equipment used by Japanese semiconduc- smaller' American competitors, Japan's in- tor manufacturers comes from the United Stites. This percentage has been declining, however, with some observers predicting tegrated electronics manufacturers can take ad- that Japan will produce 70 percent of its needs by 1985. Japanese vantage.of internally generated fundsas well firms are reportedly already designing and building fifth-genera- as somewhat lower costs for external capital= tion automated wire bonders, while U.S. firms are still working with first or second generation machines; see "Pushing for to invest in mechanized equipment. Further- Leadership in the World Market," Business Week Dec. 14, 1981, more, the Japanese GOvernment has reported- p. 61. In some casese.g., automated testing equipmentjapa- ly given preferential tax treatment for in- nese 'products have the reputation of being somewhat more reliable than those of American suppliers, largely because they are vestments in production equipment that will simpler. However, industry opinion generally holds that U.S. improve the quality and reliability of Japanese equipment is as good as or better than that made in Japan or in Europe, as well as being less expensive. See "Can Semicon- 44R.Ristelhueber, "RobotiicsThe Applications Gap," Elec- ductors Survive Big Business?" Business Week, Dec. 3, 1979. tronic News. Jan. 11, 1982, p. 60. p. 81.

24.1 Ch. 6Manufacturing: Quality, Reliability, and Automation 239 manufacturing because labor costs will become The simplest. robots have only two or three a smaller fraction of total manufacturing cost. degrees of freedom; anillustration of atwos degree-of-freedom system would be ah "iifili" Robotics that could only extend and rotate, as for tight- ening a bolt. The most sophisticated robots Industrial robots comprise a subset of pro-have seven or eight degrees of freedom, which grammable automation technologies mimick- allows them to reach around obstacles. Figure ing some of the attributes of the human work 44 shows a pair of typical robot designs. force. They are more flexiblein terms of ability to perform new tasks, or to carry out com- While robots trace their desumt from more plex motion sequencesthan other types of primitiv6 manipulators having little flexibil- programmable equipment. Because advances itye.g., with position and sequencing con- in robotics depend to considerable extent on electronics and computer science they are dis- Figure 44.Two Approaches to,the Design' cussed in some detail below. In the future, of Industrial Robots robots will be used to automate many of the operations in making electronics products now carried out by hand. In Japan, robots are already being ined to produce more robots by a subsidiary of one of the major electronics companiesFanuc, a part of the Fujitsu organization.'" The changing proportion of costs associated with electrical and mechanical components since the first industrial robots were intro- I I I I is1.3 duced in the 1960's demonstrates the importance of electronics for this technology. A few years ago, about half the direct cost of making a robot was associated with the electronics, the other half with mechanical components. Now A..1 only about 15 percent of the costs go into elec- Photo credit: Cincinnati Milacron tronics, largely because of the increases in com -. (a) This robot emulates the human arm puting power/dvailable with cheap ICs. Costs for the mechanical components have not "V ' changed greatly, but the total 'osts of robots have decreased, the mechanical parts now accounting for 85 percent of the total. Technology Industrial robots, those used for factory work, are machines that can move a manipu- lator, or end effector, at the end of a chain of mechanical links. The end effector maybe a gripping device similar to fingers; alternatively, the end of the robot arm may carry a tool, welding torch, or nozzle for spraying paint.

Photo credit: Westinghouse 47N. Usui, "Untended Machines Build Machines," American Machinist. June 1981, p. 142. (b) This robot moves rectilinearly 240 International Competitiveness in Electronics

trolled by limit switchesstate-of-the-art eration robots is similar to that for DNC and designs now operate under computer control, CNC machine tools. Companies that intend to often a microcomputer. In routine production compete in the design and manufacture of fu- applications, the robot is commonly "taught" ture generations of robots will need a broader a sequence of motions by a human operator, range of technical capability; those moving intc who leads the arm through these motions while programmable automation from other high- they are stored in memory. The machine can technology fields may have the advantage, par- then repeat them on command. Although sat- ticularly firms with experience in automatic isfactory for simple applications like spray controls and the modeling of complex mechan- painting or some forms of welding, off-line pro- ical systems. gramingin which the instruction sequence is developed independently and loaded into the Robots in Manufacturing computer when neededis a major R&D goal. Robots are usually installed in factories Virtually all robots still operate with relative- where they can cut costs (compared to human ly primitive open loop control systems.48 This workers) and increase labor productivitythe is one of the factors limiting operating speeds, same motives that drive other capital, invest- as well as the accuracy with which the end ef- ments. In many early installations of robots, fector can be positioned. Current-generation human workers wore replaced on a one-for-one robots are also burdened by complex and ex- basis, a substitution facilitated-by the robot's pensive actuators that tend to restrict perform- ability to emulate the human arm. In compar- ance. At some future time, one robot may be ing the costs of robots to those for human workable to throw a part across the factory floor to ers, one-for-one replacement has often been as- another, but this i3 far in the future. Making sumed. This is potentially misleading because robots "smart"i.e., with the ability to gather a more thorough redesign of the production fa- and process substantial amounts of informa- cility means that some robots may each replace tion, then make decistons based on that infor- several people, while in other cases several mationis a related problem. Few robots can robots might be needed to do the work of one yet make even simple decisions. person. A cost analysis comparing robots and As such examples indicate, robotics technol- people must also take account of the number ogy depends on computer technology. While of shifts planned for the facility, maintenance computer firms like IBM, Digital Equipment and repair costs, depreciation. and energy con- Corp., and Texas Instruments are expected to sumption. A robot may use a hundred times enter the market for robotsand Fujitsu and as much energy, as a human worker. General- Hitachi are already two of the leading pro- ly speaking, when production volumes are ducers in Japanmany of the robots in current small, human operators will still be the least. production come from machine tool builders. cost alternative because of the expenses asso- In the United States, for instance, Cincinnati ciated with setting up and programing the Milacron has a substantial share of the market. robotfigure 45. Moreover, at sufficiently high production volumes, fixed automation will be From the perspective of the machine tool in- cheaper because there is no need to trade off dustrywhether the portion that builds metal- performance for flexibility. In general, robots cutting equipment or the manufacturers of tend to be the low-cost alternative ,for annual hard-automated assembly equipment such as production volumes of roughly 300,000 to 3 transfer linesrobots trace their descent most- million units." .ly from NC machines. In fact, much of the technology in the control systems of current-gen- The clot in figure 46 has been well publicized by Unirnation, one of the largest robot manu- ""Government Attitudes Toward Programmable Automation." facturers in the United States. It compares the op. cit. Much of the material that follows is drawn from this "R. Allan. "Busy Robots Spur Productivity,"IEEESpectrum, paper. Sepiember 1979, p. 3i. Ch. 6Manufacturing: Quality, Reliability, and Automation 241

Figure 45.Manufacturing Costs for Robots, Hard the robot is included, but not the engineering 'Automation, and Human Operators as a Function costs for the application. For a new installa- of Production Volume tion, development costs, including programing and general debugging, can easily total twice RobotsRob the purchase price of the robot. In electronics, robots can contribute to quali- Hard automation ty and reliability by minimizing mechanical damage to delicate partswhich also cuts direct costsand by helping improve cleanliness. The latter is particularly important for large- scale ICs, where "clean rooms" with levels of Human ge for dust and other contaminants reduced to very operators which robots low levels are required. Because people bring are the contaminants into the production area with low -cost them, automation helps raise yields and reduce alternative fabrication costs. Production volume Beyond the now routine applications like SOURCE Of hce of Technology Assessment. spray painting and welding lies a great deal of scope for robots that extend or improve on human capabilities. Some of these robots will be Figure 46.Ccst Comparison for Human Operator larger than those currently on the market, and Robot Assuming One-for-One Replacement and others smaller. While a few robots now avail- TwoShift Operation in the Automobile Industry able can handle loads in the range of 500 to 2,000 lb, most are designed with lifting capaci- notot Price $40,000 Useful life-8 years at two shifts ties in the range of 50 to 200 lb. Until recently, Cost of money-11 percent Installation cost $12,300 robot , intended for low loads (e.g., under 10 Maintenance cost $1.05perhour lb) and precision work have been rare. Limita- Power cost 35 cents per hour Overhaul 40 cents per hour tions on precision ai.d repeatability have 13 Depreciation $1.25 per hour placed severe constraints on potential applica- Installation 80 cents per hour tions.," Money cost 65 cents per hour

_ Human operator hourly cost Robots and Jobs Despite the fact that robots are simply one 5 Robothourly cost type of flexible automation, with roots in a 3 number of familiar manufacturing methods i and that automation itself is as old as the in- 1960 1963196619691972 1975 19781981 dustrial revolutionit is as difficult for many - Year people to be dispassionate and objective about SOURCE n. Allan,"Busy Robots Spur Productivity," IEEE Spectrum. September 1979, p. 31. robots as it is for them to regard nuclear power as just another means of generating electrici- costs for one of their robots with the costs of wages plus fringe benefits for an autoworker, "Electronics firms need robots with high accuracy because assuming the robot to be a direct-replacement.- the small and delicate parts used in electronic equipment are so easily damaged. Nippon Electric Co. (NEC) has recently de- According to figure 46, hourly costs for indus- scribed a machine with a load capability of about 5 lb and a trial robots have gone up only slightly since claimed positioning accuracy of 0.00016 inches, an order of mag- their introduction in the 1960's, a period over nitude better than previous high-precision robots. NEC plans r r to use it in circuit board assembly, as well as wire bonding for 242 international Competitiveness in Electronics ty. The entire set of technologies for automat- equipment in U.S. manufacturing industries is ing manufacturing and services poses very real currently $80 billion to $85 billion per year, and threats to the employment opportunities and second that during the 1980's expenditures on current job skills of a large segment of the U.S. robots are expected to remain well under 10 labor force, as discussed in chapter 9. But it percent of total expenditures just on automated is the whole family that is the proper focus of equipment.51.Again, from a job displacement attention. . "While it is too early to predict the viewpoint, all types of automation must be con- full range'ange of impacts of computerized manufac- sidered. turing, it is likely thatas with most instances Figure 48 illustrates the growth in sales by of major technological change these will application expected in Japan over the period come relatively slowly and in piecemeal fash- 1980.90. While many of the robots sold in 1980 ion. Just as these impacts are likely to be ran- were for use in casting, metal forming, (i.e., dom and incremental, many will be unex- forging and stamping), and painting, the pro- pectedand to the extent that they are, peo- jections in the figure indicate that these apple and institutions will be unprepared for plications will be far outstripped by assembly, them. welding, loading and unloading of machine Market Growth toola, and inspection. Some observers predict even more rapid market growth in the United If the effects of programmable automation States than figure 48 shows for Japan. will not become visible overnight, this is in part simply a result of time scales for production When markets grow this rapidly, a good deal and installation; rates of increase will be high, of technical and market volatility can be ex- but' total penetration will mount rather slow- pected, with many opportunities for entrepre- ly. Figure 47 gives estimates for worldwide neurial firms pursuing innovaive technologies. robot sales through 1990. According to this pro- While there are no guarantees that robot man- jection, the market will exceed $3 billion by the ufacturing will follow a path similar to semi- end of the decade, an increase of nearly 10 conduCtors, it would not be surprising to see times' over the 1980 level. Other estimates range startups in the United Stated challenging es- both higher and lower. To set these figures in tablished leaders like Unimation 'and Cincin- perspective, note first that spending for capital nati Milacron. The multidisciplinary demands of advanced robotsboth hardware (microelec- Figure 47.Worldwide Annual Sales of Robots, tronics, kinematics and mechanical design, in- Past and Projected strumentation) and software (artificial intel- 1=3 C=1 ligence and computer engineering, automatic USA Japan West Other Germany Western control theory, the production engineering European skills needed to integrate robots into the countries workplace)create conditions favoring innova- 400 tion and fresh thinking. New companies may emerge to lead this industry into the third generation of robotics, just as Unimation originally a startup and now owned by a larger 200 corporationled the first and second generations.

100 "J. T. Woodward, E. P. Seskin, and J. S. Landefeld, "Plant and Equipment Expenditures, the Four Quarters of 1982;" Survey 0 of Current Business, September 1982, p. 35 (capital equipment spending); "Industrial Robotics," Emerging Issues in Science flMnr,L; .,.,tom., r AInfLsrl ol Cnitannr Ch. 6Manufacturing: Quality, Reliability, and Automation 243

Figure 48.Robot Sales in Japan by Application 800

700

600 1990 Projection

500

400

300

200

100

Casting Heat Metal Welding Painting Machine AssemblyInspection Other treatment forming (arc and tool spot) loading SOURCE: Based on data in "Robot Industry to Grow Rapidly in 19131,- Japan Report, Joint Publications Research Service JPRS U9744, May 19, 1981, p. 33.

Japanese firms are applying robots in man- process. More than half the costs of typical ufacturing more intensively than their com- batch manufacturing operations are associated petitors in other countriesfigure49.Although with scheduling and otherwise managing the most observers feel that the United States still flow of productioni.e., with software, broadly leads in the relevant technologies, there are speaking. These management and production more firms building robots in Japan-130 to control costs involve: getting the right parts, 150, perhaps five times the number hereand, materials, and supplies to the right place at the indications that Japan may be ahead in learn- right time; seeing that shop floor personnel ing to apply robots in typical factory environ- have the information (now including computer ments. As figure47indicated, future robot in- programs) they need; and ensuring that ma- stallations in Japan are e:Tected to at least chinery is available and in good repair when match those in the United States. scheduled for use. Tasks involving production planning and machine scheduling, job flows, inventory control, and the related routing and coordinaung functions might-seem 'straightfor-- While the critical technical problems in the ward, but in fact they are extraordinarily com- further development of robots center around plex; experience shows them to be among the modeling and control. the critical imulemen- most critical factors in controlling mannfantur- 244 International Competitiveness in Electronics.

Figure 49.Estimated Numbers of Robots in Use, 1980

Japan United' Sweden West Italy United France Other States Ge-many Kingdom Western countries

SOURCE 'CAM: An International Comparison," American Machinist, November 1981, p. 214. least some of these costse.g., those associated compared to blue collar labor costs. Flexibili- with materials handling, control of the inven- ty will make small-batch production more at- tory of tools, jigs, and fixtures, routirof tractive; product differentiation and product partsbut only when appropriately integrated customization will become relatively less, ex- into the production environment. Integration pensive. The resultwill be far-reiching will require a great deal of rethinking at both changes in the product and marketing strat- the design and manufacturing stagesrethink- egies of manufacturing companies throughout ing for which cheap computing power is nec- the world.- essary but hardly sufficient. The potential benefits in terms of productivity are huge, but International Trends no one anywhere in the world knows at pres- As has happened in so many other industries, ent how _to realize them. Japanese firmswhich began to manufacture Computer-integrated manufacturing will af- robots by importing technology from the fect the cost structures of marl;r industries; as United States and Europeare now quite capa- labor productivity improves. fixed costs will ir- ble of advancinc the state of the art on their Ch. 6Manufacturing: Quality, Reliability, and Automation 245

of a leasing program. The Japan Robot Leas- ing Company, Ltd. (JAROL), incorporated in 1980, is owned 70 percent by 24 robot manufacturers and 30 percent by 10 insurance companies. About 60 percent of JAROL's capital has been borrowed from the Japan Develop- ment Bank and from commercial banks. The consortium uses this capital to purchase robots, which are then leased, primarily to smaller firms.53 While the program is similar to that operated by the Japan Electronic Computer Corp. for financing computers, JAROL does more than just lease equipment; its engineer- , ing staff provides assistance in installing and programing robots. Among its other initiatives, the Japanese Government has also granted an extra 13 percent depreciation allowance to pur- "chasers of advanced types of robots, while smaller firms that buy robots for modernization or to automate hazardous jobs can take advantage of low-interest loans.54 Much more ambitious is MITI's plan for a

4-0* joint R&D program aimed at making robots -Cc smarter.55 This effort, with a proposed annual funding level of about 30 billion yen (roughly Photo credit: Cincinnati Milacron $135 million), will be organized much like other Otwelding frame assembly for computer government-sponsored R&D programs in Ja- pan. About 10 companies are expected tobe involved, plus the Electrotechnical Laboratory ns; General Electric, for example, has of the Agency for Industrial Science and Tech- o build robots under license from nology. Major thrusts planned over the 7-year While fewer than 5 percent of the schedule beginning in 1982 include: roduced in Japan during 1980 were ex- mports were comparably small), Japan improved 'sensory *capabilitiese.g pat- to export 15 to 20 percent of its robot tern recognition, force/torque sensors; ion by 1985..62 control algorithms incorporating true adaptive or "intelligent" behavior which would as in other industries, the Japanese allow the robot to operate more-or-less au- nentvia the Ministry of International rid Industry (MITI)has designed pro- builders of 53Y. Machida, "Industrial Robot in Japan," LTCB Research, o encourage and support Long-Term Credit Bank of Japan. March/April 1981, p. 4. \. number of Western European govern- "The special depreciation pro Jisions apply to robots pur- re following suit. MITI sponsored sev- chased between 1980 and 1983 that are computer controlled, nufacturing-oriented R&D programs pave six or more degrees of freedom, and meet specified standards for positioning accuracy. The added 13 percent deprecia- ncompassed robotics during the 1970's. tion in the first year means that a total of 53 percent can be writ- the agency's first steps to support the ten off (assuming the 5-year, double declining balance deprecia- industry itself was the establishment tion procedure that is normal in Japan). See "Robotics: They Are Smart and Never Need a Tea-Break," Far Eastern Economic Re- view. Dec. 4. 1981, p. 70. 246 International Competitiveness in Electronics

tonomously, making decisions based on The program is in part a sequel to previous sensory data it receives from the operating MITI-sponsored work on remote control de- environment; and vices for maintaining and repairing the radio- improved mechanical design, including active portions of nuclear powerplants. How- manipulators and mobility, the latter very ever, the robot program will be much broader. much a controls problem as well.

Summary and Conclusions In the past, Japanese electronics firms mak- assurance has often been an orphan in the ing both TV receivers arid ICsnotably mem- United States. Quality control personnel here ory chips for the merchant markethave em- have been viewed as obstacles to production; phasized quality and reliability more heavily they have had integral roles in neither design than their counterparts in the United States. nor manufacturing. Japanese firms learned This does not mean that the best performing many of the basic techniques of quality con- American firms may not have had quality and trol from American engineer s, but they have reliability as good as the Japanese, or that cap- gone a step beyond conventional practice in tive manufacturers in the United States may much of the rest of the world by, for instance, not have been as good or better than IC makers making individual workers responsible for the anywhere. It does mean that specific types of quality of their own efforts. productscolor TVs and dynamic RAMs have, in the past, been produced to higher aver- Electronics firms in Japan also invest more age levels of quality and reliability in Japan. heavily in employee training. At all levels The picture at present is less certainindeed, assembly workers, engineers and designers, reliability cannot be estimated until products foremen and supervisors, sales and manage- are well along in their service lives. It is plain mentemployees of American electronics that American firms have made major efforts firms tend, on average, to be less knowledge- to improve quality and reliabilityWith con- able concerning quality and reliability than siderable success. But it is not obvious that they their counterparts in Japan. Although many of hay. ghtmuch less surpassedthe Japa- the recognized authorities in these fields are ns::_:: who have been improving their own per- Americans, expertise is not spread as widely formance at the same time. here as in Japan. Moreover, U.S. electronics manufacturers may still to some extent be 'pay- Japanese manufacturers may have succeeded ing lip service to quality and reliability. Over in creating perceptions of quality and reliability time, the depth of their commitments will be- outstripping any actual performance margins, come more apparent. In particular, it takes particularly for color TVs; certainly the strat- time to design quality and reliability into a egies of Japanese electronics firms have par= product line. allels in other industriese.g., cameras or auto- In general, Japanese electronics manufactur- mobileswhere the emphasis placed on both ers also appear to do a better job of managing the image and the reality of quality had an im- the interface between design and production portant role in the penetration of U.S. markets. Moreover, the characteristically close w3rking For American firms to match this image de- relationships between vendors and purchasers mands top management attention. in Japan's electronics industry evidently yields benefits in quality a.ud reliability. Production Ch. 6 Manufacturing: Quality, Reliability, and Automation 247 chase a good deal of manufacturing equipment Japanwith uni..redictable effects, as for any from U.S. suppliers. new technologythe cumulative impacts of these and other types of factory automation Japanese companies automated the produc- will be massive, affecting productivity and tion of TV receivers and other consumer elec- competitiveness, the skill mix in the work tronic products earlier than most American force, and the total number of job opportunities firms. Extensive applications of robotsin in the economy. Computer-integrated manufac- electronics'and other industrieswill help the turing will shift corporate strategies in many Japanese increase manufacturing productivi- industries toward greater product differentia- ty still further, and will also improve quality tion. Japanese companies can be expected to and reliability. At present, robots remain a apply computerized manufacturing technolo- small subset of automated production equip- gies at least as fast as American firms, wher- ment with limited impact, but they will be a ever there are benefits in terms of cost, quali- key part of future manufacturing facilities. ty, worker safety, or product design and mar- And, while robots will spread rather slowly keting. through industry in both the United States and

Appendix 6AQuality and Reliability Comparisons for Integrated Circuits

This appendix summarizes the data that have during 1980, but that they continued to trail Japa- been made public concerning quality and reliabili- nese firms. Improvements by Japanese suppliers ty levels of .:hips supplied by American and Japa- over the time period covered in the table were negli- nese firms to the merchant market. The most wide- gible. Note that failure rate after burn-inparts B ly noted comparisons have come from Hewlett- and C of the tableis essentially an indication of Packard Corp., a U.S. firm that purchases large infant mortality failures, and thus more closely as- numbers of semiconductors on the merchant mar- sociated with quality than with reliability. Needless ket, and also manufactures ICs for internal use. to say, conclusions based on such results should be generalized only with care; the table refers solely Quality Levels to circuits purchased by Hewlett-Packard, and differences from shipment to shipment even from a

As indicated in table 6A-1, part A, at the end of . single manufacturer can be large. 1979 the quality levels of Hewlett-Packard's U.S. Table 6A-2 presents data gathered for OTA on suppliers were poor relative to Japanese firms. quality levels for RAMs, 4K as well as 16K. As for While Hewlett-Packard had an obvious interest in the first set of the Hewlett-Packard data, the Japa- pressing their suppliers to provide high quality, this nese 16K RAMs were superior by a large margin. data is not just another case of a purchaser play- The 4K RAM datathough limited to only one Jap- ing its vendors off against one another; the semi- anese vendor are quite different, shbwing the conductor industry has generally accepted Hewlett- American-made devices to be superior. Packard's test results as valid, although offering a Along with quantitative data on RAMS such as variety of explanations for the relatively poor showthat in tables 6A-1 and 6A-2, purchasers of ICs sur- ings by- domestic manufacturers. veyed by OTA's contractor sometimes provided The data in table 6A-1 are all for 16K RAMs; in- more general comments. One purchaser, for in: deed, most of the public discussion of quality has stance, included the fcilowing comparison: focused on RAMs, because they are bought in large quantities by many customers and have become a Percent of ICs rejected on 248 international Competitiveness in Electronics

Table 6A-1.-Hewlett-Packard Data on 16K Random Access Memory (RAM) Circuits

A. Reported March 1980. Country of Percent failing Field failure rate H-P composite manufacture incoming inspection per thousand hours) quality indexa Japan- 1b 0 0.0100 /o 89.9 2 0 0.019 87.2 3 0 0.012 87.2 Unitrid States- 1 0.19% 0.0900/0 86.1 2 0.11 0.059 63.3 3 0.19 0.267 48.1 aThis index is based on 10 equally weighted factors, of which incoming inspection and field failure rates are two; the others include such things as cost and delivery schedules. °Evidently, the three Japanese suppliers (not necessarily in order.) were Fujitsu, Hitachi. and NEC. tge American suppliers (again not in order) Intel, Mostek, and Texas Instruments. See The Quality Goes On Before the (JapaneSe) Names Goes On," Posen Electronics Letter, Mar. 31, 1980, p. 1. SOURCE: R. W. Anderson, "The Japanese Success Formula: Quality Equals the Competitive Edge." Verbatim Record, Seminar on Quality Control: ,!span's Key to High oroductivity, Washington, D.C., Mar. 25, 1980, p. 40.

B. Reported November 1980. Country of Failure rate after manufacture burn-in

Japan- 1 0.05% 2 0.10 3 0.12 Average = 0.17%a 4 0.35 5 0.25

United States- 1 0.60 2 0.50 3 1.20 Average = 0.75%a 4 0.70 aAverages are not weighted by numbers of circuits from each manufacturer. SOURCE: B. Le Boss, "U.S. Reject Rate Still Trails Japanese," Electronics, Nov. 6, 1980.

C. Reported April 1981. Failure rate after burn-in Country of First half Second half manufacture 1980 1980

Japan- 1 0.06% 0.04 2 0.13 Average = 0.25%a 0.13 Average = 0.24%a 3 0.40 0.40 4 0.40 0.40 United States-1 0.60 0.35 2 1.20 Average = 0.97%a 0.20 Average = 0.35%a 3 1.10 0.50 aAverages are not weighted by numbers of circuits from each manufacturer. SOURCE: R. W. Anderson, Seminar on Management, Productivity and Reindustrialization: East Meets West, Washington, D.C., Apr. 2, 1981..

Consistent with such patterns, an independent test- Reliability Levels ing firm noted that rejection rates following screening and burn-in were twice as high (about4per- While screening and other tests can locate defec- cent) for American-made ICs as for Japanese (1 to tive circuits and :neasure quality, determinations 2 percent). End users of ICs generally reported sim- of reliability must await field experience; long-term ilar results, several noting that they no longer'felt equivalent of average /apnoea() products. Good Japanese lots run at a re- Ch. 6-Manufacturing: Quality, Reliability, and Automation 249

Table 6A2.-Quality Levels of Japanese and U.S. Table 6A3.-Reliability Levels of Japanese and Random Access Memory (RAM) Circuits U.S. Random Access Memory (RAM) Circuits

Percent rejected on Country of Field failure rate incoming inspection manufacture (°,2i, per thousand hours) Country of A. 16K RAMs Japan' 1 0.0062% manufacture Average Range 2 0.0263 Average = 0.027%4 A. 16K RAMs 3 0.0507 Japan- 1 0.30% United S:ates- 1 0.0167 2 0.53 Average = 0.87%a 0-1.8 2 0.0687 3 1.77 0.1-5.0 3 0.088c; AverageG.125%a United States- 1 0.70% 0-2.7 4 0.107

2 0.85 Average = 1.7% 0-4.7 5 0.1268 3 1.07 0-6.2 6 0.3421 4 4.11 0-12.4 8. 4K RAMs 8. 4K RAMs Japan- 1 0% Japan 1 07% 0-5.3 United United States- 1 0 32% 0-2.0 States- 1 0.0b24 2 0.41 Average 0.53% 0-24.5 2 0.0526 3 0.87 0.1-1.1 3 0.1018 3,,,ufacies are not ,sec;ateo oy nuiroufs of cocuos from each manufacturer C. 1K RAMs r-souncE uasi,.a id hisi,3t.ehi,, ofSE31',:.011CIUCtOrSand CTVs United States! Japan- 1 0.0756% v Japan," report prepared for OTA by Consultant Services Institute.. under contract No 0331170, p 72 United States- 1 0.0667 aAverages are not weighted by numbers of chips from each manufacturer. testsare veryexpensive, and burn-in failure:; mor SOURCE "Quality and Reliability Of Semiconductors and CTVs:United States v Japan." report prepared for OTA by Consultant Services Institute. properly ascribed to quality problems. Field failure Inc., under contract No 0311170.0, p. 72. data w;sembled by I fewlett-Packard for 16K RAM's were included in table 6A-1, part A. Table GA-3 contains theremainderof the reliability data available The frequency of soft errors caused by alpha radi- to OTA. Consistent with the Hewlett-Packard re- ation can be reduced by a number of techniques. sults, this shows the reliability of Japanese 16K which Japanese manufacturers evidently imple- RAMs to have been markedly better than Arnericfm mented more rapidly than American firms-perhaps because Japanese semiconductor firms were,. Products. more willing to accept the extra costs. Onepur- chaserof 64K 12AMs reported soft failure rates of Soft Errors 10"a per hour for circuits from two Japanese manu- Failure modes for ICs can be divided into "ha d" facturers; the rates for the products of a pair of and -soft" Failures. } lard failures are repeatable md American firms wire 10-3 and 1076 failUres per final; the device no longer functions properly. In hour.3 cont rast, soft failures are random and nonrer eat- able. Alpha radiation can cause soft errors in I AM circuits, a problem that was not appreciated nntil densities reached 1.:1:. The radiation-emint d at low levels from ceramics and other materials ised in packaging It;s:--sortaitimes causes a bit ste7 ql in ,"Quality and Reliability of Semiconductors and CTVs: G.S. v. Japan," a memory cell to switch front -0" to "1"ori vice op. cit., p. 73. Several years ago, the alpha-induced soft error rate for Fujitsu's lid< RAM was reported to be three orders ofmagnitude better versa., The result is a soft error that appears 'vhen than that of Musters device. The differeoims were attributed to the de- the contents of that cell are next recalled. signs of the circuits. SeeJ.G. Pose, "Dynamic RAMS: ;Nita! to Expect Next," Electronics,May22, 1980, p. 119. Tile resistance of Mostek's IRK 'TI.MO.SOftErrors in VI.S1: Present and Future," /LT!Trans- HAM to alpha radiation has since been greatly improved, and U.S. firma :ems nn Components. Hybrids, and Afarni factoring 'Thcfmoloy. in general have adopted measures that substantially reduce sensitivity (AIM Ii noble I979, p. '177. to alpha particles. CHAPTER 7 Financing: Its Role in Competitiveness in Electronics Contents

Page Overview 253 Sources of Funds and Financial Leverage in the United States and Japan 255 :tnui Exiornal Firiancin "57 Pisk 258 1)1..' lntlimiimn and Banking Practices 26(3 Eitects mu! Financial I.VVOnigc on Tax Paynumts 262 k Absurption in Japan 262 Other Factors in Costs of Financing 263 "17.1. anti Diversification 263 Nat.os 264 Costs of Capital for Elei !rimics Firics in the United States and Japan: Summary 266 Financial Structure: An International Comparisor 268 268 280 Fraii.a 289 1,1:!-.! 292 Summary and Conclusions 294

List ofTables

- 11.,si- (1a6ital 11.S. and lapatieso Semiconductor Lii1C111,1101 1)V Chase Fitoincial Policy 7.56 acd F\ternal Sources of Corporate Financiiii.; To . JuimiR Ratios fur Selected din] Janese Electuinics Firms 258 Savi:1_Is Pates in Several Industrial ()trims 264 iimistiliold Asset; in the l'inted States and Japan. 1978 265 Adjme,tH Pates of RI! mire in the Inilmud States anti japan 266 ' \11,',-11.c;',,Ile ".S. Venture hive:anoint Activity 270 Internally Cieliorated Funds as a Percent of' 'Fotill.Capital From All Sour':-; 27:1 .,11. Ay.,ets of and Japanese Semiconductor Nlanufacturers ml). Sources of External Financing for U.S. and Japanese Corporations 282 61. Short- and Long-Term Debt as a Percent of r.till('',apitalization tom :.S.111(1 fapiinese Electronics Firms 2113 6.1Asset (:ategories as Listed on Balance Sheets al Electronics Firms in the l'Itited States and japan 284 6.t. Aver:we Assets and Liihilitimus of laparnise Commercial Banks 285 64.ilitortA Rates and Price Inflation in Japan 216 65. Household Borrowing in Japan and the United States 288 oti.lfolders of Liiihilities and With the Nonfinancial Sector in the Slates and France 9 9(1 List of Figures Page incicaso in Capital Costs fur Iligh-Volinne Integrated Circuit Production, lane 271 1,1ales of Capital Spending hy 1.1.5. and Japanese Semiconductor Firms 27,1 lila Filnity Patios for Japitnesti ElectricalfElectrimi, 282 CHAPTER 7 Financing: Its Role in Competitiveness in EJe6tronics

Overview

Declines, real or imagined, in U.S. competi- ownership relation with the issuing company, tiveness in electronics have been ascribed at but receive a fixed rate of return, as well as pos- various +.. imes and by various people to such sible capital gains (or losses).* Shareholders ac- causes as: unfait competitive tactics by foreign cept a variable rate of return in the form of div- firms, trade barriers that keep American prod- idends, as well as changes in the value of the ucts out of overseas markets, government sub- stock depending on the success of the com- sidies in other countries, and costs of capital pany. Both stocks and bonds are traded in ac- that are lower than in the United States. Low- tive secondary markets in the United States cost investment funds are saidtobe available and many other industrialized nations. In gen- in countries like Japan for-reasons ranging eral, holders of debtof which bonds are only from higher rates of consumer savings to allo- one typehave first claim on the residual as- cations of capital by gollernments or direct sub- sets of a corporation in the event of liquidation; sidies. the claims of stockholders are subordinate. This chapter deals with only this last set of Highly developed capital markets such as possible causesthose related to corporate fi- those in the United States also provide indirect nancing. Although limited in scope, the discus- financing mechanismsi.e., one or more finan- sion has clear implications for other facets of cial intermediaries are interposed between the competitiveness. For example, financing costs investor and the final recipient of funds. Banks could be lower where a protected home market are the most common intermediaries. Investors reduces risk and provides a stable foundation deposit moneysfor instance, in ordinary sav- for international operations. Government sub- ings accountswhich the banks then lend to sidies might he indirectly channeled through businesses. Other financial institutions func- financial markets as implicit or explicit loan tion in generally similar fashione.g., the post- guarantees, as well as in more obvious forms al savings system in Japan, an important chan- such as grants for research and development nel for capital that ultimately helps finance or tax havens encouraging regional develop- Japanese industry. Investment banks, insur- ment. ance companies, and pension and retirement In mature industrial economies, a vast and funds are other examples of financial interme- varied network of channels links companies diaries. seeking funds with individuals and organiza- The fundamental questions in this chapter tions that have moneys to lend or otheiwise in- deal with costs of capital faced by electronics vest. The capital markets where transactions firms in various parts of the world. Spukesraen between those seeking and those providing for American companies have often compared funds take place accommodate both direct and indirect investments, for short time periods and for long. Among the direct and long-term 'Several types of corporate bonds exist. Straight bonds carry __AL _ L a fixed interest rate, but their market price varies with economic International Competitiveness in Electronics their costs for investment fundswhether debt The interest rate thus serves a critical func- (bank loans, bonds) or equity (primarily stock tion in the economythat of the price for bor- issues)unfavorably with costs in other coun- rowed funds. This price serves to allocate tries. in particular, costs of capital in Japan are funds so that the pool of available capital goes often said to be as little as half those in the first to the most productive investments. The United States. Some observers also claim that mechanism is as follows. Managers of profit- the pool of funds potentially available for in- seeking enterprises make investment decisions vestment in the U.S. industry is too small. by comparing their costs in acquiring funds with the expected profits from the uses of these Such concerns are particularly relevant for fundsi.e., with the returns on alternative in- the rapidly growing, high-technology portions vestments. These projects might be new man- of the American electronics industry. Firms ufacturing facilities, R&D programs, or the ac- whose business centers on semiconductors, quisition of other corporations. If the antici- computers (including software), and even the pated returns are greater than the costs of ob- more rapidly expanding portions of consumer taining funds, then the investment might be electronics (e.g., electronic games) can find made using money generated within the enter- themselves with markets outstripping their prisee.g., from retained earningsor from ability to finance expansion. outside capital markets. In either case, the in- Problems in securing funds for rapid expan- terest rate is the primary factor in determin- sion riot only of production, but of R&D and ing the cost of financing the project. For ex- product developmentare compounded by the ample, if market interest rates are high, a cor- rapidly increasing capital intensity of some poration might choose to invest in securities portions of the electronics industry. Semicon- rather than in its own business. In general, less ductor manufacture is a prime example; capital attractive investment projects will be post- costs are going up rapidly, not only because poned when interest rates rise, the market serv- of escalating design cost as circuits become ing to allocate funds to other uses both within more complex, but alsc 'because new genera- the firm and among various companies seek- tions of production equipment are much more ing financing in the capital market. expensive (ch. 3). Given the predilection of U.S. firms, in electronics as in other parts of Amer- The market-driven process described does ican industry, for relying on internally gener- not always function ideally, but as a rule inated fundsi.e., retained earnings and depre- terest rates allocate funds quite efficiently. Still, ciationwhenever possible, financial mana- governments can act in various ways to influ- gers have often been hard pressed to secure ence investment decisionseither on a case- funds for growth. by-case basis or by favoring some sectors of the economy over others. Outright subsidiesand Because the chapter centers on costs of cap- loan guarantees are two of the more obvious ital, interest rates and the mechanisms by and common tools. Less visible and less direct which they are determined become one of the policies are also possible; some of these are ex- fundamental bases of comparison. The cost to plored in the discussion of financing practices the borrower of acquiring funds is the interest in Japan later in the chapter. Where govern- rate' on the loan or bond. Costs of equity, fol- ments intervene in capital markets, one con- lowing conventional practice, can be related sequence can be higher interest rates forall to costs of debt. In countries with well-devel- borrowers except those favored by the govern- oped capital Markets and -modest levels of ment. government interventionas in the United Steinsrnerket-rietermineri interest rates are To explore internationaldifferencesin Ch. 7 Financing: Its Role in Competitiveness in Eiectronfcs 255

More limited discussions of France and West where governments may try to complement Germany follow. The objective is to understand corporate strategies or to implement national the effects of financing patterns on competi- strategies. tiveness in the international marketplace,

Sour es of Funds and Financial Leverage in the United States and Japan

Many executives in the U.S. electronics in- them at lower prices, At times, U.S. firms have dustry believe the firms they manage to be con- also associated low-priced products with "un- strained in efforts to defend or expand inter- fair" practicesin international trade (see ch.11). national markets by relative costs of financing, Certainly, q broad range of business tactics and in some cases also by shortages of capital. whether or not fair within the accepted frame- The electronics industry is not alone in this work of international tradeare easier to im- concern. Other American industries, especially plement if capital is inexpensive. those facing intensified international competition, voice the same complaint, more especially The U.S. semiconductor industry has also as- if they feel threatened by the Japanese. The ar- serted that favorable access 'to funds has en- gument has been articulated bestand empir- abled Japanese manufactures to add capacity icai* supported in mcst detailby the U.S. in advance market demandindeed, to cre- semiconductor industry, largely because its ate excess capacity even in times of recession rate of expansion and changing technical char- a "luxury" decidedly unavailable to Ameri- acter place extraordinary demands on the fi- can firms. As a consequence, when the econ- nancing capabilities of independent merchant omy improves, the Japanese are better placed firms. The semiconductor industry's position to quickly move into expanding markets, while with regard to financing is summarized helovt: their competitors here struggle to build capac- to the extent possible the argument will be gen- ity and catch up. Finally, it is alleged, ample eralized to other sectors of electronicsi.e., supplies of cheap capital allow. Japanese corporations to spend lavishly on the advanced R&D computers and consumer products. so necessary in this rapidly changing field. The basic contention of the semiconductor Lower costs of capital_ together with full con- industry is straightforward, and for the most trol over their domestic market, are viewed as part directed toward the industry's primary for- primary underpinnings of Japan's global strat- eign competitor, Japan: the ability of Japanese egy. electronics firms to gain market position against American companies over the past few What are the perceived reasons for these low- years, both in the United States and abroad, has er capital costs? Two main causes are frequent- been eased by cheap capital. (The meanings ly cited, along with related structural features that attach to cost of capital will become clearer of the financial system in Japan: 1) the distinct- below.) ly different capital structures of Japanese electronics companies; and, 2) the very high rate For one reason or another, in this view, Japa- of savings within the Japanese economy. For nese corporations in many industries enjoy structural reasons; Japanese firms can tap rela costs of capital markedly lower than their tively large amounts of nonequity funds, pri- 256 International Competitiveness in Electronics revenues to finance growth. NonequitY funds, Table 51.Costs of Capital for U.S. and Japanese Semiconductor Manufacturers as Calculated it is claimed, tend to be less costly. by Chase Financial Policy The second source of Japanese advantage- - Weighted averages high savingsby increasing the pool of funds of debt and equity available to be lent, should depress interest ° costs as of rates. This would have the effect of making all June 4, 19808 types of investment capital less expensive com- U.S. companies (calculations in dollars) Advanced'Micro Devices 17.7% pared to countries where savings are a small- Fairchild Camera and Instrumentb 15.5 er proportion of gross-national-produck-Sav !Mel ...... 16.8 ings rates are discussed in more detail in a later Irrtersilc 21.1 Mos7,ekd 16.7 section; household savings in Japan run at Motorola 13.8 about 20 percent of income--nearly four times National Semiconductor 17.4 the rate in the United States. There is little Texas- Instruments 16.5 agreement on why the savings rates in different Japanese companies (calculations in yen) Fujitsu 8.8% countries vary so'much, and in particular why Hitachi 12.1 that in Japan is so high and that in the United Matsushita Electric 17,1 States so low. Variations in the average pro- Mitsubishi Electric 7.7 Nippon Electric CO. (NEC) 7.7 clivity of individuals in different countries to Toshiba 7.7 save under otherwise similar circumstances aln terms of required overall rate of return on Invested capital. b Subsequently acquired by schlumberger. appear to be a factor; so do the extent of social cSubsequen9y acquired by General Electric. welfare programs and differing tax structures. dSubsecitieritly acquired by United 7ectihologies. SOURCE:"Q.S. and Japanese Semiconductor Industries: A Financial Com Ccimbined, these two sources of financial ad- garlson,- Chase Financial Policy for the Semiconductor Industry vantage are said to give Japanese electronics Association, June 9, 1980, tables 4 and 9, pp. 5.3 and 7.6. firms capital costs barely half those of their American competitorsin 1980, about 9 percent compared to15 to 18percent for U.S. semiconductor firms.1 Such a result, if true, has implications for competitiveness in many other than any of the U.S. companies, at least for the industries. time period examined. Nonetheless, the range incapital costs faced by firms in either coun- The conclusions of the Chase Financial Pol- try is relatively large. icy study cited above are summarized in table 5.1. According to Chase's calculations, the typi- There are two major reasons for the wide di- cal Japanese manufacturer of semiconductors vergences in capital costs in table 51. First, bor- enjoys substantially lower costs of capital than rowing costs used in the calculations for Japa- merchant firms in this country. Only Matsu- nese firms were lower than rates for American shita (table51)incurs financing costs larger companies, The second primary solircQ:of difference lies in the dissimilar capital structures "U.S. and. Japanese Semiconductor Industries: A Financial of corporations in the two countries--the Comparison," Chase Financial Policy for the Semiconductor In- greater use of debt by Japanese firms, principal- dustry Association, June 9, 1980, p. 2.5. The most thorough discussion so far of the impact of corporate financial structure on ly in the form of bank loans (most American relative costs of capital, this report seeks to quantify Japanese firms with substantial debt carry this in the and U.S. financing costs with considerable care to identifying form .,f bonds). the sources of the differences, A more receot anal; ::s comparing industry as a whole in the If Japanese firms use.substantially more debt UnitedStates and Japan finds average costs of capital for 1981 to be 16.6 percent here versus 9.2 percent in Japan. See, "A than U.S. companiesas they doand if debt Htsterical Comparison of the Cost of Financial Capital in France, financing is less costly than equityaS the com- the Federal Republic of Germany, Japan, and the Uiiited States," putational methrd used by Chase assumes- Department of Commerce. April 1983. In this report, no attempt 4101-1nfrith] rnet riarivarl from a wpichtpd aviar. Ch. 7Financing: Its Role in Competitiveness in Electronics 257

would imply cost advantages for Japanese com- greater use of financial leverageand the panies, not only in electronics but in any indus- consequencesare take'i up later. try making similar use of leverage. The conclusions of the Chase study concerning the impact of debt financing on capital Internal and External Financing costs in Japan are grounded in well-accepted methods of calculation. The cost of capital for Table52illustrates something of the range a particular investment can be estimated using in international differences in corporate fi- the relative proportions of the company's nance. Japanese capital structures are heavily sources of capital as weights in the computa- weighted toward external financing. Japanese tion for the investment. For example, if a com- corporations, on the average, received less than pany pays 15 percent interest on debt instru- half their capital from internal sourcesi.e., ments, and its risk-adjusted cost Of equity (ex- from depreciation and reinvested profits. And, plained below) is 20 percentand if the debt- while Japan is at the high end in use of exter- equity ratio is 1.0then the firm's overall cost nal capital, the United States is atthe low 'end, of capital would be171/2percent. The returns relying much more heavily on internally gen- expected from a given investment can then be erated funds. compared to this estimated cost of capital. The computational method is deceptively' simple The category of external finance includes andexcept for various subtleties involved in both loanswhich in all,five countries are ex- determining the appropriate interest rate for tended primarily by banksand securities. The debt and the risk measures for equitycan be two- major categories of securities are bonds applied in straightforward fashion. (like loans, debt) and stocks, reptesenting equity holdings.2 Note that Japanese firms rely All other things equal, then, Japanese firms ,much more heavily on loans than securities would enjoy clear financial advantages, from (either loans or equity) for their external fund- their greater relative amounts of debt (higher ing; in general, companies in Japan employ leverage) so long as the interest rate on debt much higher financial leverage than do Ameri- is less than the risk-adjusted cost of equity can corporations (leverage can be defined in the normal case. Several questions follow: If several ways, perhaps the most common being financial leverage lowers costs of capital, why the ratio of debt to equity in a firm's capitaliza- don't U.S. firms emulate the Japanese by usihg tion). Table 53 compares debt/equity ratios for more debt in their capital structures? Wouldn't U.S. and Japanese electronics companies. The stockholders benefit from this choice by earn- reasons that corporations in Japan make ing higher returns? There are also potential tax benefits: since corporations can deduct interest For a standard introduction to...corporate finance, see J. C. paid on debt as an expense, but not dividends Van Horne. Fundamentals of Financial Management, 4th ed. paid to stockholders, would not greater use of (Englewood Cliffs. N.J.: Prentice -Hall, 1980). debt decrease Federal tax obligations and in-

Table 52.Internal and External Sources of Corporate Financings

Internal finance (reinvested profits, External finance Ratio of internal depreciation) Loans Securitiesb Totalto external finance United States 69.4% 12.4% 18.2% 30.6% 2.27 Japan 40.0 49.0 11.0 60.0 0.67 United Kingdom 51.4 10.3 38.3 48.6 1.06 West Germany 63.1 29.6'. 7.3 36.9 1.71 France 65.0 27.4 7.6. 35.0 1.86 1966.70. These Patterns have Probably not chanaed °really. 258 International Competitiveness in Electronics

Table 53.Total Debt-toEquity Ratios for than others. In one year, the funds remaining Selected U.S. and Japanese Electronics Firms after expenseshence available for distribution 1975 1979 to shareholders or for retention in the enter- United States prisemay be plentiful; in another, such mon- Advanced Micro Devices 81% 8% eys may be scarce or nonexistent. Stockholders Control Data Corp. (CDC) 38 20 are generally.believed to desire stable earnings Digital Equipment Corp. (DEC) 30 32 General Electric 41 25 from year to 'year, accepting greater variabili- Honeywell 65 32 ty in rate of return only if compensated by a IBM 4 17 higher average return. Intel 0 0 Motorola 28 30 In contrast to stockholders, who share in the National Semiconductor 25 37 RCA 106 125 ownership of the firm, creditors merely lend Texas Instruments 14 21 it funds; they generally have first claim on cash Japan° flow, as well as on the assets of a firm, and Fujitsu 200% 190% Hitachi 160 96 receive a "guaranteed" rate of returni.e., the Matsushita Electric 14 16, interest rate on bonds or other debt instru- Mitsubishi Electric 370 270 ments. While creditors seldom share in the first Nippon Electric Co. (NEC) 350 400 type of riskvariability in returnsthey may aThe financial data for Hitachi, Matsushita, and Toshibaas used by Chase Financial PolicyIncludes affiliated trading companies among the consolidated sometimes choose to subordinate their claims subsidiaries, while that for Fujitsu, MitsubiShi Electric, and NEC does not. See the Chase Financial Policy report cited in the source note below, p. 6.1. rather than force a firm into bankruptcy. In the SOURCES: United StaLS;Derived from annual reports; also "Financial Issues recent example of Braniff International, the in the Competitiveness of the U.S. Electronics Industry," report prepared for OTA by L. W. Borgman & Co. under contract No. 033.1550,0, airline's creditors several times allowed pay- pp. 52, 56. JapanDerived from data in "U.S. and Japanese Semicon- ductor Industries: A Financial Comparison," Chase Financial Policy ments of both principal and interest to be de- for the Semiconductor Industry Association, June 9, 1980, Appendix: ferred before Braniff finally entered bankrupt- Japanese Semiconductor Companies, Financial Statements and Sup. porting Schedules. cy. The Braniff case illustrates the second type crease aftertax profits? If so, isn't this another of riskloss of all or part of the investment it- reason to encourage U.S.- electronics firms to self, as well as loss of revenues from interest increase their leverage? (Japanese tax treatment payments or distributions of profits. This is a of interest payments is similar to. U.S. law in risk borne by both owners and creditors. But this respect.) At this point, the layperson might because creditors have first claim, they are think that Japanese firms have simply taken ad- more likely to recover at least part of their in- vantage of financing choices also open to vestment in the event of business failure. This American companies. is the reason interest rates on debt are generally lower than the risk-adjusted cost of stockholders' equity: holders of debt face lower risks Risk because they have first claim on assets. At the The answers to the questions above, and the same time they must accept a nominally fixed key to understanding the U.S. electronics in- rate of returngenerally lovirenthan that accru- dustry's unhappiness with Japanese financing ing to shareholders. (In fact, the effective rate practices", relate to a second aspect of finan- of return on bonds is not necessarily fixed, as cial decisionmaking risk. Investment deci- pointed out earlier, but this is not important sions inevitably involve risks for those who here.) supply fundswhether external funds or inter- The discussion above is necessarily schemat- nalbecause there can; e no certainty that fu- ic, and corporations- can avail themselves of ture cash flows will be sufficient to compen- other methods of financing, which fit into the sate investors. In essence, the risks borne by subordination ordering in various ways. But investors are of two types. First, cash flows are as a general rule, common stockholders come Ch. 7Financing: Its Role in Competitiveness in Electronics ' 259 . other creditors and investors have been paid. 2. Some Japanese investors are accepting This subordinated status makes shareholders risks for which they receive less compen- sensitive to the degree of leverage employed sationfor whatever reasonthan they by the firm; their exposure to risk increases desire. with higher leverage. Not only does more debt 3. Some classes of borrowers in Japan pay in the firm's capital structure tend to increase premiums for funds, these premiums the variability of returns to shareholders, but counterbalancing the low rates available added debt worsens their position in the event " to other borrowersor, alternatively, some of a forced_ liquidation.TypicallyT- common potential borrowers cannot get funds at all stockholders must be compensated through because of capital rationing. higher returnswhich can include capital ap- 4. Some risks which private investors in the preciationbefore they will accept the risk in- United States must bear are, in one way herent in greater leverage. or another, reduced for private investors As a consequence, adding more debt will not in Japan. necessarily lower a firm's cost of capita1,3 In- Each of these four possibilities will be briefly deed, neglecting tax effects, the choice of debt- examined. equity ratio, over rather wide ranges, should have little, if any, impact on capital costs. Even Risk Aversion Behaviors assuming no increase in interest rate as a firm borrows morewhich is not very realisticthe Financial and busine3s risks must be ab- lower costs of debt are generally offset by the sorbed within any system that operates to transfer funds from saverstocommercial bor- higher required returns to common sharehold- rowers. The presumption is that people have ers as leverage increases. Several cautions must an aversion to risk and, if they'are to accept be adde`d. While this conclusion is commonly such risks, they must be compensated by in- accepted as applying for U.S. capital markets, terest payments, capital,appreciation, or divi- it is not clear that it always .holds in the same dends on shares. Still, it is not necessarily true way in other countries. Furthermore, taxes do that all peopleor all economieswill exhibit matter, and the fact that interest payments identical patterns of risk aversion. Japanese in- lower a company's tax bill usually would argue vestors, for example, might demand less re- for adding to the proportion of debt in a firm's muneration for a given level of risk than capital structure. But Et some point more debt Americans. (It is also possible that the Japanese will be accompanied by higher interest rates, are less reluctant to postpone consumption since the debt itself becomes increasingly risky a possible explanation for the higher savings , for potential holders. rate mentioned earlier, with its tendency to With all of this said, how is it possible that force down interest levels.) Japanese companies can, on the average, employ debt-equity ratios markedly higher than Compensation for Risk American firms, without seeming-to bear high- The second possibility suggested above was er costs of both debt and equity? The usual re- that some individual or-institutional investors sponse holds that the Japanese financial system in Japan might be compelled to accept ,less differs from that in the United. States, and compensation than they desirei.e., less than forces that tend to raise the cost of capital as they would receive in a capital market that leverage increases are absent in Japan (or func- functioned differently. It does appear that Jap- tion differently than they do here). This implies anese bankswhich provide much of the cap- one or more of the following: italizationfor electronics firms through 1. Japanese investors exhibit risk aversion loansare accepting what are essentially quasi- cc-- esc 260 International Competitiveness in Electronics crated with equity. While banks can diversify Subsidization of Risk these risks by maintaining a portfolio of cor- The foregoing question is often answered, at porate loans, to the extent that such risks are least in part, by appeal to the fourth point systematic, diversification will be ineffective namely, that on some loans Japanese banks can ("systematic risk" is simply that which cannoi shift part of the risk to other parties. In partic- be reduced by diversification). The question re- ular, for loans to companies whose activities mains: Why do banks in Japan accept finan- are deemed to further national interests, the cial risks that would not be acceptable else- Japanese Government may effectively guaran- where in the world? This question is taken up tee the loan, at least to the extent of providing later in the chapter. protection against bankruptcy. Some observers, indeed, suggest that many loans by Japa- Preferential Treatment of nese banks are simply government loans Selected Borrowers passed throtigh the banking system. In this Some observers assert that "target" indus- view, some of the "normal" risks of debt fi- tries in Japan area selected to receive bank loans nancing are absorbed by the government rather at interest rates well below market levels.4 This than the banks; interest charges below market would imply, in the normal circumstance, that rates reflect government subsidization. other borrowers wil! pay higher than market rates; still Other potential borrowers might not As with the other points raised above, the be able to secure funds at all. The bias is usu- question of whether and to what extent the Jap- ally alleged to favor large firms at the expense anese Government subsidizes risk can be an- of the far greater number of small establish- swered, at least in part, by empirical evidence. ments in Japan, and to favor growth industries While the actual functioning of the country's even though such industries may, in the financial system is taken up in a later section, shorter term, offer rates of return both lower the Japanese economy is no different from and more variable than alternative invest- others in-that capital is a scarce resource allo- ments. In general, both semiconductors and cated by various mechanisms among an enor- computers would be classed as rapid growth, mous variety of investments. If the government long payout industriesin Japan as well as in or the banking community chooses to step in the United States. by selecting target sectors to receive capital at rates that were directly or indirectly subsi- If some borrowers are, in fact, favored with dized, the consequences are quite predictable. lower than market interest rates in Japan, this The target sectors would gain at the expense question follows: Why doesn't competition of the rest of the economyfor which credit among lenders force Japanese banks to allocate would normally have to be rationed. In other resources to the firms and industries promis- words, no country can subsidize all industrial ing the greatest returns, as in the United States? sectors simultaneously although manufacturing might be favored over agriculture, or the 4L. C. Thurow, for example, has claimed that Japanese inroads private sector over the public. In fact, there is into U.S. and world semiconductor markets are financed with no question that capital was allocated bythe funds provided by,the government-owned Bank of Japan: '.'But Japanese Government during the earlier post- the Japanese are entering this industry with massive amounts of debt capital ultimately lent by the Bank of Japan. Their aim war period; what is not soclear is whether is to jump directly into large-scale, capital-intensive techniques morethan remnants of these policies remain. of pioduction; proceed rabidly down the learning curve; sell at prices lower than those of the rest of the world; and capture most, if not all, of the market. If American industry limits its investments to those that can be financed by retained earnings, they Price Inflation and Banking Practices will simply be driven out of the semiconductor industry." See, ) "Prepared Statement of Dr. Lester C. Thurow," Productivity in nne_nnestinn that even themore sonhisti- Ch. 7Financing: Its Role in Competitiveness in Electronics 261 is the impact of inflation oc, international finan- of the cost advantages calculated for the much cial comparisons. The effects are too complex more heavily leveraged Japanese companies, to fully review here, but differing inflation rates But are they close in real, rather than nominal, among the world's economies are a major fac- terms? The answer depends entirely on long- tor in apparent differences in costs of capital. term inflationary expectations ,expectations The reason is that observed market interest which were probably considerably higher in rates depend in part on expectations by invest- the United States than in Japan during 1980.* ors with respect to price inflation. If expecta- The nominal interest rate differential favoring _ tions cliff& between a pair of countries, then the Japanese might well reverse, and favor the market interest rates will diverge from this United States, could the real rates- be com- cause alone. But to the extent that the diver- pared. gence in interest rates simply reflects the underlying inflation rates in the two economies, Differences in banking practices between the differences in costs of capital based on these two countries also affect the true costs of capi- interest rates are not "real." Only a difference tal. For instance, banks in japan typically de- in costs of capital after adjustments for infla- mand that greater compensating balances be tion would confer advantages in internation- kept on deposit against corporate loans.5 Be- al competition. cause the firm pays more for the funds it has borrowed than it receives on its deposits, this The difficulty is that future inflation can only practice raises the effective interest cost of the be projected; the presumption is that the mar- loan. Large compensating balances mean that ket mechanisms which set interest rates take the usual measure of financial leveragedebt- such projections into account. Interest levels equity ratiooverstates the true degree of lever- enter the calculation of capital costs for U.S. age. and Japanese electronics firms in table. 51 in at least two ways: 1) through the cost of equi- These observations on the effects of inflation ty computation, which is based on a "riskless" and compensating -balances emphasize the interest rate; and, 2) in the choice of interest complexity of cost of capital comparisons as rate for the cost of debt. The riskless interest applied to funds from external sources. They rate applies to investments fOr which the risk do not confirm or deny the general trends in borne by the lender can be considered negligi- table 51, though perhaps throwing doubt on the ble, at least in comparison to the risks of equi- magnitudf; of the differences resulting from the ty. Government notes, bonds, or bills are typi- Chase analysis. cal examples of "riskless" investments. 'The Chase study from which table 51 comes did attempt to compensate for varying inflationary expectations. The Japanese The riskless rates applied by Chase Financial cost of capital in yen was converted to a dollar cost using the Policy were: 10.2 percent for the United States, difference between the two interest rates cited above, said to represent "the premium investors require for receiving interest derived from the June 1980 Treasury bond rate; and principal payments in U.S. dollars rather than yen" (p. 7.7), and 9.0 percent for Japan, the yield on the most But by implying that the 1.2 percentage point difference widely traded debenture (a type of bond) on the represents dissimilar inflationary expectations, this prOcedure amounts to assuming that the riskless rates in the two countries, Japanese market-10-year issues of Nippon Tel- expressed in dollars. are equal. This seems unlikely: it would egraph and Telephone Public Corp. (NTT). The imply that the capital market in Japan is both efficient and per- analysis takes the degree of risk for these in- .fectly linked to that in the United States, neither of which is true. It is more reasonable to assume that differences in inflationary struments to be, if not zero, at least small and expectations were considerably larger than 1.2 percent in comparable in the two countfies.* mid-1980. 5y. Suzuki, Money and Banking in Contemporary Japan (New These two interest rates do not differ by Haven. Conn.: Yale University Press, 1980), p. 50. Japanese firms much; indeed, their closeness accounts for part often keep 25 percent or more of borrowed funds deposited with lending banks. Furthermore, banks are more likely to lend to firms that are already large depositors. In the United States. com- 262 International Competitiveness in Electronics

Effects of Financial t_everage capital advantage of about 1.9 percentage on Tax Paymehts points compared to firms in the United States. While the Japanese Government might sup- In fact, this example, overstates the advantage port electronics firms through a variety of cap- because the median figure (0.67) for Japanese ital and other subsidies, 014 study by Chase firms ignores the impact of absolute size. Financial Policy summarizied in table 51 is Hitachi and Matsushitawhich have less than based solely on a leverage argumenti.e., on median leverageare much larger than fhe the advantages of debt as a source of corporate others; Nippon Electric Co. (NEC), which has financing. In the absence of other sources of the highest debt-to-capital ratio-0.80is con- financial advantage, leverage provides lower siderably smaller. When the debt-to-capital capital costs primarily through its effects on values are weighted by total assets, the debt- corporate tax payments. Although these are not to-(total) capital ratio for the Japanese compa- trivial, the tax advantages that accrue to Japa- nies is 0.52, a result that is considerably af- nese firms as a result of high debt-equity ratios fected by Matsushita, which had negligible levr reduce their costs of capital by only a few erage. Using this figure for Japan. along with percentage points--probably less than 2com- the leverage value that Chase suggests in their pared to American firmS.. The reasons are out- study as a desirable target value for American firms desiring to reduce their costs of capital- .1i ned below. 0.33the comparison becomes: In order to isolate the effects on cost of capital stemming from tax shields on debt, assume Japan 0 52 (0.4)(0.1)=0.208 that interest rates. in the United States and United States .0.33 (0.48)(0.1)=0.01584 Japan are the samesay 10 percentbut that -Substracting gives 0 00496 or 0.496 %. corporate tax rates differ. For purposes of il- These two changes reduce the cost of capital lustration, use the nominal tax rates in the two advantage of the Japanese firms from 1.9 per- countries-48 percent for the United States, 40 centage points to only !-alf a point. This second percent for Japan.° For leverage, assume a ratio comparison is not necessarily either more or Of total debt to .total capital equal to 0.67 for less meaningful than the first; the lesson is that Japan and 0.17 for the the United States.* As tax advantages are quite sensitive to small a result of the tax shield created by leveraging, changes in leverage. The computation is far costs of capital would he lowered by: less sensitive to the tax rate itself. Japan 0 67 (0.4)(0.1)=0.0268 These examples show that, while the use of United States .0.17 (0.48)(0.1)=.6.00816 leverage does lower a firm's cost of capital, the Subtracting gives 0 01864 or1.864% effects are relatively small. A cost of capital 'Thatis.the tax shield created by greater lower by 2 percentage points might transla'e leverage would give Japanese firms a cost of to manufacturing costs lower by 1 percentage pointnot very significant. The difference that "This is the nominal rate for retained income in Japan; distrib- does result can be regarded as an implicit fi- uted profits are taxed at 39 percent. While nominal rates suf- nancial subsidy from the Japanese Government fice for illustration, they bear little resemblance to the taxes that via the tax system. (The question of special tax corporation:, actually pay after deductions, credit, depreciation allowances, etc. The -effective" tax rates in the two countries provisions for certain industries is independ- in the late 1970'stotal corporate taxes paid divided by total cor- ent.) porate-profitswereabout 37percent in the United States,29 percent in Japan. See H. Gourevitch, A. Wilson. and D. Culp, fax Rates in Atajor Industrial Countries: A Brief Comparison, Risk Absorption in Japan Congressional Research Service report No. 80-224 E. Dec 15, nitn, p.8. As mentioned earlier, Japan's banking sys- 'Tile 9.97 fieure is used at several mints in the summaryof tPrnnhcnrhcricicc nnrmnIlliacciimpri 1-11r chirp_ Ch. 7Financing: Its Role in Competitiveness in Electronics 263 fact the risks of bankruptcy, reorganization, tronics have been infrequentin part because and business failure are increased by the growth rates and cash flows have remained greater use of financial leverage in Japan. If so, high, allowing firms to service their debt-,---risk the frequency of failuresespecially during is clearly present. economic downturnsshould be higher than in countries like the United States where, on In the Japanese financial system, these risks the average, leverage is much lower. Indeed, tend to be shifted to, the banking community. bankruptcies in Japan have risen to rather high Banks assume quasi- equity positions by ilLcept- levels in times of general economic downturn. ing debt in hig,y leveraged firms. If a com- In 1977, for example, Japanese enterVises pany finds itself in financial difficulty, the failed in record numbers.' The rate of bank- banks are literally forced to take action aimed ruptcy that year was four times greater than at reorganization. The alternative is to proceed the comparable U.S. figure, and these failures,. with bankruptcy. When large corporations involved corporate liabilities of over $16 bil- have faced trouble, the choice, not surprising- lion, inure than Five times the 1977 level in the ly, has often been restructuringsometimes ac- United States. While 1977 remains the peak companied by infusions of even more funds. year in terms of both number of business fail- Typically the banks have forced a complete ures and total liabilities, in every year since reassessment of corporate strategy; not infre- 1976 Japan has experienced more than 15,000 quently the ailing firm's executives have been .bankruptcies (excluding small businesses) with replaced by a bank-appointed managerial team. -total liabilities exceeding 2 trillion yen (roughly Sometimes observers in the United States $10 billion)." Although bankruptcies in elec- conclude that these interventions by Japanese 'G. R. Saxonhouse, "Industrial Restructuring in Japan," Journal banks serve to reduce risks to businesses, or of Japan:?se Studies. vol. 5, 1978, p. 273. Elsewhere Saxonhouse risks to the banks, or both: To believe this im- states: "Debt- equity ratios which are four or five times the Amer- plies believing that bankers are on the average ican level result in bankruptcy rates which are also four or fine times the American level. Large Japanese firms cio go bankrupt. wiser than managers of industrial corpora- In recent years /alma has had two close to $1 billion in liabil- tions. In fact, these interventions do not lessen ities bankruptcies." See "Statembnt of Gary R. Saxonhouse financial risks, but are caused by the much Before the f:ouse Foreign Affairs Committee, Subcommittee on greater exposure of the banks. Such interven- Asian and Pacific Affairs and Subcommittee on International .Economic Policy and Trade," Oct. 1, 1980. tions are less common in the United States be- 6Only firms with liabilities of more than 10 million yen are cause American banks do not provide as great included. Figures for 1968 through 1980 can be found in "Japan a' fraction of corporate financing. 1981An International Comparison,' Japan Report,Joint Publications Research Service JPRS L/10760. Aug. 24, 1982, p: 7. those for 1981 in "Corporate Failures in Japan Last Year Fell 1.5% to 17,610," 1311 Street Journal. Jan. 15, 1932, p. 28.

Other Factors in Costs of Financing

Size and Diversification These firms have generally depended more Many of the leading international clmpeti- heavily on one or a few product lines than their tors in electroi, are larp,,..1, diversified c--,m- competitors in japan. Size and diversification, panies (ch. 4).ins is the case for Ainerican affect capital costs quite directly, with the ad- corporations like GE or IBM, European man- vantages going to big companies with broad ufacturers like Philips or Siemens, and many product lines. Such firms can absorb and IaDanese electronics firms. But other U.S. en- spread risks more effectively. 264 IntQrnational Competitiveness in Electronics own repayment. Diversified companies exhibit important such differences might be interna- more stability, hence lenders are willing to tionally. Over the years, fast-growing and prof- supply funds at lower rates of interest. Large itable small- and medium-sized firms have co- diversified firms are also, on average, evaluated existed with the giants of the U.S. industry as better risks by bond rating companies like indeed have often Outstripped them; size and Moody's 'or Standard & Poor's. Texas Instru- diversity did not appear to give RCA (ir GE ments taps a lower cost bond market than, say, much help in computers or semiconduciurs. Advanced Micro Devices; IBM lower than Dig- In dynamic, technologically advancing indus- ital Equipment Corp.; GE lower than Zenith. tries, other competitive forces far outweigh The conclusion is: costs of capital will be small differences in interest rates on bonds or higher for U.S. as opposed to foreign electron- bank loans. ics companies when the American firms are significantly smaller and less diversified. SavingsRates10 There is a great deal of variation in the sizes As mentioned earlier, international differ- of firms within the electronics industries of the ences in savings rates could affect relative costs United States, Japan, and the European coun- of capital. Within a closed economy, a high rate tries. Nevertheless, in Japan and Europe it is of savings creating an ample supply of invest- primarily larger companies that are active on ment funds will tend to depress interest rates. a world basis, more the exception than the rule Given the international linkages among capital for companies the size of Nixdorf (West Ger- markets, this simple /argument is not by itself many) or Oki Electric (Japan) to be strong inter- sufficient to relate savings levels to interest national competitors. In the United States; the rates, but may still have weight. For reasons situations of companies like Mostek, Fairchild, that are poorly understood, the savings rate in or Intersil have changed dramatically as a re- Japan has been extraordinarily high for many sult of their acquisitions by much larger con- years. Table54gives data on household sav- cerns. Still, the United States remains unique ings for the1976-79period; the figures for all in the number of relatively small electronics five countries have remained fairly constant firms that seek to compete worldwide, includ- over the past two decades. High savings rates .ing many of the new startups making semicon- have characterized both the corporate and ductor and computer products. household sectors in Japan, but it is personal In Japan, the evidence suggests that the gov- savings that have been most surprising in view ernment and banking system as in a number of interest rates that have been below prevail- of other countriesovertly discriminate among borrowers on tilt: basis of size.° But even if for- "For a general introduction to savings rates. see C. Elwell. Investment and Saving: The Requisites for E6onomic Growth. i..;;n capital markets were identical in every Congressional Research Service report No. 81-207 E. Nov. 15. respect to American marketb, and operated 1981. with no more government intervention than in this country, the larger average size of the ma= jor foreign competitorsparticularly in semi- Table 54.Household Savings Rates in conductorswould give them at least a small Several Industrial Countries' relative advantage. Average savings rate, On the other hand, larger semiconductor and 1976-79a United States 5.8% computer firms in the United States clearly Japan 21.1 have not reaped great benefits from their own West Germany 13.4 ability to tap somewhat lower cost sources of France 13.5 extern alcanital. Thus. one can Question how United Kingdom 8.3 Ch. 7 Financing: Its Role in Competitiveness in Electronics 265

ing rates of inflationas shown later (see table est, high growth strategy in the earlier postwar 64) and have also been significantly lower for years. Some observers go so far as to imply that savings accounts than for alternative invest- no resource allocation problems exist in Japan ments such as bonds." Despite this, Japanese because of a virtual glut in investment funds.12 households carry the largest portion Of their Often such assertions are linked with the target savings as cash or deposits; the contrast with industry argument mentioned earlier. If true, behavior in the United Statesillustrated in this would mean that Japan's chosen industries table 55is striking. Householders in Japan enjoy low financial costs for reasons entirely keep ai substantially lower portion of their as- apart from their high debt-equity ratios. sets in corporate stocks. But capital markets should not be viewed This extraordinarily high savings ratehalf from only one side. In tit, case, there are po- again as much as in France or West Germany, tential Impacts on the demand side as well as and nearly four times that in the United States the supply side. Growth affects both the de- (table 54)--when combined with closely con- mand for funds and the supply. Businessmen trolled savings institutions, is often said to pro- foresee abundant saleopportunities in ex- duce artificially low interest rates on loans to panding economies ar...,est to meet the new Japanese businesses. The argument is essen- demand. This places heavy pressures on capital tially on the supply side: low interest yet abun- markets. On the supply side of these markets, dant sources of capital have allowed Japanese individual consumers may experience rapid corporations to expand rapidly while maintain- growth in real income but adjust their con- 'ing low prices, especially in export markets. sumption habits more slowlymeanwhile sav- / As one consequence of rapid growth, the coming their unspent income. Thus, a case can be / panies would enjoy economies of scale, along made that Japan's high savings rate is a con- / with modern, highly productiVe manufacturing sequence rather than a cause of rapid econom- facilities. Past this point, low capital costs ic growthi-i.e., that income growth has out- would be less of an advantage but Japan's stripped cnsumption. firms could by then compete Comfortably on In fact, neither demand-side nor supply-side other grounds. perspectives tell the whole story; both are Numerous variations on this theme have needed. din/In/ Japan,Japan inflation-adjusted interest been propounded many stemming from the rates on /savings have recently been comparable Japanese Government's well-known low inter- to ratein the United Statestable 56. This table n rates of return available on "See, for example. H. C. Wallich and M. I. Wallich. "Bank- long-teirn government bonds in both countries ing and Finance," Asia's New Giant, H. Patrick and H. Rosov- sky (eds.) (Washington. D.C.: Brookings Institution. 19761. pp. (a rather/arbitrary choice) to the respective in- 260-26 I . ation rates, the difference being "real" rate of return. As the table shows, since 1S78 inves- Table 55.Distribution of Household Assets tors inijapan have received higher real returns in the United States and Japan, 1978 than,Viose in the United Sates. This suggests that lqrtificially depressed interest rates have United States Japan not recently been a source of abnormally low Cash, plus demand and savings deposits 39.2% 68.7 % Bonds 9.6 8.1 '=Response of W. Rapp, Technology Trade, hearings, Commit- Stocks, including mutual funds 23.5 10.0 tee on Science and Technology, Committee ot, Interstate an'd Life insurance 5.6° 12.6 Foreign Commerce. and Subcommittee on Inteinational Trade, Othera 22.1 0.6 Investment and Monetary Policy of the Committee on Banking, 100.0% 100.0% Finance and Urban Affairs, House of Representatives, and House _ Task Force on Industrial Innovation. June 24. 25. 26. 1980. p. 266 International Competitiveness in Electronics --,--

Table 56.-Inflation-Adjusted Rates of Return in the United States and Japan

United States Japan Longterm InflationReal rate GovernmentInflationReal rate government rate rates of return bond rate rate° of return 1975 8.2% 9.2% -1.0% 9.2% 11.9% -2.7°A 1976 7.9 5.8 2.1 8.7 9.3 -0.6 1977 7.7 6.5 1.2 7.3 8.1 --0.8 1978 8.5 7.5 1.0 6.1 3.8 2.3 1979 9.3 11.3 --2.0 7.7 3.6 4.1 1980 11.4 13.5 -2.1 9.2 8.0 1.2 1981 13.7 10.4 3.3 8.7 4.9 3.8 8Elased on consumer price Indexes. SOURCE. Based on data from International Financial Statistics, International Monetary Fund, various issues. costs of capital for Japanese electronics firms. nese financial system is described in more de- In other words, there is little evidence of any tail. Still, this practice seems largely to have across-the-board supply-side stimulus that disappeared; government capital allocations do might stem froth an ability by Japan's Govern- not now seem to provide borrowers in Japan ment to persuade people to save even at rela- with a notable edge over U.S. competitors. Gov- tively unattractive interest rates. As the table ernment financial institutions accounted for demonstrates, corporate (and recently govern- about 30 percent of all corporate loans in 1 -1502 ment) demands for funds have driven up in- but as of the end.of 1980 their share. had fallen terest rates in Japan much as in other devel- to 13 percent; during the 1970's, loans from oped economies. government institutions accounted for only This does not dispense with the possibility about 5 percent of total capital flowing into that the Japanese Government intervenes in Japanese industry." The percentage is even capital markets to subsidize target industries. lower in the electrical machinery sector, which Certain industries-or firms--could be favored includes electronics; here, government institu- by government repayment guarantees to lend- tions accounted for only 8.2 percent of out- ing banks, effectively transferring the risk of standing loans as of December 1.980. Nonethe- default from the commercial banking system less, some observers continue to hold that the to the public at large. Alternatively, the Bank Japanese Government effectively socializes the of Japan could be encouraged to rediscount ,risk of corporate borrowing.15 bank loans at favorable rates. Finally, through the postal savings system the government itself Costs of Capital for Electronics Firms takes in about a third of all savings deposits:" in the United States and These funds could be channeled to favored in- Japan: Summary dustries at interest rates largely determined by government fiat. It does seem clear that Japanese electronics manufacturers can get external capital at some- It is quite true that in early post-World War what lower rates than their counterparts in the II Japan, allocations of investment funds were -United States. But at present, this capital cost more a function of administrative control than advantage, in inflation-adjusted terms, appears relative interest rates; favored industries had access to funds at subsidized rates of interest, "E. Lincoln, "The Japanese Government's Role in Business while personal savers and small businesses Financing,"JEIReport,Japan Economic Institute. Washington. bore the brunt of the costs. This point is taken D.C.. Jan. 8. 1982, p. 12. sin infor r-vc,ormii ofrIlrflIrcs r1f tb.hna- 157Catratrihara P Pc.lrirnan and V Pararla ThP la na mPcp Ch. 7Financing: Its Role in Competitiveness in Electronics 267

age point margin applies for the total investment in a production facilitywhich is unlikely. Even then, the result-would be a pdtential manufacturing cost difference of about 11/2 percent, and might translate into a price difference of the same magnitude. Smaller capital cost differences would reduce this advantage com- mensurately. Such a 2 or 3 percentage point advantage in capital costs represents an average over many firms in both Japan and the United States. The difference in average costs of capital in the two countries is smaller than the differences in capital costs among competing electronics firms within either country. Although table 51 does not accurately portray cost of capital dif- /44 ferences between the two countriesty will il- 10 lustrate this point if taken as representative of firm-to-firm differentials. The range in costs of capital for the eight U.S. semiconductor firms listed in the table is 7 percentage points, that for the six Japanese manufacturers nearly .10 percentage points. The interfirm differences-- and the resulting competitive advantages or Photo credit: GCA Corp. -handicapsare much larger than OTA's esti- Wafer processing equipment for making. mate of the average differential between the integrated circuits two countries. Relative availability of capital for electronics to be small- -certainly less than 5 percentage firms in the United States and Japan has great- points. The sources of this advantage are multi- er potential impact. Favored Japanese electron- ple: government policies in Japan that have the ics firms seem to have little difficulty in acquir- effect of subsidizing interest rates for favored ing funds for expansion and modernization. In investments no doubt continue to account for contrast, some U.S. companiesparticularly a good deal of the margin. Except for .the tax- the smaller onesbelieve themselves subject shielding efiects of their higher leverage, Japa- to capital constraints. That is, they may find nese electronics companies do not have access themselves unable to raise as much capital as to cheaper capital because of the preference they would like at any reasonable cost. for bank loans in their capital structures. While greater leverage transfers business risks to the Capital availability is a subject left to a later banking system, it does not allow the Japanese section, but note one major difference between to avoid risks. electronics and other American industries that make this same complaint. Some domeitic steel A differenc:, in financing costs of 2 or 3 companies, for instance, have had difficulty percentage points is nontrivial but will not tracting external funding because of their L drastically alter the market postures of com- ability to convince prospective investors of the peting firms. For purposes of comparison, as- industry's potential for growth and future prof- ellTeStl n ,ninn -.r _._Tit rl-1 268 International Competitiveness in Electronics capitalparticularly compared to existing net than with absolute costs of capital, are taken worthrequired to maintain their share in a up below. The next section extends the com- rapidly expanding worldwide market. In semi- parative treatment of financing to several other conductor manufacturing, rising capital inten- countries, while carrying forward the U.S. - sity compounds the difficulty. These matters, Japancomparison. concerned more with the dynamics of growth

Financial Structure: An International Comparison

Many countries are attempting 'to build com- New corporate startups have been frequent petitive electronics industries because they be- during the industry's postwar histo-ynot only lieve them essential for a growing and healthy in semiconductors and computers, but in many, economy. Government assistance, often finan- other product lines. Hewlett-Packarda div9r,, cial, has flowed to electfonics companies: sified manufacturer of test and laboratorY GreatBritainprovidesexplicitsubsidies; equipment, calculators and computers, and a France has combined subsidies with trade pro- leader in integrated circuit. (IC) technology tection. A number of the rapidly developing through its captive operationsgot its start just countries are following suit, as outlined in before the war in a garage in Palo Alto, Calif.17 chapter 10. Still, Japan remains the primary In many respects typical of the firms that later competitor in electronics, and its financial sys- gave this region the name Silicon Valley, the tem is treated in much more detail than ihat company's founders began by designing its of France or West- Germany, the two' other first product themselvesan audio oscillator countries examined below. The focus on semi- supplied to Walt Disney Studios for the pro- conductors continues because U.S. firms in duction Fantaia. this oart of the industry have fae6d the most pronounced financing problems. Venture Capital Funding rapid expansion is a challenge that Businesses typically draw on far different semiconductor companies share with manu- sources of funds in their early stages of devel- facturers of small computers and peripherals, opment than later, progressing through a fairly software firms, and new entrants in other por- predictable sequence as they grow and mature. tions of the industry; Atari, the manufacturer This progression, which illuminates some of of video games, has reputedly been the fastest the unique characteristics of U.S. capital mar- growing technology-based company in history, , kets, is rather different from that in other coun- while one firm making game cartridges saw its tries. For purposes of illustration, consider a sales grow 1.000 percent (to $50 million) dur- startup firm with origins like those of Hewlett- ing 1981.16 In order to remain competitive, Packard or the many semiconductor manufac- firms in such markets must be able to finance turers that sprang up during the 1960's. Often growth at _rates that are literally explosive. these enterprises were formed by small groups of engineers and managers spinning off from a somewhat older company with the aid of United States funds from the venture capital market. The process is not unique te.., the semiconductor in- Sources of financing for American electron- dustry: Control Data Corp. was founded in ics companies vary depending largely on their 1958 by a grutip of ex-Univac. employees. While Ch. 7Financing: ifs Role in Competitiveness in Electronics 269 many new ventures in microelectronics, com- ated by the Small Business Investment Act of puter peripherals, and software have been es- 1958. Although. most SBICs concentrate on tablished since 1980.18 neighborhood businesses, a few have national Where rapid. growth creates expectations of outlooks. Venture capital partnerships and high returns, capitalization for new companies funds have also blossomed in recent years; the often comes from specialized financiers who U.S. venture capital industry now includes provide equity funds to the venture capital mar- about 600 firms and should continue to expand ket. Along with electronics, biotechnology is as a consequence of new ERISA (Employee Re- a current example. In such cases, ownership tirement Income Security Act of 1974) rules is typically shared among venture capital orga- allowing "prudent" participation of pension nizations and the founders of the firm. Stock funds in venture activities." In rare circum- options have been a common means of attract- stances funds, be raised through public ing talented individuals to startups, and have stock offerings, but this avenue is more likely been used to build handsome 'compensation to be ailable later in development. packages for key executives or technical spe- Verdure capital markets are highly cyclical. cialists without cutting too decply into cash One influence has been taxation of capital flow. gains. In general, high taxes on capital gains Annual returns of 25 to 50 percent over a discourage potential investors. Table 57 sum- period of perhaps 5 years are typical goals of marizes the results of a study prepared for the institutional venture capital organizations. In National Venture Capital Association, together the past, wealthy individuals of families some- with more recent data, that bears on this point. times founded private corporations for seek- The maximum tax on capital gains in the ing new and riskybut potentially highly prof- United States was reduced from 49 percent to itableinvestments. Today, corporate venture less than 30 percent in 1978. Although total capital organizations are also prominent venture investments rose dramatically, such subsidiaries of larger companies seeking to di- trends do not prove a cause-and-effect relation- versify. Corporate venture funding is more like- ship. They do suggest that the tax revision was ly to go toward second- cr third-round financ- a powerful contributing fal:Aor in the upswing. ing of young companies than to new startups, At the same time, a variety of other forces and investments tend to be larger than thoSe affect the ups and downs of venture funding. of independent venture capital organizations. The cyclicality reflects the confidence of poten- Sometimes eventual ownership is an objective; tial investors on the supply side and of poten- in other cases corporate venture capitalists are tial entrepreneurs on the demand side concern- simply seeking capital appreciation. Occasion- ing prospects for the economy and the propi- ally the funding organization provides capital tiousness of risky new undertakings. The tim- largely to gain proprietary technology. This has ing of startups depends on more than economic been the apparent goal of investments in U.S. conditions. Venture organizations look care- electronics firms by a number of foreign com- fully at the abilities. of a new firm's leaders; panies. Both Siemens (West Germany) and Fu- both capitalists and managers look for "tech- jitsu (Japan) have invested in this way. Siemens nological -windows" that offer unusual oppor- owns 20 percent of Advanced Micro Devizies, Fujitsu 26 percent of Amdahl, a leader in technology for large computers. An alternative source of venture funding, the Small P.:sinacc Invactmant rn (SRIC.1 wag rra- 190ne estimate of the Ind ictnt hrnalcrinwn is ac fnlIntvc. nriva to 270 International Competitiveness in Electronics

Table 57.Aggregate U.S. Venture-investment Activity (millions of dollars) Higher round investments New commitments (prior commitments) Totals 1977 Amount $56 $28 $84 Number of investments 126 126 .252 1978 'Amount $92 $31 $123 Number of investments 196 150 346 1979' Amount -145 $65 $210 Number of investments 29 248 538 1980 Amount NA NA 1981 Amount $500 $900 $1,400 NA - not available. 1379 ::.ota annualized from first 6 months. Data taken from 55 respondents. To OTA'sknowieige,no fulty comparable data venture Investments for later years we available. SOURCES 1977-79"Financial Issues in the Competitiveness of the U.S. Electronics Industry," report prepared for OTA by L.W. Bergman & co.under contract No. 033-1660.0, p. 9, quoting from "Survey of Ventu:e Capital Investment, 1977.1979," prepared for the National Venture Capital Association by D. J. Brophy and P. L. Schaefer o1 the Unive sityofMIchigin. 1980, -Government.industty Cooperation Can Enhance the Venture Capital Process, GAO/AFMD-82-35 (Washington, D.C.: Genert,1 Accounting Office, Aug. 12, 1982). p. 3. 1981J. W. Dizard. "Do We Have Too Many Venture Capitalists? Fortune, Oct. 4, 1982, p. 106. tunities. Some of the upsurge in investments ture capital of pbrhaps $5 billion to$6billion, in table 57 is related to booming interest in ap- pales alongside other sources of capital for U.S. plications of microprocessors; of the new ven- business and industry: bank loans, $230 billion; ture capital deals nationally since1979,per other short-term debt, $250 billion; corporate baps half have been in electronics or closely bonds, $490 billion; commercial mortgages, related faelds.20 At the peak of the most recent $280 billion; equities, $1.3 trillion.22 cyclei.e., rnid-1981---many observbrs of venture capital. markets concluded that-entrepre- Costs of Entry neurs wue having an easy time finding start,: Although a substanti it fraction of new ven- up fonds; some claimed that the supply of venture investments continue to flow into electron- ture capital considerably exceeded demand ics, in some segments of the industry entry during1981.,and that poor risks were being costs are becoming prohibitive. Among the ex- financed.21In more normal times, potential ceptions are firms able to generate cash flows startups may face a long and arduous search in other lines of business. NCR is an example: for Capital. an established manufacturer of computers and Table 57 will serve to illustrate another point: other business equipment, the firm had made venture capital makes only a small contribu- ICs exclt:ilrtly for internal consumption. In tion to the over-all funding needs of American 1981 NCR announced plans to produce semi- industry. Annual venture financing at some- custom logic circuits and certain kinds of thing over $1 billion, and a total pool of yen-. memory chips to be sold on the outside, becoming one of the few new entrants in the merchant market contemplating a fairly broad IFsues in the Competitiveness of the U.S. Elec. indost,T," report prepared for OTA by L. W. Borgmrin. product line.23 The reason is straightforward. & Co. wirier contract No. 033-1550.0, p. "Startup Fever is Costs for establishing a new seinicbnductor Spreading," op. cit. "Ser:i, for example, A. Follzick, "Few Places for Venture Capital: 22"The Perilous Hunt for Financing," op. cit. The estimates Funds Outpace Investment Opportunities," New York Times, are totals outstanding at the end of 1981. June 17, 1081. p. Dl; J. Levine. "Once Again, It's A Buyer's Mar- 2"A Surprise NCR Leap Into the Chip Market," Business ket,"Venture,June 1'482, p. 80. Week July 13, 1031. p. 22. ..

2 Ch. 7Finarr:ing: Its Role in Competitiveness in Electronics 271 firm have risen by a factor of 10 over the past Entry barriers can be higher still in main- decade; as figure 50 indicates, the end is not frame computers, where the new firms in re- in sight. The rapid increases in capital inten- cent years have entered with plug-compatible sity shown in the figure stem largely from the machinesAmdahl, Magnuson, in 1981 Tril- more expensive prodtiction equipment re- ogy. The one exception during the 1970's was quired for current-generation ICs (ch. 3, espe- Cray, established with venture funding to build cially table 2). specialized supercomputers. An added hurdle Rising entry costs are one reason why the stems from the preference by many customers 1980-83 group of semiconductor startups have for leasing rather than purchasing large com- picked narrow market niches rather than at- puters. Financing leases is a severe strain on tempting to compete in a broad range of prod- smaller companies; lease cancellations were ucts. Examples include: custom chips, or in one of the immediate reasons that Itel, a con- some cases just design services; specialized glomerate that had entered the plug-compatible device families .such as linear ICs or program- mainframe business, declared bankruptcy in mable logic arrays; and, in one case, gallium 1981. arsenide circuits. While entry via nichemar- Leasing has been a major part of IBM's strat- kets is the usual pattern in this ancilAher indus- egy in mainframes; competing firmsnone of tries, none of the semiconductor startups ap- which has assets near IBM'sface a major con- pear to be aiming at the mass produced mem- straint in financing leases.25 Not only are they ory or microp"ocessor markets: the most recent limited by their smaller size in securing funds new entrant to attempt this was Inroos, estab- at rates comparable to IBM's cost of capita, lished in 1978 with the aid of $90 million in but the risks of competing with IBM are large direct investment by the British Government.24 and v " L-nownitel's failure presumably add- =' UK 13oard to Cut Stdko in inmos," Electronic News, Mar. ing to the concerns of potential lenders. With- 29, 1982. p. 54. The National Enterprise Board, which has had- out a source of particular advantage such as considerable autonomy to fund British industry (ch. 10), provided Amdahl or Cray get from their reputations as' 50 million pounds. with Limos receiving an equal amount in loan. guarantees and grants for specific projects. technological leaders, cost and-availability of capital will remain formidable barriers for en- 50.Increase in Capital Costs for try into the mainframe computer market. High-Vailmelc.tegrated Circuit Production Line Si r--- Entry costs are also daunting at the high per- bdi, formance end of the minicomputer industry, though firms such as Prime (1972) and Tandem (1974) did :begin operations during the past decade. The microcomputer segment has still seemed attractive; the entry of IBM into the personal computer-market at the end of 1981 $100 has not seemed to dampen the enthusiasm of 2 million prospective entrepreneurs and venture capi, talists. Early Growth In the startup stage, equity capital from venture or other sources goes to purchase or lease S10 'plant and equipment and cover the initial ex- million 1970 penses of developing, manufacturing, and mar- 1975 , 1980 1985 1990 keting the first products. Later, more familiar Year

SOURCE R W. Broderson. "Signal Processing Using MOS.VLS itichr.olc.gy," VLSI Electronics: Microstructure Science. vol. 2. N. Einspruch J. T. Soina, The.Compnter Industry (Lexington, Mass.: Lex- (ed.) (Ne?. York: Academic Press. 1981), p. 203. ington Books; 1976), p. 4L 272 International Competitiveness in Electronics financial markets can be tapped. Exterrel fi- point, electronics companies exhibit financing naming is cical at this stage: the Company patterns that depend on the preferences of may be growing rapidly, with production out- owners and managers, as well as opportunities stripping sales as inventories build and distri- in relevant capital markets. Some firms offer bution channels are filled. New firms often new equity shares to the public; others issue operate below their break-even points for a shares but limit purchases to their own execu- number of years, and cash flow problems are tives or employees. Some sell bonds, often in common. addition to equity, to add leverage to the capital structure. But while financing patterns differ, Once sales have begun, local banks Ivill nor- they share a characteristic common to most of mally provide short-term loans up to about 80 U.S. industry: American electronics firms percent of net receivables, this amount being typically attempt to finance expansion inter- rolled overi.e.. the loans _rewritten at prevail- nally, even when this delays dividends inde- ing interest ratesevery 3 months or so. Long- finitely. Only if self-generated sources prove er term financing may come from incremental inadequate do companies enter external venture capital commitments; many ventuye markets. organizations prefer to participate in second- or third-round financing because they can bet- Internal and External Sources of Funds ter evaluate a company's prospects once it has products to show. While at this stage litnited Table 58 illustrates the extent to which Amer- public offerings may also be possible, stock ican semiconductor firms rely on internal sales to the general public have been less com- fundsi.e., depreciation and retained earnings. mon in electronics than in other industries. When the companies listed in table 58 have Many electronics manufacturers have been resorted to external financing, this has ranged able to finance quite rapid growth through -re- from virtually all capital stock (Intel) to substan- tained eavnings and employee stock option and tial amounts of debt (National). purchase plans: Indeed, the managers of elec- Although both new and established firms in tronics companies started by individual entre- the U.S. electronics industry will no doubt wish preneurs or small groups have often shunned to continue relying on internal funds, a number external equity markets to avoid stock dilution of factors converging during the early 1980's and the loss of close control. foretell financial dilemmas for some compa- When sale reach annual levels in the :vicinity nies. Among those common at least to manu- of $10 mi .. credit lines typicallyecome facturers of semiconductors are: more regi...... zed. Revolving credit a d term 1. Growth in unit sales over the coming years loans provide short-term financing. In addition, ray exceed even the .apid rates of the pre- banks will generally extend lease credit for :;ous decade. capital equipmentparticularly helpful in elec- 2. Revenue growth will continue to trail tronics because it reduce: the pressure to raise growth in unit sales as manufacturing funds for purchasing equ:,:nent ate time when costs and sales pricesIdecline. Historically, long-term investment requirements are ex- semiconductor prices are driven rapidly panding rapidly. Because capital equipment downward as cosis fall because of learn- Can quickly become obsolete, staying at the ing curve phenomena. The sharp drops in forefront of the technology can strain re- memory chip prices during 1981when sources. On the other hand, for firms with ade- 16K RAM (randorit .:ccess memory) prices quate cash flows, rapid obsolescence means fell from $4 each to about $17-a1' e sympto- short writeoff cycles and tax savings from matic. :depreciation. 3. Capital intel -ity will continue to rise be- In any case, firms with growth patterns that cause new pro: ..iction equipmente.g., for take them beyond the $10 million level find fine-line lithographyis much more ex- capital-more abundant-and-less-costly. At-this pensive-than-in-the-past- Ch. 7Financing: ItS Role in Competitiveness in Electrorics 273

Table 58.Internally Generated Funds as a Percent of Total Capital From All Sources

xas National Advanced Yea' Iriirunients Semiconductora Intel Micro Devices 1974 L % 76% 89% 93% 1976 79 82 79 96 1978 87 97 87 67

1980 65 60 46 . 71 aFiscal year SOURCES 1974-78 "Financiat Issues in the CompetitivAness of the U.S Electronics Industry," report prepared for OTA by L Vi. Borgman & Co. under contract No. 033-0550.0. p. 31. 1980 Annu al reports.

4.Engineering and design costs are also es- industry sources expecting considerably higher calating, due to the increasing complex- figures. Such estimates compare with capital ity of ICs. expenditures totaling $4.5 billion during the 5.Global competition, particularly from the 1970's.26 Japanese, is becoming more intense, and will be based on low prices to an even The changing character of semiconductor greater extent than in the past. Although production and marketing is not unique. forward pricing in anticipation of learn- Smaller American manufacturers of computing economies has been characteristic ers, as well as peripherals, are confronting even of competition among domestic more intense foreign competition at a time firms, pressures from the rapidly expand- when developing technologies are placing ing Japanese industry may cut still fur- severe demands on their financing capabilities: thur into sales revenues. New firms are entering markets for peripherals 6. Competition is also forcing companies to such as terminals and disk drives designed to pay more attention to quality and reliabil- be compatible with the products of established companies. Microcomputer applications of all ity, requiring costly test equipment. As IC sorts are on the rise. Computer software is one designs increase in complexity, testing of the most rapidly growing portions of the in- procedures become more time-consuming dustry. New entrants have often depended on and expensive. venture capital for their original financing, Figure 51, comparing capital spending rates andagain like semiconductor manufacturin the United States and japan over the past ersfollowed growth patterns relying on inter-_ few years, illustrates the rise-in capital inten- nally generated funds supplemented with lim- sity. Capital spending in both countries fell ited amounts of debt. , sharply in 1975 when the market for semiconductors slumped; since that time, the U.S. tread Parallels exist even in consumer electronics. has been steadily upward. Japanese spending A good deal of the production of established rates have been higher because they have been productsradios, televisions, audio equip- adding capacity faster. menthas moved abroad, taken by Far Eastern While capital spending in the United States 2°''Hungry for Capital-to Sustain the. Boom," Business Week. June 1. 1981. p. 74. J. F. r y of Texas Instruments has estimated averaged around 10 percent of sales during the spending at $25 billion.to .-Z,15 billion for the decade of the 1980's, 1970's; rates in the last3years have been while G. Moore of Intel v..,,ices the figure much higher, in the significantly great:(fig. 51). Continued in- range of $65 billton. An estimate of $30 bfllion results if sales are projected through 1990 based on the trend since 1975 (ch. crease will be difficult for many merchant 4), with capital spending assumed to remain at 13.7 percent of firms without substantial outside funding. sales, the average over the past 3 years. Spending rates may Capital needs of U.S. semiconductor manu- risesome predi6tions point to 15 percent of sales in coming years (see E. Williams. "Electronic Compo:::ents," Financial facturers during the current decade will prob- Times. Mar. 5, 1982, sec. III, p..1)but the total is much more ably be in the range of $30 billion, with some sensitive to the sales projection titan to the capital sperciinc rate._

276 Cornce:Itpieness in Electronics

Figure 51.Rates of Capital Spending by U.S. and Japanese Semiconductor Firms

t Japanese averages

...... / .... ,s -, r %or .. 'd'

J .197 1975 197f 1977 1978 1979 1980 198 Year

ndnut.. 1`+73-137'4 wniqritml ,"144eraye e.! 12 manutac!ur,r5 19731979 11 Manutanturnrs 1980 981 t4480 '7444,3t44,1 ter 11 U S mirdtant serreCnndUCtor manutantitrers. t,:j11 estimated. 1973 1979..L.IPJ, Fall '-iUnited States 1y 1377 E3uau 01 Census, 1978-1981, Departrri4)411 of Corrirnerce4Ind Semiconductor Industry A' lion Japan )1..,8').).0)111.:11 PUt)811)))1,081i,1980) 1; 203. 1980, 1981. JapanECOI70:771CJournal

competitors or transferred offshore by U.S. in consumer products; financing could become co-rtirtanies. But the broader market for coo- a problem here too if small firmsneed to :;up- sumer electronics remains dynamic, as the export rapid expansion. ample of video games showed. Along with video cassette and disk players, .as well as -home Financing Startups and Growth computers. these are precursors of an array of in 0.:3 Coming Decade electronics-based innovations for personal and But why should future financing be trouble- home use to be introduced over the next two se73e for an industry whichby all indications decades. Many of these products will be high can expect expanding markets overseas as technology items made possible by advdnces well as at home? Not all. observers believe the elsewhere in the industry. Some may come problems will be that-great; those who do gen- from small companies or-ganized by entrepre- erally point to a pair of related concerns:. neurs with strong technical backgrounds, as has happened with semiconductors, micro- 1. it may not.be possible to finance growth .computers, and software. Unquestionably, from internal sources in the proportions ______fierce_competit ion. from abroad will continue. common in-eariiern.years, thus requiring Ch. 7 Financing: Its Role in Competitiveness :7 Eitctronics 275

greater dependence on external sources of its markets. The second termthe ratio of as- capital. sets to equityis one way of measuring ti firm's Costs" of external funds will he high, and leverage, an alternative to debt-equity ratio. may place U.S. firms at a disadvantage Profitability, the third, depends on many fac- compared to competitors in Japan, The tors: product mix, competition, and labor pro- profit levels required may be difficult for ductivity, to mention just a few. American firms to reach. Industry analysts tend to focus on the asset The first point deals with the continuing abil- turnover ratio, which may fall as a conse- ity of electronics companies to fund growth in- quence of expensive capital equipment (this ternally. To expand sales, manufacturers must should, however, improve productivity). Thus either purchase assets that are more produc- to preserve existing returns to equity, firms will .tive or use existing assets more effet :tively. To either have to adopt higher leverage ratios or supplementassets, funds must come from one somehow improve profits on sales. Given the or more of the following sources: moneys accu- constraints likely to be exercised by the U.S. inLI1;1! ed through co.porate operations (retained r'nancial community, there are clear bounda- ea ii;gs plus depreciation); capital generated , les to the first optioni.e., to increasing the by lee sale of additional stock; 07 borroWings reletive proportions of debt in a firm's 'capital- of one type or mother. ization. The second possibilitygreater profit- As discussed prie4iously, borrowing w'ehout abilitywill be equally problematic if interna- parallel increases in equity alters afiee-i's tional and domestic competition continues to leverage. This, in turn, tends to increase the be stiff. variability of rettfrns to equity, increasing the There is a second difficulty. In the identity risks to owners/. In countries like the United above, note that if each term were to be held States, where capital markets are relatively constant, growth would have to involve a bal- competitive, managements choosing higher le- anced expansion of .lebt and equity. But the verage eventually confront two problems: 1) com- !eft-hand side of the equation dictates that, if mon stockholders become increasingly sensitive a given increase in equity is to be financed in- to their risk positions, and make adjustments ternallyi.e., through retained earningsthere that tend to depress stuck prices; and 2) lenders must be a proportionate increase in aftertax also may:'object, ultimately refusing to supply returns to equity (profits). (This assumes that additional debt. As capital intensity increases, electronics companies use aftertax earnings to financial managers face an intricate set of de- fund growth instead of paying dividends, a pol- cisions; icy followed by most of the rapidly growing The important relationships can be ex- companies during the 1970's.) Thus, growth 1-:,,Atssi!flits..oilows: must be accompanied by an increase in profits. There is no requirement that new capital be restricted to funds generated internally. Firms Valueof Tot-1 value I,J1( sides t-f Net profit could tap markets for both equity and addition. . . . 0; Total vt:Itte Value ui ecitta,.. Va!ott of al debt. /NOnagements, however, often object (-1 asset,: to stock offerings on the grounds that new issues are expensive." Aside from the costs of 17his expression is simply an identity. Thefirst floating the offering, before the current boom ferin on the right-hand side, asset turnover, is executives commonly cited what they per- a broad measure of Asset productivity. Defined ceived as low price-earnings (P--E) ratios in seas I. A; ratio of sales to total assets, it in,',icates curitY markets. From the perspective of man- / the efficiency with whic:la company utilizes agenient, selling new stock-under such condi- its assets to .i/enerate revenues, and deper-de tionS would have provided too little capital -on-such -fa,evee-eas_thelirmls.capitai intensity. compared to their expectations of future and the degree of competition cha.r:Y*...... ,ng growth, earnings, and presumably dividend 276 International Competitiveness in Electronics payments. P-E ratios of 2 or ;:t for electronics While equity from venture capital sources stocks were cited as indications :if "unrealis- has been more freely available since the 1978 tic:" rate of return requirementsas implying revision of capital gains tax rates, most of the that the market vas demanding returns in the firms needing capital will be well beyond this range of 33 to 50 percent. New equity issues. stage. Nor is it likely that such sources could in this view, should be delayed until the market provide enough money; the venture capital returned to more normal co7ditionsi.e., un- market is far too small. Still, there remain por- til stock prices and P-E ratios rise. tions of the electronics industry where initial capital requirements are less than in semicon- Whether or not this makes sense deperAs in ductor manufacturinge.g., computer 'soft- part on how stockholders are perceive. To s!-..e.Cialized industrial equipmentand those who believe that equity holders are, 1i where startups should find capital relatively fact, owners of the corporationand if existing abundant. Considering the effective reduc!i Ins stockholders have first rights in the purchase in corporate income taxes resulting from the of new issuesthen stock prices may appear Economu_ Recovery Tax Act of 1981, the total too low to management but the firm's "loss" venture carital pool should continue to ex- is exactly offset by the "gains" of these stock- pand. holders. That is. existing owners would be able to buy new stoc..."cheaply." The existing own- Even so, many observers predict that the cost ers would be unaffected by variations in the of funds will be too high. From this perspec- issue price based on market conditions. On the tive, investments may simply not promise ade- other hand, equity may be viewed as effectively quate rewards; American companies in parts another form of subordinated debtin reality of the industry that face mounting competition separated from any control. In this case, man-. Jrn abr--2,ad may have difficulty in attracting agers would perceive equity as "borrowing" funds from wary lenders with numerous alter- and its costs wo,ld be evaluated like any other natives, some offering better prospects for safe- debt in cost of capital calculations. Statements ty and high returns. Finally, some commenta- on the high cost of equity by managements of tors believe that the total supply of investment electronics firms oaen seem to point toward capital in tne United States is smaller than it this second belie could or should be because of a variety of disincentives affecting savings and investment Then to the seconc: pond above: Will exter- built into the American tax systein.29 While the nal funding be available at a cost manufactur- capital market will certainly supply funds in ers are willing and ablein the face of compet- an amount equal to total demand at some set itive pressty-s--tu : the moment. the of interest rates, such observers believe that availability u(fur.3 does not appear to be supply constraints artificially force these rates a limiting factor. c, could change as firms to levels too high compared to other countries. increase their leve Lending institutions typically establish standards on levels of debt Several of these matters are at the heart of considered prudentguidelines that depend on economic policy debates still before Congress; liquidity and the variability of cash flows to the the capital formation question, in particular, borrowing firm. Effective limits .on debt. there- has been widely discussed for years, and has fore, differ front co.npany to company. Ever. not been resolved by the tax policy changes in- if some companies can borrow what ,hey named, stituted in 1981.29 The issues are often complex the total volume of funds required by the U.S.. and teclsical. As a consequence, the discus- electronics industry seems so higil that other sion below concentrates on matters that are will almost certainly reed new equity. particularly relevent to:)ctronics:

"For a typical commentary, see A. Sloan and C. Miles, Slow- down at Capital Gap," Forbes, Jan.;. 1980, p. 38, For similar argumanis applied to the Ja.,.)aneira caseice 2i lee, for wimple, Capital Formation, hearings, Joint Econotr. Wallich and Wallich., op. ,it., pp. 268-269. is Committee, Congress of the States, lone 9, 1976. Ch. 7Financing: Its Role in Competitiveness in Electronics 277

Capital Supplies and Financing Costs aresubstantiallylarger thanthetypical for the U.S. Electronic.:, Industry American merchant suppliers, several of the As mentioned previously. American elec- U.S. firms already have been acquired by much tronics firmswith some prominent excep- bigger companies. Intersil may look puny com- tions tend to be smaller than their major in- pared to Hitachi ce NEC, but this is hardly true ternational competitors. And, in part beca:ise of its new owner, GE. It is quite possible that their lack of diversification results in sharper- further rationalizetions of this type will con- swings in cash flow, small companies confront the, perhaps in part to assure continued fund- higher financing costs than large integrated ing for expansicu. FAthough many of the cur- manufacturers. beret ire; on the basis of com- rently attractive candidates for purchase have pany size alone, electronics producers are now been acquired. Finally, as table 59 also likely to face higher costs of capital than many notes, the two largest captive manufacturers of their competi.ers in Western Europe or Ja- in the United States, IBM and Western Elec- pan. trc, have assets much larger than the Japanese producers. The semiconductor industry again provides a ready example. Table 59 lists total assets for For U.S. companies that are riot affiliated a sample of U.S. and Japanese companies. -vith larger firms, the question of differential While Japan's semiconductor manufacturers funding costs remains. Based on the usual assumptions concerni i-tident amounts of leverage as discussed ea.,::.er, the difference in debt financing costs bie.ween a firm the size of Intel (table 59) and companies like Motorola Table 59.Assets of U.S. and Japanese or Texas Instruments probably averages about Semiconductor Manufacturers 1.2 percentf,,c points (ir' fact, as table 51 Totalassets lustrateS, the firm-to-firm differences at any (millionsof dollars) point in time clearly depend on many factors 1979 1980 beyond size). _Ci)=11 else equal, Japanese firms U.S. merchant firms making semiconductors would probably have Advanced Micro Devices $109 $165 American Microsystems 80 81a about the same advantage as a result of their Intel 500 767 size and diversity. But, because debt accounts ntersil . 83 98' for only a fraction of a company's capitaliza- 161 Motorr)r) ...... 1,904 2,112 tion, the impact of this difference in interest National Sernici..,nduct,)r 385 561 rates is smaller, and other forces are likely to Texas Instrumems 1.908 2.414 weigh heavier in the competiCve balance. After U.S. captive producers all, the :Xmerican firms included in ta'''e: 59 IBM . ... S24.530 $26,703 Western Electric' 7,126 3.048 were f.,rmed, grew, and flourished even though Japn.-wse Companies' their capital costs were greater than such -.em- Fujitsu $2,030 $2,380 petitors as RCA, GE, and GTE, all of 1.vhor. en- Hitachi 6,790 7,450 titre(' the semico..ductor market in its early Matsushita Elec:tric 5,190 5,640 Mitsubishi Electric 4,490 4,910 years. Nippon Electric Co. (NEC) 3.110 3,560 Toshiba 6,180 6,450 The related questionwhether interest rates AC7un r1 in early 15'12 t'y GcAlir,.rich nail 1980 assets of Al 6 talhon in the Un"d States have been driven to ex- LA:. of Sebteatser 1980. in early t Inter sa purcha4,ed 5r General Elec t,", 1:090 a: nets acre. T.18 5 billion cessive lei by supply constraintsi.i too in "-%-r-tn,ea a fy; s7 7 LOH:, valved to cover in any detail. It is true, for ex- l-,:',.'n=tern Elector: only ,0 ,, not Include assets of `3.T. cperating ,asortns oche; A IT stAbsi,.;,tts ample, that in ;erest rates to carp-07a o bor- "Asset ''Gore; do not include iatillata5 i..anyaraions forn yen to cIC.asas based rotes trotEronomio .Reoorf of t,lo Pf:sidort (Wasilinoon, t) C rowers are lower in Japan for equivalent pro- ,0Q!, r,145 1979, 218 yen per del,ar, 980. 226 pen per dollar portions of debt (to the extent that equivalence ,;0151,; . IS, firms - it r.n1.1a1-e 5 li.pank.,e firmsJana Hand f-0cpit yo Toyo Kel:a Ft.;-://Ii),.;t. 1579, can be ascertaine(.l). But such comparisons, , on diff,:rent currences, con easily !ME-

2 8 2 278 International Competitiveness in. Electronics lead. In ;)art. they reflect differing inflationary changescuts in personal income taxes as well expec:at ions as In.ircored in interest rates. Ad- as effective reductions in corporate taxes justments for inflationarN,, expectations. are were directed at enlarging the supply of funds proilematical; to some extent, it might be possi- forinvestment.Unfortunately,increasing ble to make such adjushnents based on patterns capital supplieswhich all else equal will of variation in the exchange rate. But a varie- decrease the costs of investmenttends to be ty of factors other than inflationary expecta- more easily said than done.3' For instance, tax tions affect foreign exchange markets. especial- changes that affect both supply and demand ly in the short run, and the gyrations of the yen for funds may leave interest rates unchanged. against the dollar in recent years have gener- Under this circumstance, neither the availabili- ated a wholly independent source of contro- ty nor the cost of funds for Amer'in elec- versy."'' tronics firms would change. The inflationary trends reflected in high U.S. In part because of these complicating factors, interest rates can themselves deter new invest- it is not yet clear what the net effects of the ment, in electronics and in other sectors of the Economic Recovery Tax Act of 1981 on aggre- econoiny. Inflation adds to uncertainty in the gate capital formation will he. much less the cash flows that can be anticipated from in- differential nnpacts on various sectors of the vestments; these mount as investment horizons economy. Thus far, there is scant evidence of recede. High rates of price inflation tend to broad positive effects on capital investment in discourage longer term commitments; instead. industry; many observers are skepti,-tal that the they/ favor investments with relatively quick revisions to U.S. tax law will have niuch effect payoffs. American managersin electronics as on levels of personal savings, which they feel in other indo trieshave been charged with ig- are central to freeing up new investment for noring long-term growth opportunities while industry.32 Internationally, even before the concentrating on the short run. Part of this more rapid depreciation schedules and other hesitancy to commit resources is tied to uncer- reductions in corporate taxes enacted in 1981, tainties created by price inflation and the at- the United States had in place tax policies that, t impact on interest rates. That is, dif- according to at least some analyses. favored hces in short-term compared to long-term capital investment more strongly than taxation managerial behavior between American and in Japan and most of the European nations.33 Japanese corporations may not be caused .by &fierences in capital availability, or by "C .-er the postwar period, investment levels as a fraction of ferences ill real interest rates, so much as by the ....mintry's gross national product have C.,,ctuated markedly year to but with no in:idol-ice of any long-term ',lend uncertainties associated with high and variable up or down. SII! j. W. E. Conrad, and I.. I., mime). rates of interescand inflation. Policy and Capital Formation,"ederalReserve Bulletin, October 1981, p. 749. For an excellent summary or international "The UniteJtates has not"had 0 great deal differer,co-. in capital formation. s(!I! V. C. Price. "capital For. of succw;.-; ill controlling inflation over the past !nation and Investment Policy." l' dern1.;:conninies in Trans,- ,i(w. (Jul:we andAdjte:treeel )'elidesin Industrial few :years; while current economic policies ( :otintr,e.,, I.Levestm and I. W. 11/:coler !eds.) Ilsoulder, Colo.: May help to 1, et'!) down the inflation rate, ad- West vim.v Pro's, 1979). p. 135. "K. W. Aronson. -The Low. Ll.S. Hate of Savings," Neer r stinerits jrxpectations often lag well behind, Timm,,Dec. 22, aPe p (:an the eh;conics industry expect to benefit C.F. Kopits,- rax Provisions lo Itmst Cape. it Formation Vary moire directly fruit) the changes in U.S. tax law Widely in LIdosfrial Nations," 17, 131110, p. 055. adopted in 1981? A Per all, Irrniy of these The effects inter;.:Itional :1117ffelict!s in taxation on coin- petitiveness are, (1.^.1211001iIrirdy complicated. The .Si11114: '".tioo, for I'. Ilarilandl'Ilimborg "V,111! of vtFtivls ci(:111HCS thai ,oply to otbor international comparisons areit ;;;,1,Exrlintige }tats, Nt't ,inports a work, compmepled by the co:nolexitii.,. of the tax' codes iu each Betio' 1.os Anger/es 111, .iimkestnen for country :dies rf tax;',:on be compared in a rminher Aim.rican (41..0 :6nm competitive problems im an of d;f1ri.:0 wm..3-. with results depend on factors such ....ndervalo,,d von mai, ng 1;,i)a..u.st, imports cheap in the 1,;.5. projectedin:.; ,1.1) s.Even tax mirket. Ihems micro:,ap, caoot a. prohlm,a n)f re:atine the Ch. 7Financing: Its Role hi Competitiveness in Electronics 279

At thu number of nations use solese.N.:e. All equiHnent used for R&D can subsidies other than tax incentives now be wr:itt. , off over 3 years; so can any inure actively than the United States to support equipment that could, under the old law, be certain of their industries.34 depreciated in 4 years or less. All other equip- Regardless of effects on overall supplies of ment is now eligible for a 5-year writeoff.35 ci:pital, the accelerated depreciation schedules Although electronics firms probably pur- ini7)lemented by the 1931 Tax Actseem likely chase little equipment with useful lives less to place electronics firms at a disadvantage than 3 years. thisis nonetheless now the relative to other industries with which elec- minimum depreciation period. To the extent tronic. competes for funds. More rapid depre- that a company was formerly able to write off ciation lowers tax obligations and increases some of its equipment more quickly, it may suf- cash flows from new investments. Faster defer a reduction in cash flow. While this should preciation raises profits for projects thatwere be no more than a minor factor, the new de- formerly marginal or unattractive; the result preciation procedures will place some elec- should be to increase the overall demand for tronics manufacturersindeed, firms in any in- investment. funds and the overall rate of invest- dustry where technological change results in ment in industry. For most industries deprecia- rapid obsolescenceat a disadvantage with tion is inure rapid under the new lawbut not respect to industries that reap greater benefits `necessarily for electronics. from the new depreciation schedules. The lat- The 1981 Tax Act permits equipment suchas ter tend to be industries where technological that used in production or R&D to he depreci- change is slower, and equipment has a long/ ated over either a 3- or a 5-year period. Former- useful life. Steel and other heavy industries are ly, pro:Mc:Hon equipment was depreciated at examples.36 rates at ieast nominally related to actual oh- Focusing on levels of domestic savings can also lead to underestimates of impacts flowing effects policy to economic performance remains:-See. from international financial markets. Nations k.ftv.c1;; Industrial Countries: A Ilriel Cori, pareion. op. cit. need no longer rely on domestic savings for in- Sobsolies and Corapetitiveness International- vestment; international capital movements =Ire ' Iv NPA Cr.!.ii'' eon changing International large and continuing to grow. For the Unied 1,:no.try States, these long-term capital flows, both portfolio and direct investment, have been negative for many yearsi.e., the flow of funds out of the United States has exceeded the inward flow.37 This implies that rates of return are

',5Econornic Recovery Tax Act of 1981. title II. subtitle A, SSC. 201. Also see P. Oosterhuis, "High Te;;:mology Industry ;IN: TikX policy in the 11Jfills,' National journal, Jan. 1. 19113, p. 40. ,"The only analyses of differential impacts across industries that have been published are on a highly aggreg4ted basis; thus it is not clear how electronicsmuch less particular portions of the industrywill fare. The "machinery and instrumen0i- category. within which the electronics industry falls. is one of the least favored sectors under the new depreciation method:. See aurunilic Report of the President (Washington, D.C.. lzibroary 19921, pp. 122-125. particularly table 5-7; also J. C. ( ;rave.:'-'. -Effects of Ore Accelerated Cost 1-4;cove; T. Sy:item by Asset Type," Congressional Research Service,-A'.- ;. 31_ 1981. "Althowh direct investment of U.S. funds kwerscas is still running o to three times the level of forebiln investment in the !sited States, foreign holdin;ts of U.S. securities roughly ialance Prot(' Universal Insiruments American holdings of foreipn securities. Sit Statistical A, hstract 11 the iThited Slates, 1982 40 103rd u.d. (Washington, D.C.: Aut.prnated machinery for ele..:tronics assembly partment of Commerce. 1.1e,:einhet1982)i p. I

28,1 Corn & ::fl eness in Eiectronics

11: :2. erseas .tt hoe te.1.l ereFore gen- cum:nen:A banking system, and various crating more savings in tile tjnited States v.iii sectors are all hierarchically trily :':crease the rate of domestic with control over resource allocation vested in eitnital lot aiat:i.e. funds :nay flow abroad in- the Ministry of Fin -nice. AlC7ough hardly the stead. At the same time, were the cause of high :ase today, this is probably a fairif simplis- interest rates in the United State.zi simply a erepresentation of the situation two shortage of fends for capital investment, 111011- deca(Ifis ago. And even Low. at the= level of in- flow here from abroad. In- dividuni inw lment decisions, the Japanese ternal:tend capita: markets Operat:- effi- system resi). much more directly to the ...1cr:t iv'111(if'1''1.1111 t.irCITInStiinCeS. (vishes of government than does that in the United StiLes. Hut if government guidance still Japan. exists. it is a far weaker force now than 20 years. (_), and the investment climate in Japanmuch -1!, at extent have .ii1F.:11eSe electronics more fluid, ft;ice, beim aided by the unique structure of the capital markets. aI .ictor that has Now to more concrete Questions: WI:A do ;''s corporations ad- Japanese corporations utilize much gieater pr)- heel, (feattiu ei( to give tap portions of bank debt in their financial struc- vantages in niter-me:1(min cor)petition? tures than firms in the United States or West- ern Europe? (beverage in Europeanelectronics Postwar Trends in the Japanese firms tends to be grclt.'r than for American Financial System cool: ias out less that of jApai:ese.) The financial system lies changed more answer lies par)y in the historical development ihan ind.-4 over the past :33 years, and the cowl- of industrial groupings ill Japan, roost of which markets and financial institutions ere now contain one or more banks.28 Japanese Mall ti- far reinoved\from their grossly underdeveloped fiAuring companies for Many years have de- state ill the early postwar period, 'the transfor- pended much more heavit.: on close working mation of the japanesc financial system has relationships with bankseven to the extent of paolleled, first, the reconstruction of Japan's participation by banks in management deci- ,.!conoiny, and, following that. the count r('s sionsthan have firms in the United States or dramatic industrial expansion. 'rile yen has Europe (West (Thrtnany is a partial exception, become a major internotional currency, while as ';,.lcussed below). In other paof tin world, l'eko only l).1 :iv by governmental generally enter the picture onlyif emerging its a (vorld banking, center. reorganization follows bankruptcy, whereas in (live.; the ,.peed with which the Japanese finaricil system has evolved. it is easy to he "Theso 2roups, s oli d idihat.,/; 1,1,; .,(:nietiling hire holditiL, ror general background,Svf! misled Ir., images frotil the past. Yet the future, It. F. Cave:. and !,.1. Cekilsa, luchrslridl lir,14ani....iaion in Japan. e% en more than the present,should he the real op cit., As(), (;ronpin1,.., tit r .,oncet n: effects on.international competitive- pt(wil Crolyyo: 1)(.:t1 vd1 '.d.irkuting Cunsultituis,11:1 Functions ()I' the dynamics of change an illustration, consider the Hitachi groupII of IleSS ail" 1)1',J v 500 firms; ;is (,I I..td. held maiority in markets ill the t Jnited States, Jima.. and minority sharestypic:dly in the 20 to 5(1 per(amt and elsewhere. Insight into competitive trends 1,1: the remainder (I. Gres.,r. !lig!) Technology and ine.q! InthistriA A Strategy for U.S. Po/it:pm/kers, depends oil these dynami s. Siii,c(?;:nnittee on Trade, Committee on Ways and Means, House of l

Ackely and I I. Ishi, "Fiscal, Monetary. and Related Pol- icies,- Asia's Oiant, op. cit., p. 153. ..""Thu Perilous Hunt for Financing," op. cit.

2 - 1 11 - 83 - ' 282 International Competitiveness in Electronics

Figure 52.Debt-Equity Ratios for Japanese Electrical /Electronics FirmSa 120 Tclal Long-term debt/equity debt/eq,

100

20 (Long-term not available for 1980' 0 1972 1974 1976 1978 1980

Year

4tf S t:ro-,i In 1972. 54. tit-^,4 to 197478, 14 thlT.5 1980. rES 1972.78" .+oalcat ,o the L;ornpet;tiveheSs of the U.S Electronics Industry," report prepa-ec for CIA by L Ui Bc,gman Co under .,-,ntract No 033-1.5. p 471. trf)-1 Japan Company Handbook (Tokyo: Toyc, Ke.:w S:prposhalThe Orien'ai Econornps-1981).

Table 60.Sources of Exte-nal Financing for ther the six Japanese electronics Firms, nor, :-<- U.S. and Japanese Corporations cept for RCA, the seven in the United States Prcportions by source, stray too far from the patterns typical of each 1965.72 country. As ti3table shows, the Japanese I rrtr United StatesJapan ponies have depended much more heavily on Loans from banks and financial external capitala point that has been em- intermediaries 51% 89% phasized previously (table 52). The heavy reli- Bond issues ...... 37 4 Stock issues . 12 7 ance on both shor.- and long-t -rm debt in 100°4 100% Japan's electronics industry contrasts sharply SOURCE T raaru.yama. -Frnancing Japanese Business." The Conference Ronrd with U.S. semiconductor firms. Electronics May 4978. p 47 companies in Japan utilize very large absolute Ch. 7Finar,cing: Its Role in Competitiveness in Elect-hnics 233

Table 61.Short- and Long-Term Debt as a Percent Total Capitaltzelion for U S. and Japanese Electronics Firms

Wei cited average of Ai; Japanese Si x Japanese electronics four 1!.S. sRmicor.n.:clor trtc.rdilual U.S. companies. 1')79 1975 companies. 1979a rnanufactures. 197c;" OS'; IBM FicA

38 0' ii 8.1'.', C...=.3".:,.. 5.3'..:, 164'.4-. 37.5 29.0 12.7 23.1 9.0 37.2 71 7 67.0 20.8 24.0 14.4 55.5 Stlarer-,oklet 5 ec:utti 28 3 33.0 79.2 75o as.,; 44.4

To!aJ ca;)ita'dzton li.'./s 100P,S 100,..-',, 100'..". 100%

Lcr"c; 7:,-r. .:.?..);'e:',:..:,ty 123 01'6 0.15 '1.30 0.11 384

. 7,, .ii ::t.,:".::-.;.,' . 253 2C; G:5 0.32 017 0.566 ;,,S,,,.-.:C-, ani: 7,-;sr,ipa A.i.;-,_,J V.,i'.,, :).-....i,..,, ,,DiaNa7-,,a, F.,,,,,Tti,:.0,,Cl:Clc,fTP.a::1,,..rurrt !!; SC.,.,',Ci-..S Japanese companies u S anis Ja;a-ese Semics-Iclor7 _:1 tr,:iiiiiiles A +,.:,lancti Go,,,;-,a,scn.- Cr'..se F-lan.c,at Pc..i:cy tc, ine Serncc,r1CL:clo,1,111..:stry As.;,-:,.:37 ,-,-_,..,e g 1:,,eCD62 American Iimt Anni;a1 ii-.iiirts

amounts of short-term debt which is automat- maintain a state of dynamic equilibrium with ically turned over as it falls duei.e.. the iaans respect to one another which is not always at- are rew-naten. normally et an interest rate contained in countries likeJapan.Japanese capital sistent with prevailing market conditionsa rnat-ketsand those in many other countries,- pointed cut earlier, these practices seem to are narrower; nor are they as closely linked. place Japanese banks in positions of consider- Neither individual markets nor individual in- ably greater risk than would be acceptable here. vestment decisions respond as quickly or as concertedly to changing conditions. Borrowers Tables GO and 61 emphasize the extent to have few--potential sources of funds.japanese which capital structures in Japan are weiglqed corporations must use bank financing; savers toward externa! financing. These practices must rely on bank deposits or government sav- must be taken into account when calculating ings institutions. Under such circumstances, costs of capital. The American financial sys- J. panese banks have little option but to accept ternand therefore the methods of establishing that would be unacceptable in the United capital costs commonly used herepresumes State!;,_heir other lending opportunities are a mix of alterna:ive sources of financing. Not too limited. all of these are widely available in Japan; sonic do not even exist. Ledgers of Japanese companies differ from those of American firm 3 on the asset side as In the United States for example, individual well as the liability side. Table 62 illustrates portfolio holders adjust their :security positions some of these asset side differences. Japanese in response to changing market conditions, electronics companies carry much rnere cash riding off risk against potential returns. and other liquid. assets, reflecting in part re- Americanseven those with modest assets tp.:irements for compensating balances im- have become sophisticated investors, switching posed by banks; they have less tied up in ac- in and .iut of certificates of d -.)sit, money counts receivable and inventories, The fixed marke, funds. corporate stocks, and other in- assets of !apanese firmsplans, and equip- vestments in response to small swings in rela- mentare proport;anately smaller, in part tive rates of return. The escalation in real estate because of grossly understated land values; in prices during the 1970's. reflecting high rates some cases plant and equipment ,R:iluations of re!, investments in land and housing may be reduced further by depreciation rates as w auvantares, is another, example. In Jap.in, individuals do not have this range of highei than in the United States.4 investment opportunities; for example, markets "On f.heundervaluation of assets in japan, 1. Keroda and for corporate bonds barely exist. The well- Y. Oritani, "A Reexamination of the Unique Features of japan's developed capital markets in the United Stvtes Corporate Financial Sp-tic:two," lepe.noseEq.:ono:lb.(' Studies, 284 International Competitiveness in Electronics

Table U.Asset Categories as Listed on Balance Sheets of Electronics Firms in the United States and Japan (percentage of total assets)

Advanced Micro National Texas U.S. companies (1979) Devices Motorola SemiconductorInstruments Cash and marketable securities 7% 5% 2% 6% Accounts receivable, net 28 26 32 29 !nventories, net 13 29 27 18 Other current assets 7 5 2 4 Total current assets 55 65 63 57 Net plant and equipment 45 34 35 42 Other noncurrent assets 1 2 1 Total noncurrent assets 45 35 37 43 Total assets 100% 100% 100% 100% Japanese companies (1978.1979) Hitachi Matsushita Mitsubishi Toshiba Cash and marketable securities 22% 18% 150/0 22% Accounts and trade related notes received 21 15 31 22 Inventories, net 19 17 26 19 Other current assets 5 6 9 6 Total currettt assets 67 56 81 69 Net plant and equipment 16 12 13 15 Other noncurrent assets 17 32 6 16 Total noncurrent assets 33 44 19 31 Total assets 100% 100% 100% 100% SOURCES: U.S. firmsDerived from data in Moody's Industrials, 1980. Japanese firmsDerived from data in U.S. and Japanese Semiconductor Ir4rustries: A Financial Comparison. Chase Financial Policy for the Semiconductor Industry Assocta tion June 9. 1980. appendixes.

But the important point of table62is the What Role Does Japan's Government Play? large fraction of short-term, liquid assets held by Japanese electronics firms. These assets do The relatively underdeveloped state of Jap- not earn high returns. Thus, on the one hand, anese capital markets gives the government lev- lower financing costs for Japanese firms are erage over allocations of funds for investment. partially offset by lower returns on their large In the absence of a wide range of alternatives, holdings of short-term assets. On the, other both financial institutions and industrial cor- hand, the greater risks that Japanese banks ap- porations are more susceptible to government parently accept are partially ameliorated by the influence. Does, then, the Japanese Govern- high levels of these same current assets. The ment indirectly subsidize target industries result is to reduce the capital cost advantages using the banking system as a conduit? A varie- of Japanese electronics firms while helping ex- ty of mechanisms would be possiblefor in- plain how they can carry such high levels of stance, government funds could flow to the debt. banking community in the form of low-cost loans earmarked for certain uses. The funds might come from tax revenues or from private (footnote continued from p. 283) savings deposited in government-controlled in- summer 1980, p. 82. In general, depreciation rates in Japan are comparable to those in the United States except for selected institutions-such as the postal savings system. vestment categories that are favored by accelerated schedules. See, "Corporation Income Tax Treatment of Investment and In- Unfortunately, just because the possible novation Activities in Six Countries," PRA research report 81-1, routes are indirect, evidence bearing on this prepared for the National Science Foundation, Aug. 12, 1981, pp. 90-95; also J. A. Pechman and K. Kaizuka, "Taxation," Asia's question is scarce. The most useful comes from New Giant, op. cit,, p. 317, and Tax Rate's In Major Industrial the collective financial statements of Japanese Countries: A Brief.Comparison, op. cit. bankstable63.The liability side of the ledger .28 Ch. 7Financing: Its Role in Competitiveness in Electronics 285

Table 63.Assets and Liabilities nancial markets. Then does Japan's Govern- of Japanese Commercial Banksa ment still influence lending decisions? In the Value past, target industries were certainly favored (billions of yen)Percentage with low-cost capital, although the extent and Assets force of this on industrial development is much Cash 7,559 6.3% less at present than 20 years ago. The govern- Securities 14,335 11.8 Short-term assets 86,634 71.9 ment appears to act largely through informal Other 11,947 10.0 and indirect mechanisms, rather than explicitly 120,475 100% allocating low-cost funds. Economic develop- Liabilities ment goals set by government after extensive Deposits 86,302 71.6% consultation with financial institutions and in- Borrowings from the Bank of Japan 1,570 1.3 dustry have traditionally been supported by the All other borrowings 315 0.3 banks. Because of the close working relation- Short-term liabilities 18,605 15.4 ships among government, the banks, and pri- Reserves 2,210 1.9 Other 11,473 9.5 vate industry, suggestions made by government 120,475 100% officials tend to be consistent with the predis- aAs of the end of 1975. positions of bank managers. SOURCE The Japanese Financial System (Tokyo: The Bank of Japan. 1978). In nracticr,,ins flow preferentially to larger companies, r,:fr.fitt of which are associated with is of particular interestspecifically, borrow- one or eror:. of -lie major banks through an in ing from the central b :nk, the Bank of Japan. dustrial groupink, Interest rates on these loans Although quasi-independent, the government are typically {hose for small borrowers holds majority ownership in the Bank, the oper- (such discrEtthrotirm is common in all indus- ations of which are closely monitored by trial f:. :ntries). ill, t.:;41),:r.pite the higher financ- the Ministry of Finance.42 As the table shows, ing c;..)sts faced by rKe.,.e and small companies, lending by the Bank of Japan to commer- firms such as Sony Honda have beCome cial banksthe -"overloan" phenomenon highly successful (luting the postwar neriod. amounted to only1.3percent of total liabilities Some have even managea to establish their in1975(the percentage is no doubt less today).43 own industrial groupings; Matsushita', which This is too little, by itself, to give the Bank or had fewer than2,000employees before the the government much influence over the rest war, is the most prominent case in electronics. of the banking system. Nor could funds from the Bank of Japan significantly reduce costs of Costs of capital in Japan dependfar more on money to commercial banks. Such weight as broad controls exercised over interest rates the government might exercise would, there- paid by the banking sector on /deposits and fore, have to flow from other sources; some charged on short-term loans than on govern- American observers hold that informal chan- ment guidance of investment flOws. High rates nels are quite sufficient. of capital formation have been rooted in sav- The situation was different in earlier years, ingsthe mirror image of investmentas illus- when overloans were much larger; their de- trated previously in table 54! The savings rate crease as a proportion of the total liabilities of in Japan is especially surprising because for Japan's commercial banks is another indication many years the government followed a delib, of the changing character of the country's fi- erate policy of keeping interest on savings and other investments low. The effect was to pre- 42K. Haitani, The Japanese Economic System: An Institutional vent interest rates from becoming the primary Overview (Lexington, Mass.: Lexington Books, 1976), pp. mechanism for allocating capital, as would 164-165. have occurred with market-determined rates. "An overloan simply means that commercial banksindi- vidually or in the aggregateare in debt to the Bank of Japan. But if in earlier years capital rationing gave ad- See Suzuki. op. cit., pp. 1-'.3. vantages to some sectors of the economy,

2 9 u 286 ticeirVional Competitiveness in

others the cost. In general, the answer must be as follows: left free to ad- .L.5:2:h.-.6ted at the exuenSe of t;Lls`, interest rates would have been higher. On savers. i'tte other hand, if the government's actions s,..:1--ved primarily to allocate funds among sec- Table personal nr..ar$ in ja --iws of the economy, then the answer is less ob- have zere, or negative -ratesOf vi...)us. To the extent that commercial banks bor7 turn (p:1.1:!r ;.Y.ii,-,IstineAts for infle:..-1.);1-(.,3r rraur-11 rFivred from the Bank of Japan-and overloans of the rfa.-4 20 ?lependingi.oFfnrse extent lame- during the 1960's-then interest on the(',.';'51e,!-Ltcs;-,-.P-5chosen. neriod r±Tres were probably depressed relative to levels 1961-69, major categw..:ts vf person- 211,:tt better developed financial markets would al 'in the tai'..';:,--remainetl. set, particularly if overloans had been pro fixed li1 fed. current impacts are uncertain, if Longer ttarned ,.st at a'oout i-rtily because overloans haVe declined in recent the rate:IAPa ti .1.141.1Cethe real ...-.1)turr(s were years. essentially -;;tf.ri:). Sli):rter term negative',7eturns,Mvich the same --airs Vrrae dur- The Japanese Government has other means ing the 1.-!.-3-70';-;, WiCa price consider- le:; help selected industries get investment cap- ably more Large part 1-,..4::,11,1193 sud-- MI. There is, for example, the Japan Develop- den inc-,:Yec-r,iv:-::<; e...nergy pres, nr:,-tatAs. in Treitt Bank-a public corporation through 1974-4 !ri were perio;71.ca517!13,444- which moneys from postal savings and trust justeri, returnsremair, f.7/.1-: 7:ixounts can be funneled. In the early postwar years, the Bank was a major instrument of gov- 'r;./I'ViPtgSbeen ttaal..4,h ernment policy, but its influence rapidly de- the ba;.(1jr.:---.,; ,Ifyslem, appearing t dkacd; the-Development Bank provided 22 per- dusiry bow rates? The que5.7.'.:.n -of all capital invested in industrial plant can be asko...I another way.. W,-.sat tre.ncic ?.s.,..1 equipment in 1953, but only 5 percent in would interest rale.,..fi breve beeil Between 1965 and 1974, loans from freer to adjust, and had saver e'n,:z--A-1 atoir- nment financial institutions-of which alternatives in placing their funds? 7. 1.7; '1",cs:.t,elopment Bank is but one-averaged the Japanese Government has ration::ci 4. :;percent of new industrial funds.45 Still, '.mall percentage came to about $3 billion arninally, more than sufficient for major im-; pacts on international competitiveness if con- Table 64.-Interest Rates and Price Inflation in Japan centrated on well- chosen targets.

Annual average interest rates ,'.nnual Continuing Change in Japan's change in One-year Postal ca:-Ir ws Financial System Consumer Demand time Price Index depositsdepositsOrdinary 2-3 year In terms of competitiveness, and from the 1961-69a 5.5% 2.19% 5.5% 3.6% 5.5% 1970 7.7 2.25 5.75 3.6 5.75 perspective of the American electronics in- 1972 4.5 2.0 5.25 3.36 5.5 dustry. the critical questions deal with the 1974 24.5 3.0 8.0b 4.32 8.0 1976 9.3 2.5 6.75 3.84 6.75 future. Change in Japan has been rapid, and 1978 3.8 1.0 4.5 2.4 4.6 the pace may even accelerate. There are at least alnterest rates were held fixed over this period: the values given are the ceiling two reasons: set by the government. The change in Consumer Price Index Is the average for the period. bTwo-year or more. 1. Shocks to the Japanese economy stemming SOURCES: Price Index-Karst/ye( Rodo Tokel (Useful Labor Statistics) (Tokyo: from high energy prices beginning in Nihon Seisansei Honbu (Japan Productivity Center. Labor Productivity Documents Cents :., 1981), p. 136: data based on Shohisha Bukka 1973-74. Shisu (Consumer Price Indax) (Tokyo: scrItu Tokei Kyoku (Prime Min- isters Office, Statistical Bureau). Interest Rates-1967-1974, H. C. Wallich and M. I. Wallich, "Banking 4C. Johnson, Japan's Public Policy Companies (Washington, . and Finance," Asia's New Giant, H. Patrick and H. Rosovsky (eds.) (Washington, D.C.: Brookings Institution, 1976), p.261. 1976 and 1978, D.C.: AEI-Hoover Policy Studies, 1978), p. 98. Bank of Japan, Research Division, New York, N.Y. sHaitani, op. cit., p. 169. Ch. 7Financirg: Its Role in Competitivenesi in Electronics 287

2. Changing aspirations and expectations Four clear trends can be discerned in the among savers, consumers, and the generA evolution of the Japanese financial systera:48 public, along with the increasing complex- 1. Interest rates are becoming more respon- ity of Japan's maturing economy. sive to underlying conditions in capital The explosion of energy priceswith deep markets, and as a result are less subject to impacts on an economy that depends almost the dictates of government. totally on imported fuelshas stimulated a 2. Corporations, especially larger ones, are reevaluation of economic goals. The gov- developing alternative sources of funding ernment is paying more attention to efficien- and relying less on banks. cy in allocating resources, backing away from 3. Banks, partly as a consequence, are look- earlier commitments to high rates of economic ing to individuals and small businesses as growth regardless of costs elsewhere. As peo- borrowers. ple's expectations rise, Japan is devoting more 4. The Japanese Government, in response to resources to the public sector, aiming to im- trends already visible, is moving toward prove the quality of life, broadly conceived. closer links with the international finan- Although public sector expenditures remain cial community. well below levels common in Western Europe or the United States, more money is going While pressure in Japan for market-deter- toward environmental protection and a varie- mined interest rates is not new, only recently ty of social welfare programs. Defense spend- have events combined to make this outcome ing is slowly increasing, par6y in response to a virtual certainty.47 One precipitating factor 7- U.S. pressures. Finally, the Japanese are dis- has been the government's own need, follow- covering that the days are over when a few sec- ing several years of deficit spending, to enter torsgrowing very rapidlycould lead the the capital market. The national debt rose from =country's economic expansion. Future growth 11.7 trillion yen in 1972 to 62.3 trillion yen in will be slower and more balanced. 1978.48 Iii earlier years, the banking system would have absorbed bond issues floated by These trends in the Japanese economy ,will Japan's Government to finance this debt, with be mirrored in the financial setting for in- interest rates set at low levels to minimize costs dustry, where change is already well under- to the treasury. But as such bonds have become way. For example, the government has lost a larger portion of bank portfolios, and as an some of its influence over interest rates; as active secondary market for government bonds Japan takes a more active role in international has developed, bank managers have become financial markets, with funds flowing in and less willing to accept new issues at below- out in greater volumes, interest levels will more market rates. The government has had to raise closely follow those in other industrial nations. yields to levels consistent with secondary mar- The dynamics of the Japanese financial system kets. At least for government issues, a long- are carrying it steadily toward the American term bond market more typical of industrial- model of open capital markets. This does net ized economies is developing. mean that the Japanese Government will abandon its past efforts to guide the economy. Japan Banks have also sought more freedom to set will remain a nation where industrial policy interest rates on certificates of deposit (CDs); is a powerful force. But, as large Japanese firms 41. E. W. Kirby, "The Japanese and Their Changing Economic continue to expand internationally they will Environment," Business in Japan, revised ed., P. Norburry and have more lattitude for independent action, G. Bownas (eds.) (London: Macmillan, 1980), p. 85. 4'M. Borsuk, "Japan/Interest Rates: Consequences of Rates Sen- and the government will necessarily play a sitivity," Far Eastern Economic Review, Mar. 26, 1982, p. 59. lesser role in the allocation of resources. 48K i Thy, op. cit.; p. 88.

2 2 2&6 International Competitiveness in Electronics

Japanese banks, after much negotiation and the evaluate their own portfolios. Many are at- acceptance of a variety of restrictions, were tempting to develop alternative markets for permitted to issue CDs beginning in mid-1979. loans, including foreign lending and expanded By now, movement toward more flexible finan- consumer credit. The Ministry of Finance has cial markets would be hard to stop. If interest recently given banks a good deal more latitude rates are decontrolled in a portion of the econ- in making overseas loans, although informal omy, pressures elsewhere will lead to a paral- quotas still exist.5° Symptomatic of the change lel freeing of rates or else to severe distortions. is the announcement of a loan at favorable Such forces are particularly potent given the rates to Fairchild for the construction of a widespread, desire to see Japan become an in- semiconductor plant in Japan.51 ternational financial center. Lending to households is also on the up- In terms of the U.S-:--electronics induE.try, swing. Mortgages, installment buying, and movement in Japan toward market:determined other forms of consumer credit have been more interest rates should help defuse concern over:the'exception than thu rule, but bank loans for government-subsidized financing. Further housing expanded fivefold during the decade more, as Japanese capital markets become bet- of the 1970%s, and now account for same 10 per- ter developed, new financial instruments will cent of total bank credit.52 Today., as table 65 come into play. Firms will be able to secur2 indicates, households still borrow much less capital from a wider variety of sources, at least in japan than in the United States. Continuing some of which will be less susceptible to gov- movement to.vard greater consumer lending ernment pressure." In short, investment deci- and more diversified bank portfolios again sionmaking will become more decentralized, points toward capital markets in which the as in other highly industrialized, capitalistic primary allocative mechanism will be the countries. Both business executives and gov- market-determined interest rate. ernment officials in Japan have been con Internationally, Japan has made an explicit cerned over the high rates of bankruptcy of recent years. Many of these failures have been de,;sioninvolving both government and the caused at least in part by the highly leveraged financial. communityto take a more promi- positions of Japanese corporations. As a con- nent role in matters affecting the world's sequence, companies have been attempting to currencies.53 This shift reflects a number of broaden their sources of financing in both Marcom; Ir.. "Borrowers Are Eager 7o G,M Yen Loans But number and kindone of the reasons some Must Grapple With Japan's Delays, Wall Street Journal, July 7. 1902, p. 24. companies are marketing securities overseas. ""Japan Offers Loan to Fairchild for IC Plant ... " Electronics, To attract foreign investment, Japanese com- June 2. 1982, p. 73. panies will have to offer rates of return com- "Kirby. op. cit., p. 91. ""(apanese Official Says Government Wants the Yen to petitive with those in other countries and other Become Major World Currency." U.S. Import Weekly. Feb. 2, currencies. This suggests that capital costs in 1983, p. 572. japan are unlikely to diverge very far from those in other parts of the world. Finally, the orientations and strategies of the Table 65.Household Borrowing in Japan major commercial banks in Japan are shifting. and the United States As corporations seek more broadly based fi- Mortgages and consumer installment nancing, and as the profit levels of banks de- loans outstanding as a percentage of GNP cline, bank managers have been forced to re- Japan Unaed States 1965 2.3% 60.8% 1970 4.6 59.4 "Japanese industrial firms floated more than twice as many 1975 12.1 631 bond issues in foreign markets as domestically during 1980. See 1978 . 17.5 68.1 NI. Kanabayashi, "Japanese Business Borrowings Abroad Surged SOURCE: E.SakOlhara, R. Feldman, and Y. llarada, The Japanese Financial Sys to Record In Year Ended March 31," Wall Street Journal. May tem in Comparative Perspective. Joir.it Economic Committee. Congress 12, 1981. p. 36. of the United States. Mar. 1:, l682, p. 21.

el I'1 Ch. 7Financing: Its Role in Competitiveness in Electrzmics 285

The internationalization of the yen will have a major effect on Japan's financing practices. As long as the economy could be insulated from outside impacts, the government could successfully hold down interest rates. This in- sularity is breaking down as banks, private ti ms7-"".7:- businesses, and individuals take advantage of tl,,, ,sange of financial opportunities now open to them. To the extent that Japan becomes a center of international financea process already well underwaythe domestic financial system will become irrevocably linked to larger world markets. And, as the United States and other Western nations have discovered, an integrated international market renders an independent monetary policy aimed at controlling Photo credit: Perkin-Elmer interest rates virtually impossible. Electronbeam lithography system used for making integrated circuits and masks France converging events, the most important of In general, French electronics firms have not which have been Japan's continuing trade sur- emerged as strong international competitors pluses. These surpluses have led to demands nor have U.S. companies confronted insur- in other parts of the world that the yen be used mountable difficillties in competing within as an official reserve and unit of account in France, although the ingenuity of the French private transactions. Foreign-held balances in creating nontariff barriers may match that have been inc:reasing rapidly as an actin e inof the Japanese. Therefore this/ sectionand ternational market in yen develops.54 Finally, that following on West Geimanyoutlines foreign investment by Japanese corporations financing methods much more briefly than for is expanding swiftly. Japanese foreign direct Japan. investment nearly doubled between 1979 and In some ways the French financial system - 1981, reaching a level of $8.9 billion in fiscal more closely resembles the Japanese case than 1981.55 The international position of the the American.58 Like Japan, France has a tradi- Japanese is beginning to look strikingly like that tion of government intervention in what would of the United States 20 years ago, with current be private investment decisions here, and the account surpluses offset by outflows of direct governor good deal of control over nve :.,tment funds. allocat ui dWdY from some of its earlier 54Foreign hole-kings of yen reached the equivalent of about $10 practices in the late 1970's, and toward freer billion by mid- 1930still small in absolute terms but doubling markets for capital as well as goods. Although over a period t f 11/2 years. See E. W. Hayden, "Tokyo as an In- the outlines of Mitterrand's industrial policy ternational Fla anciel Center," Atlantic Community Quarterly, remain somewhat vague more in terms of vol. 19, fall 1981, p.351,which also outlines forces contributing to slackening government influence over Japan's financial mar- mechanisms than objectivesthe Socialist kets. Hayden emphasizes that continuing distortions can be ex- Government has begun to, alter a number of the pected for some years to coma. " "Japan Capital Abroad Reaches Record in FY 81." Japan Re- specific practices inherited from earlier ad. port. joint Publications Research Service IPRS 1/10( 7r, June 7n, 1982,p. 10.For examples of Japanese Invet.:.me:,:!i !., F 41 is largely based on 1. Zysman, Governments, seeA. L.Otten, "Japanese Firms Press European VF Tn Markets, ,..?ad Growth: Financial Systems and the Politics t,In- HelpProfits and Deter Protectionism." Wall St:-eat dustrial Change (Ithaca, N Cornell University Press. J83). '16,1982, p. 54. ch. 3.

294 293 Interna:ionar Compelkiver.ie:ss;n Elec:ronfcs ministrations. But regardless of swings of the financial community. At another level, the gov- pendulum, government intervention in capital ernment intervenes more directly. For exam- (and other) markets is a longstanding tradition ple, a particular bank's Joan limits might be in France, and the thrust of French policies as relaxed to allow /funds to flow to a favored firm they relate to investment is not likely to change or industry. Sometimes, companies receive radically_ graats or loans directly from the government.57 Selected firms and industries may also benefit On the whole, the French financial system from low interest rateor loan guarantees. seems more tightly controlled by government than the 'Japanese system, and certainly far The ability of the French Government to in- more so than in the United Statesin part :terve in capital allocation decisionsis because .so much of French industry is natio.n- facilita._ed becausemuch more than in the alized. One major d.fference between the United Statesthe savings function is split French and Japanese cases is that France has from the investing function. That is, the finan- been a great ideal more tolerant ..of foreign- cial institutions that take deposits differ from owned gnterprises. One-fifth of the .couratry's those that lend money. Typically, funds col- 200 largest firms fall in this category. When lected by deposit-takers are first lent to in- faced with weak domestic indwAries.includ- termediaries specializing in longer term. in7 ing, at various times, semiconductors and com- vestments. These intermediaries, in turn, lend puters the French Government has sometimes to private enterprises. Direct transactions be- chosen to allow, even encourage, foreign tween deposit-taking institutions and corporate ownership. borrowers are infrequc at. Table 66 illustrates the contrast with the United States. On the Industry in France, as in Japan, gets most of average, the deposits and loans of American itsexternal financing from institutional financial institutions are much more nearly lenders, generally banks. Relative to such countries as the United Stares or Great Britain, balanced. iV.most one-third. of the total value securities markets are small; within these of loans to nonfinancial concerns in France markets, sales of bonds far outweigh equities: "For examples in electronics. see, E. DiNlaria. "French Govt. Even the comparatively small French bond to Bolder IC Industry With Grants and Loans.- Electronic News, market serves mostly as a source of funds for Apr.. 27. 1981, p. J. government and far nationalized companies, plus financial institutions of various types. Manufacturing firms look predomiriantly to banks for their credit needs, roughly parallel- Table 66.Holders of Liabilities and Claims With the ing the situation in japan. Nonfinancial Sector in Mited Stn' France°

French industrial policy as it relates to invest- United StatesFrance ment is based on underpricing capital, then Deposittaxing institutions ks, using a mixture of formal and informal mech- savings and /loan*: Propoittion of total deposits 57.3% 71.7% anisms to ration funds to investors. Mar- Proportion of total loansrellaimis.. 51.8 48.3 ket-determined interest rates play a distinctly Long-term credi't institutions: Proportion Of total IN)05i'ls 8.2 secondary role. To some extent, lending insti- Proporlicl 7 9 32.9 4'1 tutionswhich must restrict total loans Investing T-F,iltutruns hr....surance companies, prescribed amount or be penalizedare free pension `ands): of total deposits 32.3 11:3 to choose recipients of funds. But the gt.rM. }-,rt.:ition of total loanslaalms 31.2 9.3 meat also has a voice, primarily through the Other financial Institistior6: Ministry of Finance. Proportion of total deposits q 8.8 Proportion of total toans,(clairns VI 9.5 The Ministry can act in -a number of ways. 8As of the) end of 1975. SOURCE: J. Zysman, Governments. Markets, and Growth: Financial Syolems and At one level, its influence is informally exer- the Politics rl Industrial Change (Ithaca, N.Y.: Cornell Universfry Press. cised through a network of contacts within the 1983),

2,9) Ch. 7Financing: Its Role in Competitiveness in Electronics 291 flow through specialized lona-term lending income less subject to the allocative dictates of stitutions. an indication oftheir importance. the government. Not only do successful firms get less attention simply by virtue of their com- Many of the institutions that take-deposits petitiveness, but such companies can more eas- or fend funds in France are at least quasi- ily generate capital internally, or tap interna- public; with the completion of the Mitterrand government's programwhich includes the national sources. tionalization of an additional 18 commercial Operating within these constraints, the barks, plus the country's two largest invest- French Government has used the financial sys- mfmt banking houses three - quarters of all tem to affect the country's electronics industry deposits pass through publicly owned banks.58 in two basic ways. Not only has the govern- This gives the bureaucracy many tools for in- ment supported the industry with both direct fluencing investment decisions. Even where and indirect financial assistance, butoften as financial institutions are private, the govern- a precondition for loans or grantsthe government can mediate between savers and investment has sometimes insisted that the industry ors: its most powerful weaponeven if seldom restructure. While restructuring has frequent- called onis simply the ability to undercut ly been aimed at fostering competitiveness, private lenders with public funds. maintaining employment has also been a Despite the pervasiveness of its influence motive. over investment decisions, the French Govei n- In electronics, the best known examples of ment faces severe eznstraints in employing this government- directed restructuring have been tool of industrial policy. French industrypar- associated with the Plan Calcul, which effec- ticularlyinhigh-technology sectorslike tively ended in 1976 with a merger of the com- electronicshas seldom been as competitive puter firms Compagnie Internationale pour in international markets as West German or L'Informatique (CII) and Honeywell Bull, the Japanese industry, much less American. While latter partially U.S.-owned. The French Gov- exceptions such as aerospace exist, the com- ernment promised as much as $700 million parative weakness of French corporations over a 5-year period to the new concern, along limits their ability to operate autonomously in with further assistance throw-1 purr' lases of world markets The French have lagged con- both hardware an'15,fi 10).14 such spicuously in MOS ICs and minicomputers; privy '; in France are encour- low-cost capital has not proved Tr ..ah help in ageu uehaN,e y s consistent with the building a strong technological base for the goals of the government; the carrots and sticks country's electronics industry. Indeed, several tend to be more visible than in Japan. of the larger French manufacturers are con- Franci5 now pursuing a similar strategy in trolled by foreign multinationals. Many more licm atronics. As discussed in detail in are partly owned by foralgil aliapte. 10, the aim is to develop an indigenous, American. ldii:1:=es have itadoded CII -Hon- internationally competitive industryin part eywell Bull and Matra-Harris Semicon- by encouraging joint ventures through which ducteurs. Subsidiaries of foreign-owned enter- American firms transfer technology toa prises need not depend on French financial in- French partner. One carrot here is the promise stitutions for capital; even minority ownership, of R&D funding reported to total more than which is often coupled with dependence en $300 million over a 5 year period. The U.S. foreign technology, gives substantial leve'rage partners have tended to view this as perhaps in dealing with the bureaucracy or with gov- their only route to sales within the large, ernment-controlled financial institutions. Fur- lucrative, and closed French telecommunica- thermore, French companies that prosper he- ions market. ''')uite apart from direct government aid, elec- "P. Lewis.r-rance Begins $8 Billion Takeover of Private In- tronics firms find that their status as teChno- dustry and Banking." New York Times, Feb. 1 1982, p. ',Ogical leaders compared to the rest of French

2 Fr: 292 lnternatic qal Competitiveness in Electronics industry makes it relatively easy to locate funds working relationships than typical in the for R&D or expansion. Through its pervasive United States. In particular, the stockholdings influence, France's bureaucracy can assure of banks are a major source of their very con- favored industries funding from either private siderable leverage over German companies of or quasi-public sources. the Aktiengesellschaft (AG) variety. The AG, or joint stock companies, were once far more But again as in Japan, the extent of govern- numerous in West German industry, and most ment influence on corporate financing has di- of the large concerns are still organized in this minished over the years. During the 1970's, fashion. Over the postwar period, the number companies meeting with success international- of Gesellschaften mit beschrinkter Haftung ly could deal with French banks and capital (GmbH)privately held, limited liability orga- markets largely free of government intervention. The government was more concerned nizationsgrew rapidly; banks interact less with firms and industries no longer able to closely with these. compete; to considerable extent, French in- One source of banking influence over the AG dustrial policy has been preoccupied with such form of corporation stems from its two boards sectors as textiles, steel, and shipbuilding. A of directors: the shareholders elect a super- good deal of assistance has gone to firms in visory board, which in turn appoints an ex- these industries, which have been depressed ecutive board, thelatter responsible for in France as in much of the developed world. operating management.'" The supervisory In this respect, the French Government has be- board makes major decisions on matters such haved like that in other industrialized nations, as new investments or product lines. While no including the United States.59 individual can belong to both the boards of a single company, there are no bars to simul- The French are now aggressively promo-lit, taneous service on the supervisory boards of potential technoln laders like electronics, several companies, even if they are com- hoping to e firms that might prove petitors. Officers of banks holding equity in a able to com,, internationaiiy (ch. 10). The West German firm often become di, ectors, and plans of the Mitterrand government are ex- a single bank may be represented on the super- traordinarily ambitious in the spending levels visory boards of a number of competing enter- proposed for the support of new industrial prises. About 10 percent of the members of the technologies, with iw.ich of the effort in elec- supervisory boards of the 100 largest AG firms tronics focused on semiconductors. France's are bank officersnor is this the only mech- record in attempting to promote technologi- anism by which West German banks influence cally advanced industries has thus far been mixed: disaster with the Concorde; success business activities.51 with the Airbus and helicopters, as well as Indeed, the role of banks is even more cen- nuclear power; notable progress in telecom- tral in West Germany than in Japan. There are munications; little relative movement as yet in three major reasons. The first is simply that microelectronics. Direct and indirect financial German banks are allowed to hold common subsidies may contribute to technological suc- stock, as they can in Japan though not the cess, but by the recent history in France are far from sufficient. Konica, "The Rise of the Modern Industrial Enterprise in Germany," Managerial Heirarchies: Comparative Perspectives West Germany on the Rise of the Modern Industrial Enterprise, A. D. Chandler. jr. ant! H. Daems (eds.) (Cambridge, Mass.: Hnrvard University Financial mediation in the Federal Republic Press, 1980), p. 77. again involves parties having much closer 6,See E. Hartrich. The Fourth and Richest Reich (New York: Macmillan, 1980), ch. 13; also IL Medley. "Monetary Stability "See 7.C. Price, IndUstriai Policies in the European Commun- and Industrial Adaptation in West Germany," Monetary Policy. ity (London:. Martin's Press. 1981), or an excellentdiscussion Selective Credit Policy, and Industrial Policy in Franca, Britain, of how governments in Western Europe have attempted to deal West Germany, and Sweden, staff study, joint Economic Com- with the problems of distressed industries. mittee, Congress of the United States, June 26, 1981, r. 92.

29 y Ch. 7Financing: Its Role in Competitiveness in Electronics 293

United States. But, whereas Japanese banks are soning is virtually identical to that now heard limited to a maximum of 5 percent ownership in the United States concerning japan: financ- in a single company, the holdings of German ing costs are lower because the major German banks have not thus far been restricted firms make heavy use of low-cost bank loans. (although such legislation has been considered by the parliament in recent years). The second As for highly L--veraged Japanese firms, reason relates to financial structure. German however, the magnitude of any advantage in industrial companies, again like their Japanese cost of capital will depend largely on the tax counterparts, are highly leveraged, tending to benefits accompanying a high proportion of rely on bank loans rather than bonds. On the debt in a German company's financial struc- average, firms in the Federal Republic carry ture. To the degree that inside knowledge and about four times as much debt as equity on control provide better information, the involve- their books.e2 The high proportion of debt is ment of banks in the management of German even more striking in light of the third char- companies may also lower their perceived acteristic of West German financial practice: risks. If so, the banks might choose to lend on this debt is heavily concentrated in the port- better terms. The same could be said about folios of only three banksthe Deutsche Bank, Japan, although the limits on direct ownership Dresdner Bank, and Commerzbank, all need- make it a less significant factor. In any case less to say very large. These three banks hold West German or Japanonly small reductions seats on the boards of 70 of the 100 largest Ger- in interest rates could plausibly flow from such man corporations." sources. Beyond the shares they own, banks in the The deep involvement of West German banks Federal Republic frequently control proxy in corporate financing has its corollary in a rights on privately owned shares' carried in relatively underdeveloped capitalmarket. their vaults. These shares combine with Direct While stock exchanges exist, trading volumes equity ownership to create an impressive con- are much lower than in the United States. Sec- centration of voting power. The banking com- ondary markets for other types of financial in- munity can vote more than 90 percent of the struments are rare. In fact, for most transac- shares for many of the large publicly held cor- tions in German financial markets, orders must porations in West Germany." Because major be placed with the banking community. The decisions must be approved by at least 75 per- usual modes of personal saving are bank ac- cent of the shareholders, an effective veto is counts, insurance, and pension funds. Bonds, held by one or more banks if they control only-, including those of financial institutions, ac- a quarter of a company's common stock. Even count for less than 15 percent of savings, equity more so than in Japan, banks in the Federal ownership less than 1 percent. Republic absorb clear and explicit equity risks. While banks carry great weight in the Federal Republic, government influence on the financ- West Germany's competitors in Europe fre- ing of private business is far less pervasive than quently complain over the relatively high in France or japan. Although government own- leverage of German corporations, focusing on ership of business accounts for about one-fifth capital costs, together with the major holdings of all fixed capital investmentabout the same by banks of both equity and debt.65 Their rea- as in the United Kingdompublic ownership "J. Ross-Skinner, "Germany's Powerful Banks," Dun's Review, in Germany is largely confined to the transpor- January 1979, p. 68. tation, electric power generation, coal, chem- "Medley, op. cit., p. 115. icals, and shipbuilding sectors." In recent "M. Kreile, "West Germany: The Dynamics of Expansion," Between Power and Plenty: Foreign Economic Policies of Ad- vanced industrial States, P. J. Katzenstein (ed.) (Madison, Wis.: E. Owen-Smith, "Government Intervention in the Econcmy University of Wisconsin Press, 1978), p. 191. of the Federal Republic of Germany," Governmental Intervention "E. V. Morgan and R. Harrington, Capital Markets in the EEC: in the Developed Economy. P. Maunder (ed.) (London: Groom The Sources and Uses of Medium- and Long-Term Finance Helm, 1979), p. 160. The investment figures were 21 percent in (Boulder, Colo.: Westview Press, 1977), p. 323. Germany for 1972 and 19 percent in the United Kingdom for

298 294 International Competitiveness in Electronics

years, some of the government's holdings have The picture that emerges, therefore, is one been sold to private interests. Both Federal and of close working relationships between West Liinder (state) Governments maintain owner- German industry and the banking sector. Com- ship interests in banks, but their primary con- mercial banks provide the bulk of external cern appears to be financing projects involv- financing for companies in industries like elec- ing housing, agriculture, and small business. tronics, and take a correspondingly active role Government-owned banks do not compete for in management. In terms of control over the business with private banks. While direct gov- nation's economic activities, the banks occupy ernment subsidies to industry are substantial, a central position and wield considerable amounting to several billion deutsche marks power. While a variety of political interests annually (about half as .much in dollar:.), most have recently pressed for reductions in the of the subsidies have been directed toward so- presence of the banks, thus far change has been cial welfare objectives or the support of indus- minimal. Government, in contrast, plays a less try in West Berlin and the border areas. Still, dominant role than in many other industrial- Federal funds have aided aircraft design and ized nations. Recently, the willingness of the production, along with shipbuilding, coal min- West German Government to let market forces ing, and housing construction. determine economic direction has come under strain. The eventual consequence may be a (footnote continued from p. 293) greater degree of intervention in microeconom- 1975. See Public Enterprise in the EEC. 1Y. Keyser and R. Windle ic matters (ch.10). (eds.) (Alphen aan den Rijn. The Netherlands: Sijthoff and Noord- hoff. 1978). Part 111Federal Republic of Germany. p. 3 and Part VIIUnited Kingdom and Ireland. p.

Summary and Conclusions

This chapteras several othersconcen- Although Japanese electronics companies con- trated on the U.S.-Japan comparison because tinue to utilize greater financial leverage than Japanese companies are the most effective and American firms, the advantages of this prac- aggressive competitors in the world electronics tice are marginal at best. Higher debt-equity industry outside the United States. ratios do not give Japanese electronics firms From the narrow viewpoint of financing signific: ;e, benefits in financing compared to costs, it appears that government support for Amer'''. 3n manufacturers. Japan's electronics firmsnow manifesting Other analyses have resulted in estimates for itselfparticularlyinsemiconductors and the difference in cost if capital between the computershas resulted in somewhat lower United States and Japan that are considerably costs of capital than would otherwise be the larger. The explanation lies in expectations case. But in real, inflation-adjusted terms it is concerning future price inflation in the two unlikely that this cost of capital advantage ex- countries, which other studies have not fully ceeds a few percentage pointsalmost certain- considered. To gain "real" returns, lenders ly less than 5. By itstif, the effect would be to must demand interest rates in excess of the in- make expansion somewhat less costly for Jap- flation rate. Price inflation in the United States anese firms, but the competitive advantage has exceeded that in Japan by considerable gained from this source alone would be small. margins over the past few years, and the in- Differences in costs of capital faced by firms flationary expectations of lenders have re- within the industries of the United States or flected this history. The direct consequences Japan are larger than the differences between for costs of capital are perhaps less important average costs of capital in the two countries. than the effects on time horizons for corporate 29, Ch. 7Financing: Its Role in Competitiveness in Electronics 295 investment decisions. High and uncertain fu- law. But competition is likely to hold down pro- ture inflation rates lead managements to an- fitability, thus limiting the ability of American ticipate large swings in thecash. flow returns electronics manufacturers to finance rapid ex- from capital. The longer the time horizon, the pansion through reinvested earnings. greater the possibility that, at some future Compounding the difficulty for firms in point, the returns will be insufficient to cover many portions of the industry is the rising level interest expenses. Such uncertainties bias inof capital intensity. More expensive production vestment decisions in the United States toward equipment is a fact of life for semiconductor the short term. firms. Equipment used for R&D as well as Still, even if the real, inflation-adjusted costs manufacturing rapidly becomes outdated. This of financing are not that much higher here than is not necessarily a problem so long as writeoffs in other parts of the world, costs of capital are keep pace. But the changes in depreciation great enough to create serious dilemmas for the schedules adopted as part of the 1981 Tax financial managers of American electronics Actwhich fix depreciation on most equip- firms. These dilemmas stem from the limits on ment at 5 years for all industriescould disad- debt broadly acceptable within U.S. financial vantage electronics firms. In the past, deprecia- markets, the need for rapid growth in capital- tion schedules were based, at least nominally, ization to meet expanding market demand for on the actual useful life of the investment. The electronics products, rising capital-intensity in new law shortens depreciation schedules for some parts of the industry, and heightened other industries, where plant and equipment foreign competition. often have much longer useful lifetimes. With all industries now depreciating on essentially In addition to greater capital equipment equivalent schedules, firms in electronics and costs, manufacturers of computer3, semicon- other technologically dynamic industries are ductor devices, and related products face likely to find themselves at a relative disadvan- mounting expenses for R&D and product de- tage. Their domestic rivals for investment velopment simply as a result of advances in funds benefit from greater increments in de- technology and the increasing complexity of preciation rates and hence in cash flows, electronic systems. As in the past, competition augmenting their ability to attract capital for will be strong even among domestic firms; nt:w investment. added competitive pressures will come from the Japanese. When the industry was small, and U.S. electronics firms obtain financing from new startups drove the technology, competi- a variety of sources, depending largely ontheir tion was a vital source of U.S. strength. Now size and stage of development. For many of the that the industry is maturing, the ingredients younger companies, the original source was of success are changing. Managers of Ameri- the venture capital marketwhere investors can firms are reassessingtheir business provide money to infant businesses in the hope strategiesparticularly in terms of pricing of greater returns than safer investments would while finding themselves hard-pressed to yield. Venture organizations generally expec't finance R&D and new production facilities. most or their return in the form of capital appreciation; as a consequence, their investment In :ecent years, American industry has not decisions are sensitive to taxation of capital raise:i much capital from sales of stock. To in- gains. The 1978 reductions in capital gains crease equit, without selling stock, electronics taxes were an important force in enlarging the firms must generate substantial flows of re- pool of funds avail:ible for new ventures in tained earnings from profits and depreciation. electronics. To some extent, the depreciation schedules implemented by the Economic Recovery Tax Act As successful firms grow beyond the infant of 1981 will increase cash flows available for stage, they gain access to a greater variety of reinvestment, as will other changes in the tax sources of capitale.g., lines of credit with 296 International Competitiveness in Electronics banks, bond markets. They may also be able growth in the pool of capital available for in- to float public stock offerings. But manage- vestment. If effective, such programs should ments of electronics companies, following the result in lower market interest rates. But many prevailing American pattern, have strongly of the steps taken will also stimulate demand preferred internal funding of growth. Some for funds. It has not yet been possible to companies utilize considerable debt (table 53), separate the effects on supply and demand but leverage in U.S. electronics firms for the flowing from these policies; the vital question most part remains low. of how future U.S. costs of capital will com- Such financial patternsparticularly those pare with costs in other countries, also uncer- established by merchant semiconductor man- tain, takes the whole matter a step further. ufacturers during the 1960's and 1970'swill Turning to japan, in years past capital cosh: not be easy to maintain during the 1980's. for electronics firms there may have been sig- Greater demands for capital equipment and for nificantly lower than in the United States. The R&D are combining with intense foreign com- reasons are severai. Capital has, at various petition to make the finar,cing of growth by times, been channeled to favored industrial sec- small, independent companies more prob- tors, including electronics. By controlling in- lematical. But despite the attention focused on terest rates, the Japanese Government effective- international competition, the fundamental ly circumvented the market as a mechanism problem is growthtogether with the upswing for allocating funds. But while remnants of this in capital intensity. Many once-independent control remain, government influence over fi- firms have already been acquired by larger cor- nancing decisions by banks is now much porations, at least partially in consequence of weaker than 20 or even 10 years agO. Further- their needs for capital. more, as capital markets in Japan continue to As a result of these forces, the U.S. elec- evolve they will take forms more like those in tronics industry will almost certainly be com- other industrialized countriesi.e., market- pelled to rely more heavily on external funds. determined interest rates will become the ma- This is one of the reasons for the concern over jor mechanism for capital allocation. Stronger costs of capital. Many industry leaders have linkages with capital markets in other countries expressed doubts that supplies of capital will will mean that rates of returnand hence costs be adequateor that, if capital is available at of capitalwill not be much different in Japan all, the costs will be too high, particularly com- than in the United States. Thus, even if Japa- pared with costs of capital in Japan. Of course, nese electronics firms enjoyed lower costs of funds will always be available for investments capital in years past, these advantages are likely that promise sufficiently high returns. Free to diminish in the future. capital markets will clear at some interest rate, and it is this interest rateor pricethat serves At the same time, the savings rate in Japan as the allocative mechanism in the U.S. finan- continues to be extraordinarily high, though cial system. But it is quite possible thatfrom declining somewhat at the end of the 1970's. a broader perspective than simply returns to It may continue to gradually fall, but rates'of capitalprojects that are otherwise desirable capital formation remain much greater than in will not be funded. Examples include the long- the United States even considering increasing term basic research that undergirds a high- public sector demands as a result of large technology industry like electronics. budget deficits. Moreover, Japan's Government has a well-practiced capability for intervening As a result, the question of whether interest in capital markets and steering investment rates in the United States may be prohibitive- toward favored sectors of the economy. What ly high is a difficult one. The supply-side thrust the Japanese have called "administrative guid- of programs initiated by the Reagan adminis- ance" will not simply disappear. Still, the tration was intended to produce significant changing character of the country's financial

Ul Ch. 7 Financi'sg: Its Role in Competitiveness in Electronics 297 market's means that Japan's electronics firms rowing has risen. Government bonds are be- will have less of an advantage in the future coming major long-term tradable securities. compared to their rivals in the United States Securities of many types, both domestic and and Europe--some of which may even find international, are becoming more widely avail- themselves tapping sources of capital in Japan able in Japan, and market forces are having to finance their own expansion. their effects on interest rates. The government As these trends proceed, major Japanese cor- will have more difficulty in managing invest- porations will no doubt continue to diversify ment flows, and the capital structures of their capital structures, relying less heavily on Japanese electronics firms will continue mov- commercial banks. Corporate borrowing in ing closer to those in the United States. Assum- Japan as a percentage of gross national prod- ing these trends continueand there is every uct is declining as firms diversify their sources reason to expect them toat least some of the of funds. The leverage of Japanese electronics concerns of American industry with respect io companies gradually decreased over the 1970's, Japanese financial practices should diminish. as figure 52 showed, while public sector bor-

302

99-111 0 - 83 - 20 CHAPTER 8 Human Resources: Education, Training, Manag -mend Contents

Page Overview 301 Educatipn and Training 302 U.S. Secondary Sch 001 Education in Science and Mathematics 303 Secondary Schooling Abroad, Especially in Japan 304 University and Continuing Education in the United States 305 University and Continuing Education in Japan 314 International Differences in Education and Training 317 Supply and Mobility of Labor 318 Overall Labor Market Trends, 319 Personnel Supplies for the U.S. Electronics Industry 320 The Question of Mobility 321 Organization and Management / 326 Organizational Types and Management Styles 327 Worker Participation 330 Japanese and American Management Styles: How Much Difference? 333

Summary and Conclusions ; , 334 Appendix 8A. Japanese and AMerican Managethent Styles: A Comparison 33')

Listof Tables Table No. Page

67. Engineering Graduates asja Percentage of Their Age Group 306 . 68. U.S. Degrees warded by Field 306 69. Enrollments in Japanese/Colleges and Universities 315 70. Distribution by Size of Japanese Firms Providing In-House Training 316 71.- Labor Force Growth in/Several Countries 319 72 Work Stoppages Due to Labor Disputes in Several Countries 328 73. Rankings by American Managers of Factors Contributing to Productivity 335 8A-1. Responses of Middle and Upper Managers to Questions Dealing With Communications and Decisionmaking Styles 337 RA-2. Data Related to Employee Satisfaction 338

Listof Figures Figlin? No. Page 53. R&D Engineers and Scientists in the Labor Force 305 54. Engineering Graduates of American Universities 306

201 CHAPTER 8 Human Resources: education, Training, Management

Overview

Among the questions this chapter addresses for technology-intensive industries like elec- are: How good are the people an industry de- tronics? Do managements make the best use pends on? Is the pool from which they are of the talents and abilities of the labor force? drawn big enough? How do they get their train- Are countries like Japan doing anything that ing? And, the mirror image of these: Does in- is really differentor better? In the early part dustry use their abilities wisely? of the century, these .questions were, already being asked, as part of the "scientific" study Countries without adequate human resources of management. Itis no coincidence that cannot hope to design and manufacture prod- American management experts schooled Japa- ucts like computers; even televisions are be- nese executives now known for their dedir!a. yond the capabilities of many developing econ- tion to quality (ch. 6). omies. In the-United States, peopleunskilled or skilled workers, engineers and technicians, The popular press tends to oversimplify the managersare a vital resource for electronics set of issues covered by "human resources." firms; thriving semiconductor companies have Some commentators define human resources been built around the talents of three or four narrowly, as encompassing the skills and at- engineers. titudes of the work force; this approach often But people are only the starting point. How leads to stereotyping of employees in countries talents are developed, skills utilized, depends like Japan Or West Germany. Seeing the Japa- largely on management: managers shape the nese worker as the product of a culture that organization, decide on policies, set the style rewards hard work and diligence captures part and tone. The sections that follow examine of the truth but obscures the larger institutional human resources as a factor in competitive- and economic context. Others stress rtianagii- ness, primarily from the standpoint of elec- ment techniques, often narrowly defined, as F. tronics in the United States. Matters of educa- key to labor productivity. Quality control tion and training are followed by an examine.. cies are the best-publicized current exam;.!?. tion of management practices. One of the ques- While certainly critical in the utilization c tions addressed is: To what extent does the firm's human resources, management should vogue for Japanese management represent any- also be viewed as part of a broader picture. thing-new and different in the American con- Management practices themselves reflect a text, as opposed to a reemphasis of themes that mix of schooling and experience shaped by t-th. have always been present? The comparisons structure of work and organization within a on education also focus on Japan, in part be- society. cause of the recent publicity given to that coun- This chapter views the American labor force try's lead over the United States in numbers and the electronics industryfrom two of engineers graduated. damental perspectives. First, workers brir48 Such topics are particularly appropriate at with them a set of skills largely acquired pcitir a time when rates of productivity growthhave to joining a company. The question then is to slowed in the United States. Is the education compare education and training in the United and training of American workers appropriate Statesparticularly of white-collar personnel,

305 301 . 302 Internation veness in Electronics but also blue-collar employeeswith that of the with which the United States has experimented men and women who staff foreign electronics largely for reasons of social welfare, a strong firms. Second: Will there be enough appropri- case can be made, for an enhanced Federal role ately trained people to meet the needs of a rap- in training and retraining programs to support idly expanding U.S. electronics industry? the competitiveness of growing high-technology industries like electronics. Labs': -.T.obility is a separate but related issue. A growing industry, such as semiconduc- The other perspective on human resources tor manufacturing, may be able to meet its in this chapter relates to corporate manage- manpower needs by attracting workers from ment. Contrasting the practices of Japanese other parts of the economy. Within the in- and American managers shows many of the dustry, one semiconductor firm may be able lessons of effective management to be univer- to lure employees from its competitors. Mobili- sal, the unique character of Japanese manage- ty has traditionally been high in the United ment sorneffiing of a myth. Nonetheless, there States for those with knowledge and experi- are lessons to be learned from firms in Japan, ence. as well as from successful organizations in the United States. Competitive firms here and But what of those left behind by technologi- abroad tend to share a common trait: man- cal change? To a considerable extent, other na- agement practices that give employees a say tions have used retraining programs as, instru- indecisions affecting their work, along with ments of public policy for enhancing employee support for skill development. Emphasis on mobility and aiding those whose skills are out- employee participation and human relations of-date. This has been less common in the can contribute to productivity and worker sat- United States, where mobility and continuing ' isfaction, but conclusive evidence linking par- education depend on individual initiative. , ticular management techniques (such as quality Leaving aside questions of remedial education control circles)here or in any countryto and the training necessary for entry level jobs, competitive success is conspicuously lacking.

Education and Training

The U.S. electronics industry is built on the draftsmen and nondegree designers, 'produc- capabilities of production workers, skilled tech- tion forelnen, field service installers and repair- nicians, and white-collar managers and profes- men, computer system operators, and techni- sionals. On the shop floor, blue-collar employ- cal writers. White-collar workersmany with ees operate semiconductor fabrication equip- college degreesperform functions ranging ment, assemble computers or TV sets. Much from plant management to accounting and fi- of this work is essentially unslcilled, meaning nancial control, business planning, and legal that a typical job can be learned in a few hours. advising. Engineers and scientistssome with Techniciansgrey-collar employeesoften advanced degreesdesign and develop prod- with vocational school training, play an impor- ucts, plan manufacturing processes, specify tant role both on the factory floor and in reproduction equipment, and carry out R&D proj- search and development (R&D) laboratories. ects in fields ranging from solid-state physics They maintain, troubleshoot, and repair sophis- to computer architectures. All of these skills ticated equipmentand sometimes fabricate are essential to a competitive industry, not just itas well as testing and inspecting com- those of the well-educated and well-paid proponents and systems. Technicians also build fessionals; grey-collar technical workers, in and help develop prototypes of new products. particular, have a critical place in technol- Other employees with specialized skills include ogy-based organizations. Some jobs depend Ch. 8Human Resources: Educatioa, Training, Management 303 much more heavily on formal education and clear, beginning at secondary levels where training than others, but it is fair to say that students avoid courses in these subjecto. Only better skills and abilities at all levels will add one-sixth of U.S. secondary school stu( -nts, for to the competitive ability of an enterprise, as example, take courses in science or mathemat- well as adding to peoples' upward mobility. ics past the 10th grade. Technology, as opposed The United States has maintained a lead in to science, is totally lacking in secondary many fields of technology and science since schools, despite the abundant evido:ace of pub- World War II, in large part because of the lic fascination with technological achieve- excellence of the educational system here. ments. Indeed, few people seem to distinguish Nonetheless, other advanced industrial nations technology from science, hence misnomers provide their work forces with training in tech- such as science fiction. nology, mathematics, and science that on the The NSF/DOE report. along with many average is probably more intensive. It is easy others, also points to apparent shortages of to forget, in the publicity that surrounds Nobel entry-level computer professionals and several Prizes, the Apollo program, or the nascent bio- types of engineers, and the difficulties of sec- technology industry, that competitiveness rests ondary schools, vocational institutes, commu- on the skills and abilities of great numbers of nity colleges, and universities in finding and people whose contributions will never be pub-a retaining qualified teachers in the physical sci- licized or even acknowledged. At a time when ences, mathematics, engineering and computer literacy levels in the United States decline as science-, and in vocational programs. More- those elsewhere rise, and the Soviet Union over, equipment used for teaching laboratory graduates five times as many engineers, it courses in engineering and the sciences is makes sense to look at the fcundations for the years out of date and in short supply. In the Nation's human resources as well as the pin- future, American industry, T ,cularly high- nacles of its achievements. technology sectors like electr,, A;s, may sim- In fact, the evidence of U.S. weakness in ply not have an adequate supply of employees technical education and training is strong and with the kinds of skills needed 1,) inaintain U.S. continuing to mount.1 The best people and best competitiveness. educational institutions in the United States are probably as good as ever, maybe beter. But the U. SecondarySchool Education breadth of capability that once distinguished in Science and Mathematics the U.S. labor force may be diminishing. The Falling mathematics, and science enrollments National Science Foundation/Department of in American high schools indicate that, while Education (NSF/DOE) report cited above con- there is a small group of students who want cludes that American achievements in basic re- and get advanced courses, the great majority search remain unchallenged, but that the aver- avoid these subjects when they can. Average age high school or college graduate in this scores on national tests of achievement in country has only the most rudimentary knowl- mathematics and the sciences are lower than edge of mathematics or science., The trends are a decade ago. Students who elect to take Ad- ,"Science and Engineering Education for the 1980's and Be- vanced Placement Tests in science or mathe- yond." National Science Foundation and Department of Educa- matics make about the same scores as in the tion. October 1980. See also Today's Problems. Tomorrow's Crises: A Report of the Natiohal Science Board Commission on past, indicating that the core of serious stu- Precollege Education in Mathematics. Science and Technology dents gets good preparation; but overall, Scho- (Washington. D.C.: National Science Foundation, Oct. 18. 1982): lastic Aptitude Test (SAT) scores fell for 18 con- Science and Engineering Education: Data and Information, NSF 82.30 (Washington, D.C.: National Science Foundation. 1982),. secutive years until holding steady in 1981.* and Science Indicx ',r5-1980 (Washington. D.C.: National Science Board. Natio, Science Foundation, 1981), chs. 1 and According to the Educational Testing Service, Princeton. N.J., 5. The U.S.-U.S.S.R. comparison in engineeringswaduates comes mean scores in 1981 for co./lege-bound high school seniors were from p. 209 of the last-mentioned report./ 424 for the verbal portion of the SAT and 466 for the mathematics 301 304 International Competitiveness in Electronics

Some of the de ;line can be attributed to the quired to complete 2 years of mathematics plus greater percentage of students who now attend 2 years of science. Competition for entry into college and thus take .tie tests, but an advisory' the best colleges is intense; Japanese students panel convened to examine the SAT concluded choose rigorous electives and spend much that, since 1970, other factorsincluding lower more time on homework than their American educational standards and diminishing motiva- counterparts. Those who wish to attend college tion on the part of studentshave been much study mathematics each year, moving beyond more important.2 trigonometrythe point where many U.S. high school curricula still stop.5 The stress in Japa- Fewer American high school students are nese secondary schools on science and mathe- electing mathematics and science courses, par- matics for all students is far from unique. The ticularly the two fundamental physical sci- Soviet equivalent of the. American high school ences, chemistry and physics; of those who do curriculum includes a heavy dose of course- elect science, more chose the life sciences. work in these areasfor instance, 2 years of While the majority of U.S. high school grad- calculus. West German secondary school stu- uates have taken biology, only about a third dents, even those who wish to specialize in have had cheinistry; the fraction drops to about fields such as the classics or modern languages, one-tenth for physics.3 The situation is repli get extensive training in mathematics and sci- cated in high school mathematics, where only ence; by the same token, those planning tech- One-third of U.S. graduates take 3 years of nical careers receive their liberal arts educa- coursework. Regardless of how good their tion in high school. Neither curricula, nor aca- grades may be, three-quarters of American demic standards vary as widely among West high school graduates do not have the prerequi- German schools as in the United States.° site ,bourses to enter a college engineering program.4 What this means for industries like In Japan, large numbers of students v. ho do electronics is not only that the average i';(711 not go to college get technical, vocational, or school graduate is unprepared to v semiprofessional schooling as preparation for neering or one of the physical :.)ciences ,n col- jobs in industry where they will work with and lege, but may be unable to enter a career call- provide support for engineers and scientists. ing for middle-level technical skills without a The result is a large pool of well-prepared can- good deal of additional training. didates for entry-level grey-collar jobs.? The investments that students in Japan make Secondary Schooling Abroad, in science and mathematics yield measurable Especially in Japan benefits. On a number of international achievement tests, Japanese students score consistent- U.S. enrollments in science and mathematics ly above their counterparts in other industrial contrast starkly with the picture in japan. Not nations.° Nonetheless, secondary education in only do about 90 percent of Japanese high school students graduatecompared with 75 °M. W. Kirst, "Japanese Education: Its Implications for Eco- nomic Competition in the 1980's." Phi Delta Kappan, June 1981, to 80 percent in this countryhut all are re- p. 707. Only about 30 percent of U.S. high schools offer cLIculus, and fewer than 10 percent of American high school students (footnote continued from p. 303) take the subject; see Hurd. op. cit.. and Science and Engineer- pertion, identical to 1980 scores. In 1966. the means were 466 ing Education: Data and Information. op. cit., p. 59. for verbal and 492 for mathematics. While testing criteria may 6Engineering Our Future: Report of the Committee of Inquiry not have remained precisely the same over this period, the down- into the Engineering Profession (London: Her Majesty's Station- ward trend is unambiguo:Is. ery Office. January 1980), p. 219. Also. D. W. Sallet, "Education 2"Science and Engineering Education for the 1980's and Be- of the Diplom Inge-nieur," Journal of Engineering Education, yond,- op. cit.. pp. 107-108. vol. 59, June 1969, p. 1105. 'P. D. Hurd. "Falling Behind in Math and Science," Washing.: 'S. B. Levine and IKawada, Human Resources in Japanese ton Post, May 16, 1982, p. C7. See also Science and Engineer- Industrial Developmenat (Princeton, N.J.: Princeton University ing Education: Data and Information. op. cit., pp. 57, 59. Press, 1980), pp. 74, 80. Engineers in Japan are evidently sup- "Engineering: Education. Supply/Demand and Job Opportu- ported by many more technicians than in the United States.- nities." Electronic Industries Association, Washington. D.C., Oc- °R. S. Anderson, Educatit.:1 in Japan (Washington. D.C.: U.S. tober 1982. Government Printing Office, 1975), p. 130.

30d Ch. 8Human Resources: Education, Training, Management 305

japan has major weaknesses. The most obvious programs have been turning out greater num- is the strong traditional emphasis on rote learn- bers of engineers since 1967. In 1981, Ja- ing and imitation, coupled with a depenr!ence pan graduated 75,000 engineers compared to on textbooks and lectures rather than demon- 63,000 here, despite a population half that of stration and learning-by-doing (in reality, U.S. the United States. The margin is a little greater education is probably no better in this regard). for electrical engineering graduates-25 per- Critics of the system argue that this stunts the cent.1° developrnen' of creative abilities.° Academic competition L-t hpan is, furthermore, so in- As figure 53 shows, the United States once tense that the Jar nese. Ministryof Education held a commanding lead in the proportion of exprc3sed concern that other aspects of engineers and scientists in the work force. child development are being negleCted.- Despite While the advantage over other Western na- the undoubted validity of some of these Criti7 tions probably still exists (various countries cat- cisms, the fact remains that high school stu- egorize scientists, engineers, and technicians dents in Japan receive training in science and difterently, making comparisons. ambiguous), mathematics that is, on average, more exten- it has narrowed greatly, And, as table 67 sive than in the United States. Even for stu- demonstrates, engineering graduates are now dents who do not go on to technical or profes- a smaller proportion of their age group in fir sional jobs, such training contributes to quanti- United States than in Japan or West Ger- tative skills, precision in thinking, and to an Inatlycountries where a far greater fraction of engineers in any case devote their efforts to understanding of the physical world. ,;.it;11tt backgroun, 1 helps people to comprehend the commercial rather than defense industries. technologies that their dEily lives depend 'nn. In the future, their employment opportunities may depend on this as well. '°The 1981 breakdown by disciplines is not available fc,r Japan, but in 1980, 19,355 B.S level degrees were awarded in Electrical and computer engineering, compared to 15,410 in the United States. Figures for Japan are from the Ministry of Education, University and Continuing Education those for the United States from P. Doigan, "Engineering and Technology Degrees, 1981," Engineering Education, April 1982, in the United States p. 704, and P. Sheridan, "Engineering and Technology Degrees, In some respects, the Japanese and American 1980," Engineering Education, April 1981, p. 713. educational systems are opposites. The Japa- nese concentrate their efforts on precollege training where the United States is weak. On Figure 53.R&D Engineers and Scientists the other hand, the quality of university educu- in the Labor Force

tion in Japan is much inferior. In a very real U, sense, the American system of higher education must compensate for secondary schooling "2 1422°3o vr-3, 0.7 that is generally poor. 0 "2 0.6 N CD Although this comparison may be qualitative- fr; 0.5 West Germany a)C ly valid, it begins to break down in terms of Em0a .4 numbers. While the United States continues to a) cr 0.3 produce more Ph. D.s in science and mathe- "6 .E CM .00.2 matics than Japan, Japanese undergraduate aCh 0.1 °See, for example, the assessment of M. IV3gai, former Japanese CD CM 1-2 Minister for Education: "Higher Education in Japan," Japan `mCD Quarterly, vol. 24, 1977. p. 308. While many Japanese are quite a- 1967 1969 1971 1973 1975 1977 19791981 self-conscioLs about their country's supposed lack of innova- Year tion and originality in engineering and the sciences, the prod- aLower bound estimate. uct developmeids flowing in recent years from Japan's industries SOURCE: National Patterns of Science and Technology Resources, 1982 show great creativity in the application of technology. (Washington, DC: National Science Foundation. 1984 P. 33. 3 0J 306 International Competitiveness in Electronics

Table 67.Engineering Graduates as a Percentage Figure 54.Engineering Graduates of of Their. Age Groups American Universities 70.000 United States 1.6% Japan 4.2 60.000 West Germany' 2.3 France 1.3 50,000 United Kingdom 1,7 aFirst devee graduates, Including foreign nationals, In 1978, except for West 73 40.000 Gerrn.Jiny and France, where the percentages refer to 1977. In the United States, a significant fraction of engineering graduates are from overseas: in 1982, 8 percent of B S. degrees in engineering wont to foreign students, 29 percent of M.S. 30,000 iiiigiees, and 40 percent of Ph. D. degrees. See P. J. Sheridan,'"Engineering and Technology Degrees. 1982," Engineering Education. April 1983, P. 715. SOURCE: Engineering Our Future: Report of the Committee of Inquiry into the n20 000 Engineering Profession (London: Her Majesty's Stationery Office. ' January 19801, p. 83. Doctoral degrees (:.&00

0 EngineeringEluctition 1965 1970 1t:5 1980 1982 Year As table 68 indicates, graduates in engineer- SOURCES: 1965-79"Data Related to the Crisis in Engineering Education," American Association of Engineering Societies, March 1981, p. ing, the physical sc:iences, and mathematics in 17. 1980P. J. Sheridan, "Engineering and Technology Degrees, 1980," the United States accounted for steadily fall- Engineering Education, April 1981, p. 713. ing proportions of new degrees at both under- 1981P. Dolgan, "Engineering and Technology Degrees, 1981," '7ngineering Education, April 1982, p. 704. graduate and graduate levels during the 1970's. 198ZP. J. Sheridan, "Engineering and Technology Degrees, 1992," The number of degrees in the mathematical sci- Engineering Education, April 1983, p. 715. ences, including statistics and computer science, actually fell between 1970 and 1980. in 1982."At the graduate level, the trends are quite differentbut not encouraging. The num- In engineering, undergraduate enrollments ber of master's degrees in engineering has in- have jumped since the mid-1970'sand the creased slightly over the past decade, but the number of graduates has followed, as shown numb.n. of Ph. D.s has declinedone reason in figure 54leading to overcrowded classes, for faculty shortages in engineering schools.. overloaded faculty, and severe pressures on the Figure 54 illustrates the trends at both B.S. and quality of education. The number of full-time Doigan, "Engineering Enrollments, Fall 1982," EnginCer- undergraduates enrolled in U.S. engineering ins.; Education,October 1983, p, 18. At the bottom of the most schools went from about 20,000 in the early recent trough, in 1973, 187,000 students were enrolled in engi 1970's town all-time high of more than 400,000 veering; by 1982, the total was 403,000.

Table 68.U.S. Degrees Awarded by Field Total as percentage Physical of degrees awarded Engineeringsciences Mathematicsa in all fields 1960: B.S. 37,808 16,057 11,437 175/D M.S 6,989 3,387 1,765 17 Ph. D. 786 NA NA NA 1970: B.S. 42,966 21,551 29,109 12% M.S 15,548 5,948 7,107 13 Ph. D. 3,620 4,400 1,222 31 1980: B.S. 58.742 23,661 22,686 10% M.S 17,243 5,233 6,515 10 Ph. D. 2,751 3,151 963 21 NA ., Not Available. alncluding statistics and computer science. SOURCE: EngineeringData Related to the Crisis in Engineering Education," American Association of Engineering Societies, March 1981, p. 17; Physical Sciences and MathematicsNational Patterns of Science and Technology Resources 1981, NSF 81-311 (Washington, D.C.: National Science Foundation, 1981j, pp. 78-80.

, t 1 Ch. 8Human Resources: Education, Training, Management 307

Ph. D. levels. Not only have doctoral enrollments failed to keep up, but about half of all Ph. D. engineering candidates are now foreign nationals; many of them leave the United States after graduation.* An important cause of declining enrollments of Ph. D. candidates in engineering has bean the high starting salaries that holders of new bachelor's degrees commandin 1082, about $26,000. Swelling demand by industry for engineers has attracted undergraduates to the field, at the same time siphoning many off from the pool of prospective graduate students. To someone who might otherwise consider a Ph. D. followed by a teaching career, the re- Photo credit: General Motors wards of immediate employment can seem Engineer holding bracket designed with much more attractive than several years of low- computer assistance paying stipends or graduate assistantships, then the salary of a junior faculty member. Even when funds hive been available to hire While pay for college teachers has always been new faculty, good candidates are rare because well below that in industry, the other attrac- of the low numbers of new Ph. D.s and the un- tions of an academic career have diminished competitive salaries offered by universities. in these days of overcrowded classrooms, out- Estimates of the number of unfilled faculty dated equipment, and limited research funding. positions in U.S. engineering schools have been Poor facilities and an escalating student-to- in the range of 1,400 to 2,000about 10 per- faculty ratio are leading to declines in the qual- cent of the total number of faculty positions in ity of education provided in American engi- engineering." Furthermore, universities can no neering schools. For many years, the propor- longer depend on graduate students to relieve tion of programs in engineering and computer some of the load on regular faculty by assisting science that were unconditionally reaccredited in classroom teaching and laboratory instruc- during periodic reviews held steady at about tion. 70 percent, but in 1981 only 50 percent of the The equipment problem is equally serious. programs examined received full accredita- While faculties do their best with the resources tion.12 This sudden change indicates the gravity available, it is difficult to teach a digital design of the problems facing engineering education laboratory with equipment from the analog era. in the United States. And laboratories, as well as classrooms, have The most common and most serious causes become overcrowded as undergraduate enroll- of declining educational quality are faculty ments have climbed. Quality suffers when stu- shortages and obsolete laboratory equipment. dents have less contact with faculty, as well as less exposure to up-to-date laboratory equip- 'See note to table 67. In 1982, 1,167 of 2,887 engineering ment and computing facilities. Many univer- Ph. D.s went to foreign nationals; both industry and universities have b come heavily dependent on foreign-born engineers, es- "As of the fall of 1982. a survey of U.S. engineering schools pecially at the doctoral level. Figures on graduates reflect earlier reported 1.400authorized and budgeted faculty positions vacant. enrollments; currently, nearly 50 percent of Ph. D. candidates of a total of about 18,000. The number should be regarded as in U.S. engineering schools are foreign nationals. a lower bound because few universities have increased the num- ""Adequacy of U.S. Engineering Education." Emerging Issues ber of authorized faculty positions at rates commensurate with in Science and Technology, 1981 (Washington. D.C.: National growth in undergraduate enrollments. The most severe problems Science Foundation. June 1982), p. 60. Programs with deficien- are in computer specialties. See J. W. Geils, "The Faculty Short- cies may be reexamined after a shorter than normal interval or age: The 1982 Survey." Engineering Education, October 1983. placed on probation. p. 47. 311 308 International Competitiveness in Electronics cities, hurt by past slumps in engineering en- limited their enrollments, they have been able rollir.,.mts, are reluctant to put scarce funds into to raise Ale average quality of incoming stu- ,expansion to meet what may be a transient de- dents while keeping expansion to manageable mand. More fundamentally, universities have rates. Engineering departments at schools like had great difficulty in adjusting to shifting stu- Stanford or MIT have also been able to retain dent choices at a time when total enrollments their faculties. One of the dangers implicit in have stopped rising. Tight budgets have caused responses by industry or the Federal Govern- programs in the sciences as well as engineer- ment to the problems afflicting engineering ing to fall behind the times." education is that resources may flow dispro- The well-publicized situations at large, State- portionately to the top-ranked, research-ori- supported schools such as Iowa State Univer- ented universities. Of the nearly 300 colleges sity and the University of Illinois, typical of the and universities that offer engineering in the institutions that form the core of the U.S. sys- United States, it is the middle tierboth public tem of engineering education, are representa- and privatethat turns c,-ut the vast majority tive.15 Iowa State simply ran out of facilities to of graduates and faces the most serious prob- handle enrollment increases in computer engi- lems. neering, despite operatiny on a 6-day schedule. Because of overcrowding, students were Supply and Demand warned that they might not be able to complete Even though enrollments are still climbing, their programs in 4 years. Transfer students and the number of B.S. graduates in engineer- at Illinois must have a grade-point average of ing has been going up at about 10 percent per 4.2 on a scale of 5 to enter engineering, while year (fig. 54), it is not at all certain that the the universitywide requirement is only 3.25. number of engineering graduates in the United Shortages of facilities and teaching faculty States will meet future needs. As discussed in forced 16 of 30 large American engineering more detail later in the chapter, there will al- schools to adopt some form of restriction on most certainly be entry-level shortages at some the number of students they admit." times in some specialtiese.g., computer engi- neeringand the shortfall in Ph. as for teach- Only the elite universities have been largely ing is bound to continue; according to one esti- spared such problems, and even these have had mate, there is a current shortage of 3,500 doc- trouble attracting enough good graduate stu- toral-level engineers in industry beyond that of dents. But because the best schools have always Ph. D.s for university faculties.17 ""Science and Engineering Education for the 1980's and Be- yond," op. cit., pp. 68-69. Courses in physics and chemistry also While the rapid rise in engine ering enroll- depend on laboratory and computer facilities. For a discussion ments has led to fears by some that the United of laboratory equipment shortages with the emphasis on research States might be headed for an oversupply by needs, see "Obsolesrmnce of Scientific Instrumentation in Re- search Universities," Emerging Issues in Science and Technol- the 1990's, such concerns seem overstated if ogy, 1981, op. cit.. p. 49. only because many graduates of engineering The nine State-supported engineering schools in Texas have programs move on to other fields. Competent reported equipment needs totaling $88 million, about 70 percent of this for undergraduate teaching laboratories. The situation engineers have virtually always been employ- in Texas is probably fairly typical; an extrapolation to Ole United able in the United States, regardless of econom- States as a whole risults in an estimate of about $1 billion for ic conditions. Nevertheless, the American labor new laboratory equipment in engineering alone. See "$1 Billion for IrMructional Equipment," Engineering Education News, force contains nearly 11/2 million engineers, June 1982, p. 1. and some portions of the engineering commu ""Engineering Education Under Stress," Science,1Sept. 25. nity deny the reality of the current "shortage," 1981. p. 1479; C. Phillips, "Universities in U.S. Losing Ground in Computer Education," Wall Street Jan. 14, claiming that what industry really wants is a 1983. p. 1. large pool of entry-level people to help keep ""Universities Limiting Engineering Enrollments," Engineer- ing Education News, March 1981. The limitation's are based simply on numhers; as at Illinois, qualified student. are being turned ""National Engineering Action Conference." EngMe.?ring away. Education News, April 197,2, p. 1. Ch. 8--Human Resources: Education, Training, Management 309

salaries of mitkareer engineers low. There is suits. Microprocessor applications courses a good deal of truth to this. Entry-level short- may require equipment that schools cannot af- ages arise in part because employers prefer to ford. Few universities have adequateresources IliNe new engineers with fresh skills at lower for computer-aided design in any of the fields pay. This is an easier and perhaps cheaper way of engineering. At the same time, such diffi- of meeting their needs than coupling the culties can be viewed as similar to those that experience of midcareer engineersmany of have always existed. It has never beeneasy to whom find themselves with increasingly ob- give students a sense of the development effort solescent skillswith well-designed continuing that goes into an airplane ora nuclear power- education programs. plant. In this sense, the adaptations required Regardless, at least some specialtiesseem by the emergence of large-scale integrated cir- bound to face continuing shortages by almost .cuits or cheap computing power are nothing any criterion. These 'specialties include a new. number that are particularly relevant for the future competitiveness of the U.S. electronics Industry Initiatives industry; most notably, entry-level computer To help meet the needs of their members, two professionals are expected to be in high de- of the trade associations in electronics have mand well into the 1990's. Programs of instruc- established programs to support engineering tion in computer science and engineering still education. The American ElectrOnics Associa- tend to be small and underdeveloped. Someare tion has asked for money to augment faculty in engineering schoolsoften within electrical salaries and establish chairs in electrical engi- engineering departmentsothers in schools of neering, as well as to expand fellowship pro- arts and sciences, where computer science may grams for students; the Semiconductor Indus- be associated with mathematics departments. try Association is funding research, thuspro- Many teaching departments lack the critical viding indirect support to both students and mass that would help them thrive, not surpris- faculty members through stipends and salaries, ing in a field which did not exist 25 years ago. as well as money for equipment." In computer science, the United States graduates only 250 Ph. D.s each year, a number A different approach has been taken by Which has been decliningone reason comput- Wang Laboratories, which manufactures mini- er science and engineering faculties are suffer- computers and office automation equipment. ing greater proportional shortages of faculty The Wang Institute, located near Boston, of- than (other) engineering departments." At fers a master's degree in software engineering present, qualified software engineers are in through its School of Information Technology. short supply; although people with many types Initially endowed by An Wang, the company's of training can fill jobs as applications pro- founder, the Institute is now an independent, gramers, there are far fewer candidates for jobs nonprofit organization. With a curriculum de- in the design and development of computer- signed to give training both in the technical based systems themselves. aspects of computer software and in planning, management, and human relations, the school Other new and/or specialized fields suffer which graduated its first students in :1982 similar problems. Perhaps half-a-dozen Ameri- grew directly out of the inability of companies can universities have the facilities needed to like Wang to meet their personnel needs. Tui- design and fabricate large-scale integrated cir- tion is about $8,000 per year, less than half the actual cost of the program; the difference is "Seventeen percent of faculty positions in computer science and engineering were vacant at the beginning of 1982, versus about 9J:cent for engineering as a whole. See "Universities ""AEA, SIA to Vie in Fund-Raising Efforts," Electronic News. in U.S. Are Losing Ground in Computer Education," op. cit.; Nov. 16. 1981, p. 36. Companies in other industries have bbgun "The Faculty Shortage: The 1982 Survey," op. cit. parallel efforts.

313 310 International Competitiveness in Electronics covered by endowments and cont1'ibutions.20 ics requirements, the focus on engineering sci- Because of the emphasis on job-related skills, ence came at the expense of engineering de- candidates for admission must have at least 2 sign, as well as manufacturing and production years of professional experience, in - addition processes. This was also a time when the to a B.S. degree in an appropriate field. The spread of digital computers made numerical Wang Institute is one of a number of experi- solutions to many previously intractable prob- ments presently underway in nontraditional lems a reality, further strengthening the move- training in specialized technical fields. ment toward analysis at the expense of synthesis. In recent yearthere has been something University-Industry Relations of a swing back. Despite these and other examples of new and Industry has always wanted graduates who close relationships between business and edu- can go to work immediately, while acknowl cational institutionsfor instance, the indus- edging tha virtues of theoretical preparation in try-supported Center for Integrated Systems at the engineering sciences as preparation for ad- Stanforduniversity-industry relations, in the vancement and for continuing education. De- United States as in most countries, tend to be mand for the "old-fashioned," practically ori- uneasy. Tensions between the theoretical learn- ented engineer has led to a proliferation of cur- ings of faculty members and the more practical ricula in what isusually called engineering concerns of private firms, particularly smaller technology. companies that do not engage in much research, are common. This also holds for pro- Engineering Technology fessions such as businos administration; to Technology Programssome 2 years in some extent it applies to the sciences as well. length and leading to an associate degree, In engineering, these tensions have deep his- others full 4-year B.S. courses of studyrepre- torical roots; by the first decade of the 20th cen- sent an attempt by American colleges and uni- tury, the academic perspective had largely won versities to equip entry-level employees with out over the shopfloor orientation that many immediately applicable job skills. Graduates of in industry had advocated.21 Later, between the these programsmore than 26,000 at the asso- wars, U.S. engineering education began to stag- ciate and bachelor's levels in 1981 (40 percent nate. During World War II, numerous R&D ef- of the total in engineering)can be thought of fortsincluding many in electronicswere as paraengineers; they get less extensive and spearheaded by scientists (particularly physi- less rigorous training in mathematics, the sci- cists) with engineers filling subordinate roles. ences, and in engineering science, but consid- This lesson was one of several pointing to the erable exposure to routine technical prob- need for reevaluations of engineering educa- lems.22 While B.S. technology graduates are bet- tion. "in 1976, 16,685 associate degrees and 5.721 B.S. degrees in The resulting turn toward theory led to "engi- engineering technology were granted in the United States: in neering science" as the core of undergraduate 1982, the figures were 17.198 for associate degrees, and 8.325 at the bachelor's level. The 106 data are from "Engineering and curricula. In the post-Sputnik period beginning Technology Degrees, 1976," the 1982 from "Engineering and in the late 1950's, engineering schools empha- Technology Degrees, 1982," op. cit. sized quantitative, analytical skills even more. Well over 200 schools have technology programs, about the same number as for engineering. of the associate degrees Accompanied by a strengthening of mathemat- are awarded by community college and vocational-technical schools. In a university, B.S. programs in technology and engi- toInformatiqn supplied by Wang Institute. See also M. A. neering may be offered by different colleges, particularly where Bengs. "A Unique Institution for High-Tech Training." Boston the technology curriculum has grown out of an industrial arts Globe, Mar. 2, 1980. and "Institutionalizing the Students of Soft- setting: alternatively, both programs may be found in a "Col- ware." Computer Design, December 1982. p. 189. lege of Engineering and Technology." Many faculty members Calvert, The Mechanical Engineer in America. 1830 -1910: in engineering have resisted the movement mc'vard technology Professional Cultures in Conflict (Baltimore. Md.: Johns Hopkins education. feeling that it detracts from the proiBssion and threat- University Press. 1967). ens their own image. Outside the community of educators, agood Ch. 8Human Resources: Education, Training, Management 311 ter prepared for advancement and for creative Some offer up-to-date training in needed spe- work than technicians, their upward mobility cialties, while others are accused of turning out is considerably less than for engineers. In at people for jobs that have already disappeared. least one sense, the problems that have hit engineering education are more serious still for Public 2-year and community colleges face technology: the practical, hands-on experience chronic problems in funding their programs that these programs seek to provide depends and retaining faculty.23 Even in Silicon Valley heavily on equipment similar to that actually (the area near San Francisco where so many used in industry. electronics firms have located)which is now getting a great deal of attention as a model for Community Colleges and Local Initiatives industrial developmentthese institutions have never been well- integrated into the local In recent years, community colleges have ex- environment, and seem isolated from industry panded more rapidly than any other segment as well as from the mainstream 3f university - of higher education. Many offer engineering oriented education.24 Despite she concentration technology, as well as preengineering pro-, of E.ciectronics companies, the six community grams that send students on to universities. colleges in the area have faced severe shortages Moreover, - ommunity or junior colleges and of equipment for student laboratories, and a vocational-technical schools train many of the relationship with industry in which each group technicians who take jobs in U.S. industries seems generally supportive of the other but in like electronics. While the number of students which the various parties do not always man- earning associate degrees in technical fields age to communicate or cooperate effectively. has grown in the last decade, there is little information on the quality of these programs. Only a few States or localities have thus far attempted to meet needs for technology-based deal of confusion persists concerning the role and functilea vF .,"Science and Engineering Education for the 1980's and Be- technology educationnot surprising when associate programs yond," op. cit.. p. 93. graduate men and women trained essentially as techni.zians, "E. L. Useem. "Education and High Technology Industry; The While a B.S.-level technologist is much closer to an engineer. Case of Silicon Valley," Institute for the Interdisciplinary Study Differences in academic standards among technology pro- of Education. Northeastern University, Boston, Mass., Septem- grams may be even greater than in engineering. In contrast to ber 1931, pp. 12-18. Useern's study, based on more than 100 inter- countries like West Germany, where all I:echnical universities views, finds t'iat neither educators nor companies are respond- are held to similar standards, ';uality in American engineering ing very well to the region's educational needsparticularly at. and technology programs var,r;s widely, even among those that the secondary level. are fully accredited.

. . I f t's LC.r

_...... _1 1._._ 1 7." LZ.1.5 a so 00 I 5. 331 tc..tto III 1I4 E-02 Photo credit: © Ted Spiegel, 1983 Photo credit: © Ted Spiegel, 1983 Results of test shown In adjacent photograph. Many Computer-assisted test to determine design data engineering schools have had difficulty In acquiring on metal fatigue modern equipment of this type 315 312 International Competitiveness in Electronics

education and training outside their college or courses and related noncredit programs vary university systems. North Carolina has estab- widely in quality and rigor. Some hold to high lished a Microelectronics Center linking sev- standards, others offer little more than can be eral universities and industry at Research Tri- gleaned from trade magazines. angle Park near Raleigh; the State has also set Although many firms offer on-the-job train- up the North Carolina School of Science and. ing and continuing education for their engi -. Mathematics for high school students with un- neers, scientists, and technicians, it is impossi- usual ability. But this is more the exception ble to generalize concerning the extent and ef- than the rule, and budgetary constraints in fectiveness of such efforts. Some companies many localities may limit further development allow employees to spend several hours per of nontraditional alternatives. week taking college courses on company time; others will pay tuition provided the student Continuing Education gets a good grade. Some develop in-house pro- Ongoing education and training for engi- grams. Others refuse any assistance, relying en- neers and other technical workersincluding tirely on individual initiative. Some organiza- those without degreesis a vast and amor- tions have a companywide policy covering con- phous activity. Colleges and universities enroll tinuing education; in other cases, decisions are large numbers of students in graduate or con- left with lower level supervisors. In many com- tinuing education programs, some of whom panies, internal training programs are intended take only a Env courses while others actively primarily for new employees; in other in- pursue degrees. Such programs, many of stances, firms organize or support programs which cater to part-time and evening students, aimed at a broader slice of their work force. face problems paralleling those of undergradu- While extensive in-house training is most com- ate engineering and science curricula. The low mon in large corporations, smaller electronics visibility and lack of prestige of continuing firms have also been turning to such efforts to education aggravates the difficulties; faculty help meet their manpower needs. turnover is high, quality uncertain. For exam- Beyond case-by-case insights, the overall di- ple, enrollments in New York University's mensions of company-run training programs adult education courses in computer program- in the United States are largely unexplored; the ing have been increasing by 20 to 25 percent American Society for Training and Develop- per year, but budget limitations havemade it ment estimates that business and industry alio, difficult to purchase needed equipment, as well cate some $30 billion to $40 billion a year to as to find and keep competent faculty.25 This education and training, but little information situation is replicated in private and public col- is available on how such moneys are spent, and leges and universities throughout the country. by whom, or just what is counted in arriving Many professional societies are active in con- at the total.26 tinuing education, principally through short Looking more narrowly at engineers and sci- coursessometimes offered in conjunction entists perhaps 10 to 15 percent of those with with universitieson topics of interest to their at least a bachelor's degree are taking further members. Current favorites in electronics include microprocessor-based systems, and pro- "Information from the American Society for Training and De- graming in newer computer languages like Pas- velopment. Also "Addition to the Record: Statement of Anthony cal. Colleges and universities offer their own P. Carnevale, for the American Society for Trairtingnd Develop- short courses, as do private, profit-seek' ng en- ment," Projected Changes in the Economy, Population, Labor Market, and Work Force, and Their Implications for Economic terprises. Some companies have operated Development Policy, hearings, Subcommittee on Economic De- educational armse.g.,, RCA Institute. Short velopment, Committee on Public Works and Transportation, House of Representatives, Nov. 18-19, 1981, p. 233. Carnevale notes that 35 percent of firms surveyed by the Conference Board z5C. Anders, "Colleges Faltering in Effort to Ease Critical Short- offer remedial programs in reading, writing, and arithmetic for age of Programmera," Wall Street Journal, Aug. 24, 1981, p. 15. their employees. Ch. 8Human Resources: Education, Training, Management 313 academic coursework at any given time.27 The Japanese electronics firms. In contrast, smaller great majority are probably recent graduates, U.S. semiconductor manufacturers may have .pith many of these pursuing advanced degrees annual personnel turnovers of 35 percent or on a part-time basis. more; such companies are understandably re- In electronics, a number of larger U.S. semi- luctant to devote resources to training men and conductor firmsfor example, Intel, which has women who may then jump to competing also been a leader in the Semiconductor Indus- firms. High turnover rates in electronics hold try Association's research cooperativehave for grey-collar technicians as well as engi- instituted programs of on-the-job training for neers.31 personnelit a variety of levels. Almost all In contrast, organizations like Bell Labora- semiconductor manufacturers evidently pro- tories can finance continuing education for em- vide some training, but the intensity and length ployeeseven send them back to school full of the programs vary from company to com- timewith less fear that people will quit once pany. Continuing education for electronics and they have a new M.S. or Ph. D. in hand. In part, computer engineers in Silicon Valley is readily this is simply because many engineers and sci- available for students who can meet the en- entists view Bell Labs as an exciting and pres- trance requirements of schools like Stanford tigious place to work; for some, it is more a goal or the University of California at Berkeley; than a stepping stone. Bell supports part-time Stanford has also pioneered interactive televi- graduate study at local universities, as well as sion links enabling it to offer in-plant courses full-time, on-campus programs at a number of throughout the area. However, local electronics leading engineering schools; the company also companieswith a few notable exceptions like offers a variety of in-house training activities, Hewlett-Packardhave had little involvement plus a tuition reimbursement plan for employ- with secondary-level public education in Sili- ees pursuing undergraduate or vocational con Valley, and most of the interactions with training.32 universities seem to center on the elite institu- While it is thus true that a considerable num- tions where faculty consulting strengthens ties ber of American engineers and scientists elect with industry.28 to take additional academic coursework, many Semiconductor firms are not alone in offer- individuals make such commitments independ- ing in- housetraining to their employees. Some ently, often with little or no encouragement electronics manuf,, cturerc report that, during from their employers. Smaller companies, and their initial year on the job, new employees some large ones, often feel that they cannot af- spend up to 7 hours per week in formal course - ford to support such effortsi.e., that the pay- work.29 Hughes Aircraft's training program for back would be insufficient. As the NSF/DOE new graduates in electrical engineering is espe- study notes:33 cially comprehensive.36 Bell Laboratories and At present, continuing education is a frac- IBM are also well-known for continuing educa- tionated, uncoordinated set of operations in tion and training; low employee turnover at which academia, industry, professional soci- large companies like Hughes or IBM favors in- eties, and individual entrepreneurs pursue vestment in human resources just as it does in their own individual paths in response to what they perceive as their individual needs. There "The figure was 13 percent for 1978. A somewhat larger number took non-credit courses of various kinds. See "Battelle Study has been virtually no Federal support for con- Shows 80 Percent of Organizations Support Continuing Educa- tion for Scientists, Engineers," Information From Battelle, Mar. "See R. W. Comerford, "Automation Promises To Lighten the 5, 1980. FieldService Load," Electronics, Apr. 7, 1982, p. 110. These turn- "Useem, op. cit., pp. 9-12. over rates are good evidence that personnel shortages have been ""Employment in the U.S. Electronics Industry, Volume I," real and acute. prepared for OTA by Sterling Hobe Corp. under contract No. ""Educational Opportunities at Bell Labs," Education Center, 033-1210, p. 234. Bell Laboratories, Holmdel, N.J., 1981. "See R. Connolly, "Companies Still Short of EEs," Electronics, ""Science end Engineering Education for the 1980's and Be- May 19, 1982, p. 105. yond," op. cit., p. 96.

99-111 0 ='83 - 21 31? 314 international Competitiveness in Electronics

tinuing education, in part because the cysts of engineering and science, and their quality, at industrial programs have been regarded as a time when the United States already finds business expenses. that it does not haVe enough skilled profes- Because retraining for midcareer engineers sionals to staff its commercial industries or and skilled workers will be increasingly impor- meet its military needs. tant in the years ahead, particularly in view of the aging of the American labor force, the Fed- University and Continuing eral Government may need to reconsider its in- Education In Japan volvement. If Japanese secondary school students study The Government already plays an important mathematics and the sciences more extensively role as direct employer. In 1978, nearly 90,000 than their counterparts in the United States, engineers worked for the Federal Government, at the university level Japan's educational sys- about 6 percent of the country's total engineer- tem is inferior. Postsecondary education ex- ing labor force. Many are employed by the De- panded rapidly over the postwar period; many partments of Defense and Energy, and the Na- private colleges were founded, some with low tional Aeronautics and Space Administration. standards. While the small group of elite uni- The Air Force alone faces a shortage of over versities provides more rigorous training, they 1,000 engineers; if the Nation's defense budget have faced the same criticisms as Japanese sec- is to increase during the 1980's as planned, de- ondary schoolsexcessive reliance on rote mand for engineersboth within the Govern- learning and the acquisition of facts, rather ment and among defense contractorswill than more general skills in analysis and synthe- swell even further. Some have argued that engi- sis. In neither science and mathematics, nor neering manpower shortages could jeopardize in engineering, does the quality of university- the country's security.34 As one response, the level education in Japan match that in the Defense Communications Agency is planning United States. a National Science Center for Communications and Electronics, intended to help cope with the Engineering shortfall in the defense community. Funded Electrical engineering students in Japan with the aid of corporate contributions, the spend many hours in the classroom and labora- Center, to open in 1983, will develop educa- tory, and take a series of courses rather like that tion and training courses for participating sec- of Americans, but Japanese companies contin- ondary schools and universities.35 ue to find graduates unprepared to go to work, Over the past several decades, nearly half of while faculty members point to major weaknesses in curricula.35 Programs in engineering all U.S. engineering students have received and computer science leave little room for un- financial assistanof /7-.1e sort or another from structured or independent learning. Electives the Federal Government. Funds for laboratory are limited. Students tend to work in groups; equipment intended for teaching and research, according to critics, this fosters conformity at as well as for curriculum development, have the expense of creativity and individual initia- come from Washingtonfor science and en- tive. The education that Japanese college stu- gineering, principally through the National dents receive outside their technical fields, Science Foundation. Tight Federal budgets for moreover, is less demanding than here. Defi- education may have the unfortunate conse- ciencies in higher education are among the rea- quence of shrinking the pool of grduates in sons that Japanese companies place great stress "" "Testimony of Gen. R. T. Marsh, Commander of Air Force Systems Command," Engineering Manpower Concerns, hearing, Committee on Science and Technology, House of Representa- "S. Tubbs, "Electrical Engineering at Kyoto University," Engi- tives, Oct. 6, 1981, p. 11. Perhaps one-quarter of the country's neering Education, May 1982, p. 812; T. Sugano, "Preparation engineers work in defense-related fields. of_New_Electronics Professionals in Japan: Note for Presenta- ""The NSCCE: A New National Program," Electronics, Dec. tions Given at the Japan Society Meetings of May 1, 1981, in 29, 1981. Palo Alto, Calif., and of May 4, 1981, in New York." Ch. 8Human Resources: Education, Training, Management 315

on internal trainingthey must, simply to bring United States, but not in Japan, where business new employees up to a satisfactory level of schools are virtually nonexistent and lawyers competence. form a miniscule part of the labor force. Far fewer engineering students in Japan go on to the graduate leveleither M.S. or Ph. D. Continuing Education and Training than in the United States. Table69illustrates the stress on undergraduate training and the Despite the self-criticism that Japanese level comparatively small numbers in graduate at their institutions of higle.er education, the school. In 1980, undergs, ,iduate enrollments in performance of the country's engineers and Japanese engineering programs were almost scientists across many fields, along with the the same as those in the United States, but the demonstrated competitiveness of Japanese cor- number of graduate students was about four porations in high-technology industries like times less.37 Those students who do choose to electronics, demonstrates that the system, tak- attend graduate school find thatas in under- en as a whole, functions well. Indeed, some of graduate programscourse work and research the self-criticism appears to be no more than are less rigorous than in American universities. a mechanism for urging people and organizations to greater efforts. As the figures in table 69 suggest, while academic competition is keen at secondary levels, Deficiencies in universities are at least par- with Japanese students vying for places in the tially offset by, informal mechanisms for self- most prestigious universities, postgraduate education, as well as company-run training training brings few rewards. Because corpora- programs. Western observers repeatedly note tions in Japan rely heavily on in-house training that men and women in Japan are voracious to impart job-related skills, and because re- readers with a strong penchant for self-study. search does not have the prestige that it car- The average Japanese not only spends more ries in the United States or Europe, Japanese time reading than the average American, but engineers have little incentive to go on to grad- more of what he reads is job-related. The spread uate school. Patterns are similar in other pro- of quality control methodologies through fessions. Graduate work in business or law is Japanese industry, outlined in chapter 6, de-. a popular road to career advancement in the pended heavily on self-study through books, magazines, and radio and TV broadcasting. "American engineering schools enrolled 72.600 M.S. and Ph. D. students in 1980. about 40 percent on a part-time basis The national broadcasting company, NHK, "Engineering Enrollments, Fall 1982," op. cit. Although only transmits nearly a hundred educational pro- 337.800 undergraduates were enrolled in Japanese engineering grams to attentive audiences each week, in- schools compared to nearly 400.000 in the United States, the retention rate is much higher in Japan. Once admitted, Japanese cludingthe popular "science classroom" students face far fewer hurdles than Americans, and a higher series. percentage graduate. Japan's Government also provides free train- Table 69.Enrollments in Japanese ing for recent high school graduates, as well Colleges and Universitites as for workers who need improved skills before they can join or rejoin the labor force. Over the Number of students, 1980 years, courses taken by those in the second cat- Junior Technical college college UniversityM.A./M.S.Ph. D. egoryadults already in the job market--have Engineering ..20,100 46,300 337,800 14,900 2.400 expanded greatly. They fill many of the same Physical functions for smaller companies as the ink.., science 54,600 3,740 2,590 All other house training programs conducted-by large programs ...346,100 1,349,100 17,160 13,210 corporations. Data collected by the Ministry of Total 366,200 46,300 1,741,500 35,800 18,200 Labor indicate that more than 200,000 trainees SOURCE: "Kagaku Gijutsu Benran" (Indicators of Science and Technology), Kagaku Gilutsucho Keikakukyoku (Science and Technology Agency, were enrolled in publicly supported vocational Planning Bureau). 1981, pp. 100.103. programs in1977,although the content of these

313 316 International Competitiveness in Electronics

programs has been criticized for not keeping For example, corporations rely on in-house up with the needs of industry.3° training to help buila sense of loyalty to the group and to the ofg-mization." The widely Company-Run Training Programs remarked cooperative spirit of Japanese employees is no accident. Internal training and continuing education comprise an integral part of organization and Well-known features of Japanese organiza- management in larger Japanese companies. tional structures such as quality control circles This is perhaps the most fundamental differ- also serve a training function, one in which the ence between the Japanese and American ap- informal elementsand the stress on inter- proaches to technical education. While many group cooperationare at least as important American corporations engage in such activ- as any knowledge imparted. In an unusually ities, Japanese programs are much more com- comprehensive program in a Japanese automo- prehensive. Developed in part to compensate bile plant, engineers teach other employees in for deficiencies in formal education, training a "workshop university."40 After completing an has evolved to complement employment pat- extensive program of after-hours study-2 terns in which many employees spend their en- years or more, with no special remuneration tire careers within a single organization. the workshop university graduate is rewarded with a certificate from his section chief. The Of course, not all Japanese firms or workers aim is not only to improve individual skills, but fit this pattern. Table 70 shows that big com- to keep employees intimately involved in day- panies provide much more training than small. to-day matters that affect productivity and One reason is that managers are generally ro- manufacturing efficiencyranging from work- tated within large organizations, a practice place organization, job flows, and task descrip- often accompanied by study programs. More tions to interpersonal relations. important, long-term employment within a single firmsometimes called "lifetime" employ- Among the most systematic of the industrial mentis the rule only in the major corpora- training programs in Japan have been those'de- tions (and then only among male employees). veloped by leading manufacturers of electronics and electrical equipment. Since the 1920's, While training programs within Japanese firms such as Mitsubishi Electric and Matsu- companies generally impart specialized skills shita have been known for recruiting promis- e.g., computer programingthey serve other ing young employees directly from high school, purposes as well, purposes that may seem and giving them extensive and formalized in- paternalistic or coercive to Western observers. house training.'" Such programs emerged in response to shortages of qualified wt. .7'ers in "H. Shimada, "The Japanese Employment System," Japan In- stitute of Labor, Industrial Relations Series, Tokyo. 1980, p. 21. the ,3ftermath of World War I. Japa:.,...,overn- mer0- fostered universal primary education, but

Table 70.Distribution by Size of Japanese Firms Providing InHouse Training 3°For a detailed analysis of training within a Japanese bank, see T, P. Rohlen, For Harmony and Strength: Japanese White- Size of company by Proportion of companies with Collar Organization in Anthropological Perspective (Berkeley, number of employees training programs (as of 1974) Calif.: University of California Press, 1974). In the bank studied by Rohlen, some of the training programs emphasized technical 1,000 or more employees .. 95.1% skills while others were directed at "character building." Both 500.999 , 85.3 varieties were designed to help integrate workers into the cor- 300-499 75.9 porate community. Over the course of a year, about one-third 100.299 58.8 of the staff went through one or more programs at the bank's _30-99 26.3 own training institute. **-5-29 10.1 For a comprehensive treatment of training practices at Toyota, 41.3% All firms. see R. E. Cole, Work Mobility, and Participation (Berkeley, Calif.: SOURCE: H. Shimada, The Japanese 'Employment System, Japan Institute of University of California Press, 1979). Labor, Industrial Relationsseries,Tokyo, 1990. Based on data from Ministry of Labor, Jlgyonai Kyolkukunren Jisshi Jokyo Chosa {Survey **Work Mobility, and Participation, op. cit., pp. 183-184. of Intra-FirmVocationalTraining and Education), 1974. "Levine and Kawada, op. cit., p. 267. Ch.8Human Resources: Education, Training, Manag&n:ont 317

during that period gave little attention to sec- International Differences in ondary schooling. Vocational training was !eft Education and Training to the private sector, where companies designed their own programs to train the workers A principal conclusion from the preceding needed for expansion and industrialization. If sections is that, while American universities the government had pursued a more compre- continue to provide an excellent education for hensive manpower policy, including the sup- this country's engineers and scientistsas wit- port of secondary and vocational education, nessed by the large numbers of foreign gradu- Japanese firms probably would not have moved ate students who come here to studythe aver- so far in this direction. age American high school or college graduate is poorly prepared to function in a technologi- Initially, then, internal training was a direct cally based society. Compared to their counter- response to shortages of skilled labor, and ef- parts in a number of other advanced industrial fort was directed at blue- and grey-collar work- nations, American students get less training in ers rather than managers or engineers. Despite mathematics and science, and even if they vast improvements in Japanese secondary ed- study these subjects, learn virtually nothing ucation since the 1920's, most large companies retainindeed have continued to develop about technology. these programs. Many operate their own edu- Deficiencies in mathematics are particularly cational institutes; Hitachi, for instance, main- serious. Mathematics acts as a filter at the en- tains two, sending graduates of technical high trance to many careers. Although the impor- schools for year-long courses of study.42 The tance of mathematics to the practice of engi- company, which is not untypical, also offers neering is sometimes exaggerated, a high level a large number of specialized training courses of competence relative to the average is needed on an ad hoc basis. Hitachi has given more to complete a degree program. A student who than 1,000 over the past two decades (some does not master algebra and trigonometry in many times); they include foreign languages high school drops immediately into the class and topics in management, with specialized of those needing remedial work; he or she will subjects such as international business avail- not be admitted directly into a university pro- able for executives.43 Students in a typical gram in engineering or science. Those with de- course spend 30 hours in the classroom and ficiencies who try to catch up often fail. Part twice that on outside assignments; the average of the problem is simply that as many as one- skilled worker or technical professional at fourth of high school teaching posts in mathe- Hitachi takes two such courses a year. matics are currently vacant, and a comparable fraction are filled by individuals only temporar It is difficult to compare the direct costs of ily certified to teach, many of them marginally such activities with the corresponding benefits qualified at best.45 Industry has hired away to the firm. But even in large organizations many high school mathematics teachers at at- with extensive training programs, such as Toy- . tractive salaries, in part to fill vacancies for ota, expenditures reportedly total less than 1 computer programers and systems analysts. percent of salaries and wages.t, The returns- -tangible and intangibleappear substantial. The American educational system also does a poor job of preparing those who do not go to college. Even among high school graduates, "R. Dore, British FactoryJapanese Factory (Berkeley, Calif.: functional illiteracy is common (estimates for University of California Press, 1973), p. 651. the population as a whole range around 20 per- "M. A. Maguire, "Personnel in the Electronics Industry: United States and Japan," prepared for OTA ,inder contract No. cent). Vocational education and training vary 033-1360, pp. 54-55 On training for managers in Japan, see T. Amaya, "Human Resource Development in Industry," Japan In- "'IA Science Dean iJescribes Teaching as in Sorry State," New stitute of Labor, Industrialized Relations Series, Tokyo, 1963, York Times, Mar. 6,1982, p. Cl. Shortages of teachers in science pp. 21-24. as well as mathematics appear to be worsening; see Science and "Work Mobility, and Participation, op. cit., p. 185. EngineeringEducation: Data and Information, op. cit., p. 7. 32i 318 International Competitiveness inElectronics widely in quality; excellent programs and in- The demand for grey-collar technical em- adequate ones can be found virtually side-by- ployees in industries like electronics is high; side." Other countries have developed more one study has estimated a growth rate in the coordinated and comprehensive approaches to United States of nearly 18 percent per year, vocational training, with benefits both to indi- faster than the projected growth in demand for vidual workers and to industry." engineers." But in pointed contrast to coun- Skilled technical workers are a vital resource tries like Japan or West Germany, the Ameri- for the U.S. electronics industry, and deficien- can educational system has not responded in cies throughout the middle levels of the Ameri- any large-scale fashion to these needs. In Ger- can labor force could constrain the future many, fully 60 percent of the labor force has growth and development of semiconductor and specialized training in grey-collar technical computer firms, as well as cowponies in other skills, while in the United States the figure may high-technology fields. Technicians, designers be as low as 10 percent." and draftsmen, and field service personnel A scarcity of adequately trained technical must be literate, have basic quantitative and workers could be just as serious a problem for technical skills, and, ideally, understand some- American industries like electronics as con- thing of the logic of the systems they work with. straints on capital investment or a stagnating Without such abilities, they cannot use ad- overall economy. Labor mobility has tradition- vanced production and R&D equipment to ally been a mechanism for opening manpower greatest effect, nor exercise sound judgment bottlenecks; indeed, the U.S. electronics indus- in the technical problems they face on a day- try already depends heavily on foreign-born to-day basis. Individuals without these skills though U.S.-educatedengineers. The next have little upward mobility; an assembly line section looks more closely at the structure of worker needs at least some quantitative facility the U.S. labor market, particularly mobility. to be able to move into jobs such as machine repair, quality control and inspection, or shopfloor supervision. ""Technical Employment Projections of Professionals and Paraprofessionals, 1981-1983-1985," American Electronics Asso- ciation. May 1981; see also "Testimony of Robert P. Hender- son, Chairman and C.E.O., Itek Corp.. Lexington, Mass.," Fore- '"See, for example. G. W. Wilbur. "Va. Vocational Education casting Needs for the High Technology Industry, hearing. Sub- Seen As Hindrance to Development." Washington Post. Nov. committee on Science, Research, and Technology, Committee 29, 1982. on Science and Technology, House of Representatives, Nov. 24, "The extensive system in West Germany is described i 1"- 981 tional Training in the Federal Republic of Germany (Bri. ''S.Praia, "Vocational Qualifications of the Labour Force Commission of the European Communities. 1978). See aib.o S. in Britain and Germany," National Institute Economic Review, Hutton and P. Lawrence, German Engineers: The Anatomy of November 1981, p.47; response of R. H. Hayes. Business Man- a Profession (Oxford: Clarendon Press. 1981). pp. 94.95; and I. agement Practices and the Productivity of the American Econ- M. Geddes, "Germany Profits by Apprentice System." Wall omy, hearings. Joint Economic Committee. May 1 and 11. and Street Journal, Sept. 15. 1981. p. 33. lune 1 and 5, 1981, p. 46.

Supply and Mobility of Labor Shortages of men and women with knowl- ployment" has been a policy goal for many edge and skills at .a time of high overall unem- years, the upward trend of the unemployment ployment poirt to weaknesses in U.S. labor rate over the past decade has combined with market policies, including manpower training tion and Training Policy," Projected Changes in the Economy, and adjustment assistance." While "full em- Population, Labor Market, and. Work Force, and Their Implica- tions for Economic Development Policy, op. cit., p. 33. During 5°For a relatively comprehensive, and critical, analysis of labor 1981, perhaps I million U.S. jobs wont unfilled, while,10 million market policies in the United States, see R. I. Vaughn. "The lob people were without work. See K. Sawyer, "Learning lobs in Development Administration: A National Employment, Educa- School," Washington Post, July 28. 1982, p. 1.

322 Ch. 8Human Resources: Education, Training, Management 319 sporadic shortages of workers having specific Overall Labor Market Trends skills to create a new circumstance, one to The labor forces of Japan and the United which the Federal Government has failed to States expanded swiftly during the 1960's, respond. largely as a result of postwar baby booms. Over its history, the United States has seen Table 71 shows the rates of increase in both periodic labor shortages, for both skilled and countries to have been considerably greater unskilled workers. More recently, it has begun than in Western Europe. Japan's labor force to seem thateven if the general quality of grew from 49 million in 1966 to 56million in American education were to remain highthe 1979, while that in the United States went from labor market might simply not be able to supply 79 million to 105 million over tne same peri- the right numbers of people, in the right places, od.51 Although Japan has experienced some at the right time. There are a host of reasons labor shortages, the relative abundance of for such concerns, ranging from changing at- working-age men and women in both Japan titudes toward work, to the aging of the U.S. and the United States contributed to economic population, to local constraints such as high expansion during the postwar period. Younger housing costs.° As the work force ages, and the workers made up an especially large propor- .needs of the U.S. economy shift, retraining will tion of Japan's labor force during the 1950's. be the only way to utilize people's talents fully. During the 1970-80 period, both countries continued to experience rapid increases in their This section asks whether the development working-age populations (table 71); growth in of the U.S. electronics industry will be con- their labor forces will slow during the 1980's. strained by limited supplies of engineers and computer scientists (overall employment trends Rising employment levels in industrialized are examined in the next chapter), together economies over the past two decades have been with a related question: Are the high levels of accompanied by shifts toward tilt.; service sec- labor mobility that have characterized some tor; agricultural employment has declined, parts of the U.S. electronics industry essential with manufacturing roughly stable or declin- for continued growth and competitiveness? ing slowly (see ch. 5, fig. 32). Japan has been The comparative neglect of training and re- something of an exception, withrise in in- training in the Unit 'd States stems in part from dustrial employment couple_ Alla sharp the ease with which companies have been able drop in agriculture; both the industrial and the to hire new employees with needed skills; this service sector grew as a result of migrations in turn has reinforced tendencies for workers from the farm. Such trends will continue as in- to move from job to job in search of fresh opportunities or higher pay. The labor market in "Labour Force Statistics 1968-1979(Paris: Organization for Japan functions much differently. There, the Economic Cooperation and Development, 1981), pp. 18.19. system emphasizes long-term employment (for some) and loyalty to the firm; mobility is low. Management practices in Japan have sought Table 71.Labor Force Growth in Several Countries to compensate for the weaknesses of such a system, while taking advantage of the stability Average annual increase it brings; rather than looking for new people in labor force to revitalize faltering efforts, Japanese firms 1960-70 1970-80 United States 1.80/0 1.5% redeploy those they have. Japan 1.3 1.3 WeSt Germany 0.3 0.7 France 0.8 1.1 United Kingdom 0.2 0.3 SOURCE: 1960.70W. Gelenson and K. Odake, "The Japanese Labor Market," Asia's New Giant, H. Pc..ick and H. Rosovsky (eds.) (Washington, *ln Silicon Valley, a housing shortage has driven prices so high D.C.: Brookings Institution, 1976) P. 590. that semiconductor firms have found it difficult to hire from out- 197480World Development Report, 1980 (Washington, D.C.: The side the area; few candidates can afford to move. World Bank, August 1980), p. 147. 323 320 International Competitiveness in Electroni:s

dustrial employment in the advanced countries tions throughout the Nation. Other estimates slowly shrinks relative to services. have ranged up to 25,000.53 NSF's projections It is perhaps understandable that, during a for engineers together with scientists indicate period of rapid overall labor force expansion that `he total supply of new graduates should and continuing movement into services, the meei the demand by the end of the decade. U.S. Government paid little attention toman- However, NSF may be overestimating the ex- power policies: the economy was growing rap- tent to which scientists can function in engi- idly; periods of high unemploymentwere neering jobs; in any case, shortages are an- viewed as transient; people could take advan- ticipated even by NSF in the computer field, tage of a relatively broad range of oppor- for statisticians, and in several engineering tunities. The situation today is much different: specialties. About one third of the 1.4 million aggregate expansion has slowed; the skills job openings in science and engineering over needed by industry are more specialized;un- the 1978-90 period are expected to be computer employment has become persistent. Current related (including programers). Despite NSF's unemployment is especially troubling because relative optimism, other forecastsadmittedly it is caused in part by mismatches between the often conducted by or for industry, and thus capabilities of people looking for work and the perhaps skewed by the preference of compa- jobs available; in such circumstances, more nies to be able to pick and choose when hir- rapid, oggregate expansion may do little good, ing new employeeshave projected massive and may even be impossible if growth indus- shortages of engineers, perhaps as many as. tries cannot hire the people they need. 300,000 by 1990.54 All such forecasts should be approached with considerable skepticism. None of the methodologieswhether based on Personnel Supplies 'for the simple trend analysis, on survey techniques (as U.S. Electronics Industry for many of the engineering manpower stud- In the United States, shortages of softWare ies), or on econometric models (as used by the engineers and semiconductor designers have Bureau of Labor Statistics)has a good record been heavily publicized over the past few years. for pfojecting employment; there are too many Not only has demand been higheven through imponderables. the deep recession of 1982but warnings of While forecasts and projections can warn of longer term shortfalls have been common. One possible future shortages, insight also comes educator predicted that American schools will from current levels of unemployment in some graduate a cumulative total of 70,000 new B.S. occupations. Unemployment rates have been r'iegree-holders in electrical engineering and remarkably low in technical fields. During :omputer science over the period 1982-85, 19'63, when overall U.S. unemployment aver- while nearly 200,000 will be needed.52 As dis- aged about 7 percent, only 0.6 percent of com- cussed in the next chapter, demand forcom- puter specialists found themselves out of puter service technicians is expected to dou- work." The unemployment rate for engineers ble during the current decade, with job open- "Henderson, op. cit., p. 63. ". ings for programers and systems.analysts go- "Henderson, op. cit., p. 66; "Science and Engineering Educa- ing up almost as fast. tion for the 1980's and Beyond." op. cit pp. 48-50, 60; M. A. Harris, "Manpower Surveys Continue to Disagree," Electronics, A number of job-market surveys and esti- July 28, 1983, p. 108. NSF concludes that interfield mobility mates of aggregate demand for engineers have particularly influxes of those trained in mathematicswill mitigate but not ellminate the shortage of computer specialists. One been conducted in the recent past. The Labor potential problem is that even if the total supply of engineers Department estimated that in 1980 there were roughly meets the demand, small firms with limited resources 17,000 unfilled entry-level engineering posi- may still be unable to hire new people. Wetional Patterns of Science and Technology Resources 1982, NSi 82.319 (Washington, D.C.: National Science Foundation, ""Congress Warned of Shortages in Electrical, Computer En- Marrh 1982), p. 68. This amounts to only 2.000 people. While gineers," Electronic News, Nov. 23, 1981, p. N. The rather alarm- unemployment rates for professionals of all types are normally ist estimates were those of K. Willenbrock, Southern Methodist well below the overall unemployment rate, the 0.6 percent figure University. for computer specialists is unusually low. 32., Ch. 8Human Resources: Education, Training, Management 321

(as a group) in 1980 was less than 1 percent; in mathematics, engineering, and the physical engineering unemployment averaged 1.8 per- sciences; fewer than one-third of those work- cent over the decade of the 1970's, a period that ing as computer professionals have degrees in included the aerospace "collapse," when the computer fields." High turnover rates are part unemployment rate for engineers reached 2.9 of the same picture- as noted earlier, turnover percent." Aggregate unemployment levels dur- has been rapid among both engineers and tech- ing the decade averaged 6.2 percent, more than nicians in the U.S. electronics industry. three times as high. Regardless of uncertainties in the projec- The persistence of unemployment rates far tions, then, few people are worrying that the below the national average indicates that an United States will have too many engineers in "oversupply" of new graduates in engineering the years ahead; capable individuals with train- is unlikely. And, while mathematicians and ing in engineering comprise one of the most physical scientists, as well as engineers, may employable parts of the labor force. The pros- sometimes have trouble finding the jobs they pects of shortage are real in the sense that var- consider most desirable, men and women with ious projections differ mostly in the magni- training in such fields can move into a wide tudes of the shortfalls predicted. variety of occupations; many scientists eventually find themselves practicing engineering. In contrast to the wide public awareness of It is hard to argue that the United States could potential shortages of engineers and computer have too many graduates of science, mathe- scientists, supplies of grey-collar manpower matics, or engineering curricula.. have received remarkably little attention. Thus, it is impossible to discuss needs for techni- . Data on salaries and job offers for new engi- cians, service personnel, and other skilled neering graduates provide additional evidence workers in quantitative detail. But the situation of high demand. In 1981, engineers made up for machinists illustrates the kinds of problems only 8 percent of new college graduates, but to be expected. The Bureau of Labor Statistics received more than 65 percent of all job of- estimates that annual job openings will average fersand at starting salaries twice as high as 22,000 over the near future; meanwhile, in 1978 for those in the humanities.57 Salary offers to only 2,300 machinists completed- registered engineers and scientists rose at higher rates programs of apprentice training.80 than for other categories of graduates throughout the 1970's. Demand remained high even The Question of Mobility during the recession of 1981-82.58 Lateral mobility helps moderate sporadic-- Another indicator of personnel shortages is shortages of workers with particular sets of mobility across disciplinesthe number of peo- skills. just as clearly, individuals can only move ple who switch to fields other than those in within a limited realm; a surplus of physicists which they got their formal education. Much might help compensate for a scarcity of com- of the demand for computer specialists has, puter engineers, but few biologists Would be been filled by men and women with training able to function in such jobs.

"Science Indicators-1980,op. cit., p: 320. The peak year for 59"Science and Engineering Education for the 1980's and Be- unemployment among engineers was 1971. yond," op. cit., p. 39. This refler_ta in part the slow development 5713. Abelson, "Industrial Recruiting on Campus,"Science, of academic programs in computer science and engineering. sem, ri, 1981. p. 1445. The data comes from a survey by the 6°S. Qualtrough and J. Jablonowsld, "Filling the Need for Skilled College Placement Council covering more than 60,000 offers to Workers," American Machinist, June 1579, p. 131. But see also new recipients of bachelor's degrees, The salary data also points N. H. Rosenthal, "Shortages of Machinists: An Evaluation of out the big differences between indusWel and academic start- the Information,"Monthly Labor Review,July 1982. p. 31. Al- ing salaries. though the electronics industry employs machinists, far greater "lr. 1982, two-thirds of computer and office equipment firms numbers work in heavier manufacturing industries. Regardless surveyed by NSF reported difficulty in hiring electrical and com- of the statistics, a good deal of anecdotal evidence bears out the, pu'er engineering graduates, as opposed to 95 percent in 1981. difficulty that manufacturing firms of all types have had in find- See "EEs Still Needed, Though Shortage Has Eased, Says NSF," ing journeyman machinists, tool and die makers, and other Electronics,Jan. 13. 1983, p. 69. skilled craftsmen. 322 international Competitiveness in Electronics

American workers move within and across Labor Force Mobility in the United States technical fields further and more frequently The United States draws strength from the than their counterparts in other industrial na- mobility of its labor force, not only in mod- tions. Managers and technical professionals erating skill shortages, out as a stimulus to in- change jobs much more often in the United novation, technology diffusion, and entrepre- States than it_ Japan; mobility is greater in neurship. New firms in rapidly growing seg- Europe than in Japan, but still considerably less ments of electronicssemiconductors, com- than here.'" In the U.S. electronics industry, puter software and peripheralsare often built turnover has been high among unskilled work- around engineers and managers who leave one ers, where unions have been week, as well as company to start another. On the other hand, among those whose abilities have been in high rapid staff turnover, as pointed out above, demand. works against company-run programs of edu- The effects of labor mobility cut several cation and training. In part to counteract the ways. It is li!`le solace to a firm losing key peo- attractions of entrepreneurial ventures, a ple if they start a new enterprise that con- number of large and successful American elec- tributes to U.S. competitiveness. At the same tronics firmsincluding Hewlett-Packard, time, organizations with low rates of person- Texas Instruments, and IBMhave adopted nel turnoverin any countrymust guard personnel policies aimed at retaining their against stagnation, find ways to generate new employees. Likewise, merchant manufacturers ideas; this is one of the reasons for internal such as National Semiconductor and Intel at- training and job rotation programs in Japan. tempted to maintain staffs and avoid layoffs The pluses and minuses of high or low rates during the semiconductor sales slump of 1981- of 1 lior mobility depend on factors such as 82. In this regard, American electronics man- rates of technological change, current eco- ufacturers are quite consciously emulating nomic conditions, and corporate strategies. their Japanese competitors. Patterns of mobility across industries and Still, white-collar mobility has been a sine countries depend, among other things, on in- qua non of the more dynamic merchant semi- centives such as promotion policies and wage/ conductor firms, which have competed aggres- benefit packages; managements have consid- sively for both technical and managerial talent. erable latitude in tailoring these to enhance or Silicon Valley manufacturers have offered a discourage turnover. Government programs wide range of benefits, including extensive dealing with adjustmente.g., training and recreational facilities, to recruit white-collar retraining, unemployment assistancealso act professionals. Some have even paid bounties as incentives or disincentives. While gener- to employees who bring in new people. Pros- alizations emphasizing cultural differences are pects for rapid advancementand the lure of sometimes advanced to explain mobility pat- someday getting in on the ground floor of a terns in the United States as compared to Ja- new organizationhave helped attract mana- pan, examining incentivesand the ways in gers and engineers, as has the California set- which public policies affect themprovides a ting. The mobility of talented people has helped sounder basis for understanding. Although diffuse electronics technology, contributing to Japan's labor force tends to be less mobile than rapid commercialization of new developments that of the United States, a good deal of varia- which in turn has helped build an internation exists across industries and firms in both tionally competitive industry. countries. The lawsuits occasionally filed against ex- "'On West Germany, see German Engineers: The Anatomy of employees by firms seeking to prevent leakage a Profession, op. cit., p. 48ff. of their technology are among the more strik-

326 44'

Photo credit ,,ny electronics technicians get their o in the military illustrations of the relation bet, mobility and technology diffi 'sunsuccessful 1968 case Lives who went over to Fairc y example. In 1980, Intel sue ner employees who left to stet ied Seeq; the basis of Intel's ed in a negotiated settlement, mployees intended to base iness on trade secrets deali gn and manufacture of lark nmable ROMs (read only al action to prevent technolog

Russell, "Seeq Loses Bid for Rehearing." :5. 1982, sec. 1, p. 50. Ch. 8Human Resources: Education, Training, Management 323

an extreme case; more commonly, firms adopt positive programs of rewards and incentives to keep valuable employees. Again, the semiconductor industry has been a leaderhelped by a working environment that many employees find stimulating. Of course, features that help retain people also serve a company well in attracting new employees. Turnover has also been high among unskilled blue-collar workers in many parts of the U.S. electronics industry. In domestic semiconduc tor plants, production employees tend to be female and ethnic. According to one estimate, women make up 40 percent of the Silicon Valley work force, and are heavily concentrated in lower paying jobs; three-quarters of assemblers are women." In contrast to the' mobility of top-echelon managers aid technical professionals, turnover among unskilled production workers is associated with a lack of skills; they can be laid off during business slumps and replaced later. Mobility in Japan The stereotype of Japan's "lifetime" employ-, ment system contrasts sharply with patterns in the U.S. electronics industry. According to the popular view, the Japanese system ensures job security until retirement. Also part of the vet,1983 stereotype is a sequence of promotions based tining largely on seniority rather than merit, with employees waiting patiently to move up the pay scale, assured of their ultimate reward. These arson- aspects of the Japanese system have been Moto- viewed as integral parts of a company-as-family ;tex- model, making unions in the American or as an European style superfluous. "Enterprise un- lup of ions," organized on a company basis rather many than by trade or occupation, have been seen vhich as part-and-parcel of a socioeconomic milieu at the characterized by harmony among workers and their managersthis in turn leading to low interfirm h the mobility coupled with high employee motiva- po- tion and productivity. While pieces of this ies).82 model are visible within Japan's economy,, it )ws is ,312. Howard, "Second Class in Silicon valley," Working Pap?r,I, September-October 1081, p. 25. See also M. Chase, News, "Semiconductor Firms Get Mixed Review on Safety in Study by California Agency," Wall Street Journal. Jan. 1, 1902, p. 6.

327 324 International Cornpet:Vvaness in Electronics applies to only a minority of the labor force; although corporations in Japan attempt to ad- moreover, the stereotype obscures crucial de- just to business downturns by reducing work- tails that affect the working lives of all Japa- ing hours before laying off regular employees, nese. when recessions deepenas in 1974-75they reduce employment levels at rates quite corn- To begin with, labor relations were far from parable to those in Europe, if not the United smooth in Japan as recently as the 1950's. Fur- States. Smaller Japanese companies have sel- thermore, lifetime employment is typical only of large Japanese companies, and many of dom been reluctant to cut back their labor these encourage their employees to retire at a forces. relatively early agecommonly around 55 or Nor does the stereotyped picture of seniority- 60. Afterward, many "retirees" must find new based wages in Japanese corporations (the workwhich may turn out to be a part-time or nenko system) hold up under scrutiny. Or)? lower paying job with a subsidiary of their study finds that promotion is based on a "col.. former employerbecause retirement benefits promise" between seniority and ability, the par- are low.°' Moreover, in small firms especial- ticulars varying considerably across firms.°7 ly, but also in large enterprises, Japanese work- Smaller, more rapidly growing organizations ers do leave their jobs. Horizontal mobility tend to emphasize meritocratic promotion, i.e., movement from one firm to another with- while older, established firms remain less will- out advancementis fairly canmon among ing to single out talented individuals from younger Japanese workers, particularly those others of their age group. Age and ability are, with low skills. Women seldom have much job furthermore, weighted differently depending security or upward mobility, much less the or level, with progress in the upper ranks a many temporary employees that are another stronger function of ability. Clark concludes feature of Japan's labor market." Women are that the ambiguity built into Japanese promo- generally encouraged to resign upon mar- tion practices encourages people to do their riagecertainly at childbirthand if they best: while promotion has generally been auto- return have no seniority. The 3.4 million tem- matic after a certain period of service, there porary and day workers; men and womenac- is also the possibility that outstanding perform- counting for about 6 percent of the work ance will be rewarded with rapid advance. forceare the first to be let go in the event of And, although the nenko system may appear recession. Temporary employees provide flexi- to underpay well-trained and able young work- bility to cope with economic slumps without., ers, over the longerterm they can expect to at- laying off regular workers. The proportion of tain salary levels well above those in their age temporary employees in Japanese manufactur- bracket who have lower skills or less education; ing firms has increased markedly since the re- salary profiles for older male workers in Japan cession of the mid-1970's." Furthermore, show considerable spread. Finally, as the Japanese labor force has aged, "Japanese electronics firms, along with the rest of Japanese employment practices have begun to change. industry. have been under some pressure to extend retirement ages. In the mid-1960's, retirement in the major electronics firms With the fraction of older workers increasing, was generally compulsory at 55 to 57 for men, perhaps 50 for salary competition for the best qualified recent women. By the mid-1970's, many companies had extended these graduates will intensify; recent surveys of hir- ages by about 5 yeirs. See S. Takezawa, et el., Improvements in the Quality of Working Life in Three Japanese Industries ing suggest that, in -:ertain fields, including (Geneva: International Labour Office, 1982), pp. 66-87, 95. electronics, shortage''f younger employees "A. H. Cook and H. Hayashi, Working Women in Japan are likely. Data oomph,by the Ministry of (Ithaca, N.Y.: Cornell University Press, 1980). See also F. Ginsbourger, "Japan's Dark Side," World Press Review, July .711. Clark, The Japanese Company Haven: Yale Univer- 1981, p. 32. sity Press, 1979), pp. 115-118. On the nment of the nenko mC.-y. Lin, "Wage-Price Behavior Under External Price Shocks system, see T. Inapmi, "Labor-Manage, ' Communication and Productivity Slowdown: A U.S.-Japan Comparison," Discus- at the Workshop Level " Japan IMMitute of L. Industrial Rela- sion Paper No. 402, Economic Growth Center, Yale University, tions Series, Tokyo, 1983. InEq.,,e:ni also inclue data on pro- April 1982, p. 22a. motion patterns (pp. 10-14). 323 Ch. 8Human Resources: Education, Training, Management 325

Labor indicate that younger Japanese workers and of themselveswith enhanced competi- can choose between two or three entry-level tiveness; nor should high mobility be consid- jobs, but those aged 55 and over must win out ered more "modern" than low (or vice versa). over 5 to 10 other jobseekers to find a posi- High mobility in the United States goes with tion." As a result of such trends, wage com- other aspects of the U.S. economy, just as low pression for older employees seems likely to mobility is consistent with Japan's socioeco- intensify, retirement ages may be extended fur- nomic environment. ther, and the role of seniority in promotion Public policies influence the choices made decisions will diminish. Generational conflict by corporations among the options outlined between younger employees, for whom high above. Government support for training techni- demand will push tip salaries, and older work- cians can enlarge the talent pool. Vigorous ers who stand to lose by comparison, may fol- manpower policies, designed to support ex- low." If and when such events come to pass, panding sectors of an economy, will stimulate the features that now. make the Japanese em- interfirm and interindustry mobility. Tax ployment system seem unique will stand out writeoffs for company-run programs of educa- less. tion and training would encourage intrafirm mobility. High turnover rates have made Amer- The United States and Japan Compared ican corporations wary of investments in train- While the contrasts are often exaggerated,' ing or retraining that may pay off to their com- Japanese and American employment practices petitors. "Talent raiding"so characteristic of do lead to quite different patterns of mobility. American semiconductor firmsoften be- How do these interact with the structures of comes the alternative. the electronics industries in the two countries Employment practices in the United States to affect international competitiveness? Firms may change as a result of the demographics of in each nation have alternatives in seeking the aging, just as the aging of the Japanese lnbor people they need. One option is to hire employ- force is altering patterns in that country. As ees away from other companies, an approach the U.S. population grows older, continuing more likely to be successful in the United education for those in midcareerblue- and States. An alternative is internal recruitment grey-collar workers, as well., as white-collar iritrafirm mobilityin conjunction with re- professionalswill become a necessity. When training, an avenue particularly appropriate in the labor force was expanding rapidly, employ- a system such as Japan's, where people tend ers could count on new graduates to fill many to identify more strongly with the corporation of their needs; this is less true today, particular- than with a vexation or profession. Still another ly in light of current inadequacies in technical method of coping with shifting occupational education. American firms may find them- needs is to alter or expand the potential pool selves emulating the internal training efforts of new entrants. Despite the vitality that the of their Japanese competitors, with manage- U.S. electronics industry has drawn from em- ment practices designed to enhance a com- ployee mobility, there is no need to associate pany's human resources becoming critical either high mobility or a lack of mobilityin elements of corporate strategy. The remainder of this chapter turns to questions of manage- 6,Clark. op. cit., p. 32. ment and the organization of the workplace, 0,R. E. Cole, "Changing Labor Force Characteristics and Their askingamong other thingswhether there Impact on Japanese Industrial Relations," The Paradox of Prog- ress.E. Austin (ed.) (New Haven, Conn.: Yale University Press, really is a uniquely Japanese. approach to 1976). p. 194. management.

323 326 International Competitiveness in Electronics

Organization and Management As the end of the 19th century brought rapid economic growth and technological change to American, industry, new management techniques arose to deal with shopfloor organization. The old ways, developed during the days of hand work, proved a poor guide to factory production using mechanized equipment, particularly in the emerging mass production industries Frederick W. Taylor, founder of the scientific management school, is the best remem- . bered of those who pioneered new methods.7° Taylor began as an engineer at an ironworks, and his approach to managementincluding plant layout and job flows, and the man/ machine interface--reflects the bent for rationalization associated with his profession. While Taylor himself, and the techniques he developed and advocated around the turn of thecen- tury, showed considerable appreciation for the human element in factory work, many of his .77 y followers carried scientific management to the extreme of treating people as another variety of machine. Scientific management still bears kiz; this stigmaand "Taylorisrn" is a dirty word for many who associate it with the Chaplin of Photo credit: Wash Modern Times. Integration of programmable robots into the -factdry environment poses a new set of probler Taylor believed that,for every task in for manufacturing industries manufacturing, there was an optimum method that could be "scientifically" discovered. By reducing each job to its essential elements The idea of a scientific means for org, employing, for instance, the techniques of what ing factory work attracted American busi: has come to be known as time-and-motion men. New machine tools, the assembly studythe workplace was to be rationalized mass production of durable goods like bic3 and productivity maximized. Although Taylor home appliances, and automobiles, prese thought that this approach should also increase a rapid succession of new problems; ir cooperation among workers, one of the chief trialists eagerly embraced Taylorism criticisms Ol scientific management has always means of dealing with them. The manage] been its rather mechanistic conception of the science movement springing from Tay individual, leading to an emphasis on simple, early work has continued to thrive an repetitive tasks. spread internationally; ft still shapes curri and textbooks in American business scj

70F.W. Taylor, The Principles of Scientific Management (New and industrial engineering programs. York: Norton & Co.. 1911). N. P. Mouzelis, Organization and, Bureaucracy: An Analysis of Modern Theories(Chicago:Aldine The human relations approach to mar Publishing Co., 1967), gives .a useful historical overview of var- ment was developed .primarily by induE ious approaches to organizational management. 'psychologists, beginning a decade or two Ch. 8Human Resources: Education, Training, Management 327

In contrast to the engineers who espoused Tay- ades, and now graduate more than 50,000 men lorisin, the human relations school stressed and women each year. In contrast to Japan and peoples' attitudes and motivation as keys to Western Europe, where top managers tend to productivity and manufacturing efficiency. move up from the ranksand a few individuals Studies in the human relations vein explored still reach high levels having started on the the workplace as a social organization and the shop floorAmerican firms, especially the individual employee as a member of the group; larger, publicly held corporations, have tended practitioners saw their goal as fostering an ami- to bring new employees directly into manage- cable working environment, one built around ment-track jobs. Typically graduates of aca- the existing shopflbor culture. While advocates demic programs in business administration, of scientific management tended to be anti- some of them fill staff positions, others move union, the human relations school accepted quickly into middle management. Thus the _unions as an integral part of the social system management professionwith its extensive of the factory. network of specialized academic programs has become a principal vehicle for transmitting Just as the reductionist tendencies of scien- and validating the techniques used in Ameri- tific management have been criticized, so the can business. human relations approach has been faulted for its stress on harmony to the neglect of the real Management training in this country pre- conflicts of interest characteristic of work life, pares people for work in hierarchical organiza- and for overemphasizing small group behavior tions. Distinctions between those who plan and while failing to deal with the organization as those who do the work are more sharply drawn a whole. in American corporations than elsewhere; this divisionand the equally sharp distinctions be- Variants of these two attitudes toward man- tween those responsible for production, or agementwhich reflect contrasting theories of "operations," and the rest of management organizationcontinue to proliferate. The two has increasingly come under scrutiny and criti- are based on fundamentally different notions cism." In contrast, Japanese and European of what makes organizationswhether factory, business practices are rooted in on-the-job ex- store, or officefunction, and hence on meth- perience and company-run training programs. ods for improving their operation. At present, Management institutes exist, but are typically the human relations approach has become oriented toward the needs of midcareer ex- identified with the popular view of Japanese ecutives seeking fresh perspectives. management, but both schools have American origins. This is not, of course, to say that While the ideal types of "American" and Americans cannot learn from foreign experi- "Japanese" management are exaggerations that ence. Organizations in other countries have fail to capture the variety existing within the adapted management practices originating in two countries, they nonetheless point to differ the United States to their own needs, and it ing conceptions of the nature of modern orga- may be time for a reverse flow into American nizations. The Japanese model is based on corporations. authority stemming from tradition and social- ization; the American approach is less personal and more legalistic. Central to the Japanese Organizational Types and model are group decisionmaking, cooperation Management Styles between labor and management, and long-term tenure in an organization viewed as analogous The Manager as Professional In the United States, management is a disci- "See ch. 6. The following pair of articles in the July-August 1981 issue of the Harvard BusinessRevieware typical examples pline with its own graduate schools and ad- of this criticism: R. H. Hayes, "Why Japanese Factories Work," vanced degrees; M.B.A. programs increased by p. 56, and S. C. Wheelwright, "JapanWhere Operations Real- an order of magnitude over the past twodec- ly Are Strategic," p. 67.

331 328 International Competitiveness in Electronics

to a family. Ideally, these result in a well- Group decisionmaking as embodied in the integrated system, with human resources as the traditional Japanese approach is a good fit with firm's most important long-term asset." corporate organizations that .offer individual employees considerable security and involve Decisionmaking in Japan them with the company outside their immediate duties and working hours. Company Symbolic of the Japanese approach is the housing and recreationalfacilities, group ringi seido (approval system), through which outings and even vacations, along with inter- middle-level personnel obtain sanction and approval from the top echelons by circulating a nal training programs, can all be viewed as in- document to which each person affixes his seal centives for building loyalty among a fairly im- or signature." The process yields systeniatic mobile labor force. In, best light, the system is but slow "bottom-up" decisionmaking. A deci- "wholistic" in orientation; in worst light, it is, sion is final once the company president adds a sophisticated brand of industrial paternal- his seal; since many individuals participate, ism." The widespread acceptance of company there is considerable communicationif not rather than craft or tide unions and the 'com- always true consensusand, once the outcome paratively few days lost to strikes in Japan has become apparent, little uncertainty. Con- (table 72) indicate that this labor-management trasts are frequently drawn between the systemoriented toward consultatiod'and con- tendency of "individualistic" Americans to formityhas worked to the benefit of the cor-. continue pushing their own views, even after porations that have designed and implemented contrary decisions by upper management, and it. As table 72 illustrates, large numbers of the Japanese casewhere, as the saying goes, workers participate in strikes even in Japan, "when the train leaves the station, everyone is but little time is lost because work stoppages on board." The point is that whatever disagree- are short, often serving functions that are at ments precede the ringi decision, they are sup- least partially symbolic. posedly buried afterwards, the policy fully supported by all. Contrasts With the United States While authority in a Japanese company, is Extensive involvement with the company vested in 'the president, employees at many outside normal working hours and groUp deci- levels participate in the consensus-building sionrriaking diverge markedly from patterns in process. Not all of them have the precise and the United States, whererather than spread- well-defined responsibilities that characterize, ing responsibility for decisions through the job descriptions in an American corporation. organizationtop management is expected to Ambiguity attaches to organizational structure "As late as 1976, more than one-quarter of Hitachi's male em- in Japan, rather than to people as in the United ployaes still lived in company housing; the figure had been nearly States. The Japanese system does not involve 40 percent in 1967. See Improvements in the Quality of Work- --much bargaining among managers'anagers and subor- ing Life in Three Japanese. Industries, op. cit., p. 69. dinates, nor is it participative in the sense often used in the West. In contrast to U.S. practice, where management decisions and business , planning get detailed attention, the ringi system Table 72.Work Stoppages Due to Labor Disputes allows people throughout the firm to agree on "49 Several Countries (1978) / Total number of Total number of generalities, with the specifics to be worked out participants in employee work-days later. work stoppages lost United States 1,600,000 39,000.000 "N. Hatvany and V. Pucik, "Japanese Management and Pro- Japan 660,000 1,360,000 ductivity," Organizational Dynamics, spring 1981, p. 8. West Germany 490,000. 4,280,000 "For a detailed description of the ringi seido, see M. Y. France 1,920,000 2,200,000 Yoshino, lapLn's Managerial System: Tradition and Innovation United Kingdom . 1,040,000 9,400,000 (Cambridge, Mass.: MIT Press, 1968), pp. 254ff SOURCE: "Japan, An International Comparison 1990," Ketzai Koho Center, p.49.

332 Ch.8Human Resources: Education, Training, Management 329 provide leadership. Corporate cultures in the styles as clear-cut as the stereotypes suggest; United States give pride of place to strong- in both countries, firms have identities that willed executives who leave their mark on an may vary from division to division as well as organization.1. Power, 'status, and privilege also- changing over time. The wholistic orientation attach to Japanese executives, but with real dif- of the Japanese style carries strong paternalis- ferences. The ability of managers to show im- tic overtones, with discrimination against mediate resultsprofitability over the next women and minority groups a fundamental quarter being the current target of criticsis part of the system." And, although Japanese central to the American model, group effort to management is sometimes viewed as people- the Japanese. Well-defined and often narrow oriented, personal interactions are marked by responsibilities, centralized authority, rigidly pervasive if subtle status distinctions. Paternal- hierarchical organization chartsplus the pos- ism does lead to job security for some fraction sibility of swift promotion and high rewards of the labor force in Japan, security which is characterize the "results-oriented" manage- less common in the United Sta't'es. The Ameri- ment styles of American firms. Ambiguity is can approach, while often assumed to maxi- viewed as undesirable; expertise cultivated; mize opportunity, does so in part by encourag- men and women enter the firm as specialists ing competitionsome would say to excess in accounting or finance, marketing or stra- among individuals seeking advancement and tegic planning. Individualism istolerated, personal gain. but within well-defined boundswitness the "white-shirt syndrome" still hanging over com- Comparisons of American and Japanese panies like IBM." The comparatively high management often focus on particular tech- levels of personnel mobility in the United niquese.g., quality control circles in Japan, States, and the tradition of adversarial relations management by objectives in the United between unions and management, are part of Statesrather than the schools of thought, such this picture. as scientific management or human relations,. from which these techniques derive. Burim- A further difference between Japanese and provements in management seldom result from American management practicesdiscussed the isolated adoption of some technique. This in more detail in chapter 6is the emphasis quotation from a Japanese engineer points out companies in Japa, place on manufacturing the difference between technique and under- and its integration \-ith the rest of the firm. lying attitude:" Toyota's much-noted system of just-in-time One difference I find hard to explain to my (kanban) production and inventory control is Western colleagues is that we do exactly the only one example. Since Japanese managers same things that the industrial engineer does tend to rise relatively slowly through the ranks, in Detroit or Pittsburgh; but it means some- with periodic lateral moves, stress on manufac- thing different. The American industrial engi- turing is perhaps natural. In contrast to the neer lays out the work for the worker. Our in- situation in the United States, where produc- dustrial engineers are teachers rather than tionmore especially quality controlhas lit- masters. We try to teach how one improves tle prestige, is even viewed as a dead-end job, one's own productivity and the process. What a number of Japan's top corporate executives we set up is the foundation; the edifice the began their careers as quality control or worker builds. Scientific management, time manufacturing engineers. and motion studies, materials flowwe do all Both Japanese end American approaches to 7°Even the more bemused commentators on the Japanese mod- management have their strengths dnd weak- el. such as Ouchl, note its racism and sexism. See W. G. Ouchi, nesses. Few corporations exhibit management Theory Z: How American Business Can Meet the Japanese Chal- lenge (Reading, Mass.: Addison-Wesley, 1981). p. 91. 77Quoted in P. F. Drucker, "What We Can Learn From .Japanese " "Life at IBMRules and Discipline, Goals and Praise Shape Management," Harvard Business Review. March-April 1971. p. IBMers' Taut World." Wall Street Journal.Apr. 8. 1982. p. 1. 117.

333 99 -111 83 - 22 330 International Competitiveness in Electronics

that, and no differently from the way you do Participative Mechanisms it in the States. But you in the States think that this is the end of the job; we here in Japan In the United States, industrial democracy believe it is the beginning. The worker's job has been associated with collective bargaining begins when we have finished engineering the by labor unionsan interpretation of worker job itself. participation neither so encompassing as in It is too easy to write off such statements as Western Europe nor quite so narrow as in empty philosophizing. Japan. While American unions have continued / to bargain over wage-benefit packages, Euro- pean workers have succeeded in extending WorkerParticipation their influence over workplace and organiza- The past decade has seen continuing intorest tional decisions. In some contrast, quality con- in industrial democracy, more so in Western trol (QC) circles were developed by managers Europe than in the United States," Stemming in Japan as tools for improving labor produc- at least in part from persistent economic prob- tivity and product quality. Most of the interest lems, some companies and some governments in QC circles among Americans has alsonrig- have experimented with methods for increas- inated with management. If American working the involvement of the labor forcepar- ers, particularly in companies with strong ticularly blue-collar employeesin decision- unions, have sometimes been reluctant to em- making and work design. One aim has been to brace QC circles, quality-of-work-life programs moderate wage demands. This section outlines have found, a better reception with labor. several of the modes of employee participation The worker participation movement in West- that have evolved in Europe, as well as the ern Europe is based on two presumptions: first, _quality control circles originating in Japan. The that labor is just as important to production as purpose is to capture some of the variety of capital; second, that blue-collar employees have foreign approaches, and ask how such mech- the right to be represented in corporate deci- anisms might help the productivity and com- sionmaking. Participatory mechanisms include petitiveness of American industry. A large work groups at the shopfloor level, work coun- number of specialized techniques for redesign cils at the plant or enterprise level, collective df the working environment and employee in- bargaining, labor representation on boards of volvement have been developed, both here and directors, employee-owned enterprises, and overseas; no attempt has been made to describe worker representation on socioeconomic ad- any except quality circles, which are covered visory bodies to governments. Beyond these because they have attracted so much atten- direct involvements, publicly owned compa- tion." nies are a longstanding fixture on the European scene, with governments paying more or less "Much of. the material.on European countries in this section attention to their management depending on is based on A. L. Ahmuty, "Worker Participation in Manage- political pressures and economic conditions. ment Decision-Making in Western Europe: Implications for the United States." Congressional Research Service Report No. At the shopfloor level,: work-life programs.-- 79-136E, Apr. 23, 1979. See also P, C. Roberts, H. Okamoto, and give employees a voice in determining how in- G. C. Lodge, "Collective Bargaining and Employee Participa- tion in Western Europe, North America and Japan," Tim Tri- dividual tasks should be performed, with the_ lateral Commission. 1979. aim of increasing job satisfaction as well as im- "For an overview of a number of these, sce R. M. Kanter, proving productivity. Employee involvement "Dilemmas of Participation: Issues in Implementing Par- ticipatory Quality-of-Work-Life Programs," National Forum, in work methods can be viewed as a reaction spring 1982, p. 16. Several case studies can be-found in J. A. against the scientific management tradition, in Fadem. "Automation and Work Design in the United States, which an experttypically an industrial engi- Working Paper Series No. 43. Center for Quality of Working Life, Institute of Industrial Relations, University of California, Los neerhas full responsibility for task design. Angeles. 1982. Sometimes, work-life programs reduce produc-

33,1 Ch. 8Human Resources: Education, Training) Management 331 tivity (as traditionally measured), a sacrifice benign. West German workers have remained that firms like the automobile manufacturer more interested in power over matters such as Volvo appear to have accepted in the interests hiring and firing practices. of employee satisfaction. (Volvo replaced a number of assembly lines with batch assembly Blue-collar employees in the United \States operations, giving workers more variety.) have restricted their attempts to influencecom- pany policies and decisions to the traditional At the enterprise or plant level in Europe, concerns of labor-management relations.\Un- work councilsindependent of unionsgive ion officials have been ambivalent about mov- employees a voice in codetermining a firm's ing beyond questions of wages, benefits, and future. Labor representatives\on these councils working conditionsprobably fot fear of losing participate in financial and other business deci- some of the bargaining power that comes with sions, although at the head of the agenda tend an adversarial stance.81,In contrast to Western to be matters like personnel policy, health and Europe, participation by'American workers on safety, and shopfloor organizational practices. boards of directors has been ralemostly American-owned companies in West Germany brought on by circumstances such as Chrys- have seldom been comfortable with codeter- ler's recent financial plight. Although the many mination; in the United States, the few labor- plant closings in industries like steel have led management committees that have been estab- to proposals that employees purchase facilities lished tend to have a much narrower fous, and scheduled to be shut down, few such plans to be viewed primarily as vehicles for enhanced have gone forward. communication. One of the best knowi is the National Committee to Improve the Qiility of While collective bargaining is virtually uni- WOrk Life, established by the United\ Auto versal in advanced market economies, there are Workers and General Motors in 1973. Current many differences of form and substance. In economic conditions may motivate more uch Japan, about 95 percent of all unions are orga- experiments. nized on an enterprise basis.82 In addition to ._ collective bargaining between unions and man- One of the most far-reaching experimen s in agement, negotiations take place each spring employee participation has been instituted in between groups of, firms and unions. The West Germany. In the early 1970's, the Min- "spring Offensive" is most visible in the steel; istry of Research and Technology, in coopera- electrical machinery, shipbuilding, heavy mation with the Ministry of Labour and Socialf- chinery, and automobile industries, as well as fairs, began a program aimed at the "huarrr.\ public corporatio (where a special mediation ization of work." Based on the Work Counci s committee decidGs on the settlement). Wage Act passed by the German parliament in 197 decisions during the spring offensive help set the premise is that government should not only patterns for smaller firms. Still, compared with safeguard employee health and safety, but un the. United States or many European nations, dertake to improve opportunities for individual labor in Japan has little real power. development and participation in decisionmaking.8° In general, the response of workers to Quality Control Circles, these initiatives seems to have been less positive than for earlier programs of codetermina- QC circles have been heavily publicized as tion, particularly in industries like electron- mechanisms for worker participation. Quali- ics where the workplace is already relatively ty circles are relatively autonomous, composed "For an evaluation of labor-management committees in the °°"Research on the Humanization of Work," Action Pro- United States, see K. Frieden, "Workplace Democracy and Pro- gramme of the Federal Minister for Labour and Social Affairs ductivity," National Center for Economic Alternatives, Wash- and the Federal Minister for Research and Technology," Doc. ington, D.C., 1980, p. 31. No. 2181/74e. See alsoProgramm Forschung zur Humanisierung "' "Labor Unions and Labor-Management Relations," Japan In- des Arbeitslebens,Der Bund,.sminister fur Forschung and stitute of Labour, Japanese Industrial Relations Series, Tokyo, Technologie, 1979. 1979. 335 332 Intornational Competitiveness in Electronics of a small group of workersperhaps a dozen But even in Japan, QC circles are sometimes typically led by a foreman or senior employee." perceived as a coercive management tool. In Japan, financial incentives play a relatively Overenthusiastic accounts of quality control minor role, without the emphasis on prizes for circles in Japan sometimes give the impression suggestions or improved performance that of a panacea; in reality, Japanese firms vary some American firms have adopted. QC meet- widely in the extent to which they utilize QC ings in Japanese companies are often held out- circlesregardless of commitment to circle ac- side normal working hours, and workers may tivities, they are only one manage ent tech- not be paid for their time. Although the circles nique among many. now work on job-related problems beyond quality control per see.g., production meth- Over a hundred American firmsincluding General Motors, Ford, and Genera Electric ods, worker trainingthey grew out of the have experimented with QC circ es, but the postwar stress on quality inspired by Ameri- question of whether or not they v ill work as cans such as Deming (ch. 6). The contribution well in the United States as in Japan has not made by Japan's business leaders was the ex- been answered. Certainly there are obstacles pansion of quality control to involve participa- here that do not exist in the typical Japanese tion by virtually everyone in the firm. Employ- organization. In the U.S. context; for example, ee training via circles, for example, is intended monetary incentives may, be essential; the to reduce the need for specialists in quality as- Lockheed program is typical in that employees surance and production engineering. As dis- are not expected to meet after hours, or without cussed in chapter 6, the quality and reliability extra pay." Experience also shows that Ameri- of electronic products depends on factors rang- can middle managers must be persuaded to ac- ing from engineering design to relationships cept and support the QC approach, else they with suppliers; while the quality of many Jap- may perceive the circles as challenges or as im- anese goods is now excellent, iL would be a mis- plicit criticisms of past performance. take to attribute this to any one technique such as the QC circle. Unionized firms add another dimension. Cole notes that even in Japan enthusiasm Where QC circles have been introduced into within a QC group tends to wane, and circles American companies without the consultation need to be periodically revitalized. It would be and support of union leaders, they have not no surprise to find a Hawthorne effect at work been successful. Organized labor remains am- in many of the success stories involving QC bivalent; AFL-CIO spokesmen have felt that QC circles (i.e., a situation in which any of a wide circles could be a tool for breaking up unions, variety of changes in the workplace environ- and the evolving attitude appears negative.85 ment would improve employee motivation and Japanese firms with plants in the United. productivity, at least temporarily). The effec- States have generally introduced circles grad- tiveness of QC circles also depends on the ua'Icr avid with considerable care, if at all. tent of employee identification with the com- Quasar, owned by Matsushita since 1974, did pany; members participate more fully if they not install its first circles until 1982; the com- feel that. their work is recognized and ap- pany plans to have 25 in operation by the end preciated within the organization. A group- of 1983.88 QC circles in Japan function in a con- oriented Japanese corporation is more likely to foster such attitudes than many of the Amer- ican firms now experimenting with QC circles. 64"Quality Cc-ntrol Circles Save Lockheed Nearly $3 Million in Two Years," Quality, May 1977, p. 14. "Cole has carried out the most systematic studies on quality °sR. S. Greenberger, "Quality Circles Grow, Stirring Union circles. The discussion below is based largely on Work Mobili- Worries," Wall Street Journal, Sept. 22, 1981, p. 29. ty, and Participation, op. cit.. pp. 135ff. Also see R. E. Cole, "Will "Information from Quasar. Thus far, the company views its QC Circles Work in the U.S.?" Quality Progress, July 1980, p. experience in the United States with QC circles as successful, 30: Improvements in the Quality of Working Life in Three but perhaps. not so successful as in Japan. For examples of other Japanese Industries, op. cit., pp. 76ff; and Inagami, op. cit., pp. experiences in electronics, see J. D. Couger, "Circular Solutions," 31-34. Datamation, January 1983, p. 135. Ch. 8Human Resources: Education, Training, Management 333 text that includes enterprise unions, a relatively parent, there is only so much they can import immobile work force, and seniority-based wage and implement. increases. Not all the elements in the Japanese approach or in QC circles themselves are like- The one respect in which Japanese-owned ly to prove attractive to workers and managers firms in both the United States and Japan stand in the United States. out is their emphasis on employee motivation and participation, and on diffusion of responsibility through the ranks. The survey results Japanese and American Management indicate that the anecdotal evidence on Japa- Styles: How Much Difference? nese concern for employee motivation reflects Do Japanese firms operating in the United a genuine distinction: in terms of the models Of States exhibit a distinctive management style? management style outlined earlier, the Japa- Or in adapting to the new setting do they act nese approach is closer to the human relations more like American firms? Keeping in mind pole. At the same time, the range in behavior the structural differences that have been out- across both Japanese and American firms is lined,, how different are management .styles wide. even within japan from those in the United. Japanese-owned firms stress communication States? By comparing a foreign subsidiary both and personal interaction both horizontally and to its parent and to local competitors, variables vertically. At least some, aspects of consensual of ownership and geography can be separated. decisionrnaking have been transported to the This section presents the conclusions of a study United States. One can question the extent to of managerial differences among U.S.' and which Japanese managers accept and act on Japanese firms. The survey sample included the information received through these com- upper and middle managers ',Pm: 1) Japanese munication channels, as opposed to using them subsidiaries of American companies, 2) Japa- to manipulate opinion and impose top manage- nese-owned subsidiaries in the United States, ment decisions. Nonetheless, in employee sur- and 3) both American and Japanese firms in veys, managers in Japanese-owned firms both their home country. Appendix 8A, at the end here and in Japan were more often described of this chapter, explores the data on national as sensitive to others and accessible to subor- differences in management style more system- dinates than managers of American.owned atically.87 companies. This in itself contributes to employ- The survey results show that American- and ee motivation and satisfaction. Japanese-owned electronics firms do not di- Such behavior patterns can be associated verge greatly in management style. In many with the human relations school of manage- respects, managerial practice§ were more ment. The principal contrast with American- closely associated with geographical location owned firms is along the in formal dimensions than with ownership; i.e., Japanese-owned of organizational behavior; there was little dif- firms in the United States acted more like ference between the U.S. and Japanese firms American firms, U.S. subsidiaries in Japan surveyed in terms of organizational hierarchy more like Japanese companies. In itself, this or formal lines of communication. The distin- should be no surprise, given that foreign sub- guishing features of Japanese management ap- sidiaries everywhere are mostly staffed by local pear to be rather intangible, matters of attitude people. Even if upper managers come from the more than method. U.S. subsidiaries of Japanese companies have e'App. 8A. together with the summary here, is based on a report prepared for OTA by M. A. Maguire. It includes an independ- generally found this emphasis on human reia- ent analysis of data from a project directed by R. T. Pascale. tions and employee participation to work well: The subset dealing with electronics has been of primary interest Typically, the firms surveyed have modified to OTA. For a discussion based on dl the data, including other industries, see R. T. Pascale and A. G. Athos,The Art of Japanese management techniques imported from Japan Management(New York: Simon & Shuster, 1981). to fit the American context without abandon-

337 334 International Competitiveness in Electronics in3 the human relations thrust. Furthermore, (he precise degree to which human relations- some of 'the best performing American firms priented management contributes to the per- display a similar concern for employee par- formance of particular companies, it does ap- ticipation, with the, implicit goal of giving in- peal to be a common trait in well-managed and dividual workers a stake in the suctess of the competitive organizations in both countries. enterprise. While it is impossible to cIL;iermine

Summary and Conclusions

Commitment to the development and utiliza- ciencies in mathematics are most serious; these tion of human resources is closely associated disqualify people at an early age from a broad with corporate success, and, through this, with range of career opportunities, depriving the industrial competitiveness. In electronics, U.S. Nation of a vast potential resource. manufacturers have had difficulty filling critical positions in engineering; a concurrent For those qualified for admittance, programs shortage of skilled technicians, while not so in engineering,athematics, and the sciences well publicized, could prove as serious a bot- offered by Ame ican colleges and universities tleneck. At present, the United States seems in both underg aduate and graduateremain danger of falling behind other countries at unsurpassed. Nonetheless, they have slipped training people in the skills needed for high- relatively; di,A2',.aering schools, in particular, technology industries like electronics; deficien- are suffering :k1:-om a lack of qualified faculty cies exist in both public and private sectors. and from inadequate and obsolete equipment, Education, provided first and foremost by the The press-es of expanding undergraduate en- public schools, determines the skills and to a deterioration in the capabilities that people bring with them to thp qr°ilLiTtillit; Jducation, Continued low enroll- work force. The abilit'to continue learning ments of Ph. D. students mean that the short- on the job as well as offalso depends on the age of engineering teachers will continue; what quality of that formal edocation. While some might have been a transient problem is rapid- American firms provide or encourage contin- ly turning into a serious long-term concern, uing education and training for their employees, others do little or nothing. Moreover, the average American worker is Inadequacies in the education and training less prepared than his or her counterpart in a of the American labor force are growing more number of other countries for productive em- serious. Beginning at secondary levels, the ployment in industries like electronics. As a preparation of Americans in science and result, the United States is heading toward mathematics is simply not on a par with other more shortages c` skilled blue- and grey-collar industrialized nationse.g., Japan. A smaller workerstechnicians, designers and drafts- fraction of U.S. college students major in men, engineering aides, field service person- technical fields. While many American univer- nel. Lik9wise, many white-collar jobs are filled sities are, at the moment, limited in their ability by people with little understanding of mathe- to handle greater numbers of engineering stu- matics, science, or technologyand with lit- dents, a more fundamental problem is the rel- prepakation for comprehending technical atively small fraction of the college-age popula- subjects even on a lay basis. Meanwhile, un- tion qualified to enter such programs. The employment in the United States has been ris- typical U.S. high school graduate is not only ingin part the result of a mismatch between poorly prepared in mathematics and science, what people are able to do and what needs to but uninformed concerning technology. Deli- be done.

330, Ch. 3 Human Resources: Education, Training, Management 335

One way private firms can compensate for spective employeesto which end a number deficiencies in formal education is to establish of the more successful electronics companies, in-house training and retraining programs; in in the United States as well as Japan, have addition to such efforts, many Americar, firms developed management systems that empha- support continuing education outside the com- size employee participation. Giving individuals pany. The incentives for such efforts, however, a voice in decisions that affect them increases are lower here than 'Western Fir., .pe or motivation and commitment to the organiza- because of .liouldly of the U. labor tion. r ire. The freqc nncy with which Americ ns take new jobs heightens the risk that the com- Despite the vogue for Japanese management pany will lose its investment. Nonetheless, a techniques, the human relations approach is number of U.S. electronics companies have de- in no way unique to Japan or to Japanese cor- veloped ambitious employee training efforts, porations; the similarities among competitive and the semiconductor industry is developing firms in Japan and th'?tilted States are more programs in conjunction with universities that striking than the differences. Specific mech- will help to educate new people, as well as sup- aniSms, such as quality control circles or labor- porting the R&D base. Despite their promise, management committees, appear of secondary such initiativta will l by thein.,,01,,,,5' he suf- importance compared to less tangible signs of l-icient to nic,ut. the skill requirements of the attentiveness by management to the attitudes t.lectronics industry in the years ahead, much and talents of employees. less the broader moods of the U.S. economy. While many U.S. corporations have developed their own brands of human resources-ori- Coved,;nit :. Unit Mate, and tocalhas traditionally carried the ented management, others could profit by more major responsibility for education-and train- attention to worker participation; Americdn ing; expanded public sector programs for train- managers seem to be gradually realizing that ing and retraining appear necessary for build- they may be underutilizing their employees. ing the competitiveness of American industry. Table 73 shows that executives of U.S. firms As demographic forces tilt the labor force rank employee participation as the most impor- toward greater proportions of older workers, tants'ingle influence on productivity. Whether retraining will be essential if the talents of mid- they act on such beliefs is another matter; but, career employees are to be effectively utilized. of the forces that affect competitiveness, man- As U.S. industry continues to advance techno- agement is the most immediately amenable to logically, workers who find themselves...dis- change by individual companies. A renewed placed by structural.change will'be dependent commitment by American companies to the de- On retraining to find productiVe employment velopment and utilization of human resources elsewhere. As job opportunities shrink for could pay large dividends in international com- those with limited skills, men and women with petition. poorer educations, and without the developed. ability to learn on the job, will more and more Table 73.Rankings by American Managers of find themselves unemployable. Given the com: Factors Contributing to Productivity petition and mobility characteristic of the American economy, the private sector cannot Factor Average ran ka reasonably be expected to provide the needed Employee participatton programs 3.61 Better communications 4.11 training and retraining; only government bod- Better labor-management relations 4.45 iesat all levelscan take on this responsi- Increased training 4.46 bility., Quality improvement , 4.81 Increased automation 5.02 The efforts of private industry begin with the Productivity incentive programs 5.13 Cost reduction programs 6.01 people available in the labor pool. In large Increased R&D 6.28 measure, the art of management lies in max- a8ased on a scale of .1 to 10, with 1 being the most effective and 10 being the least. imizing the contributions of existing and pro- SOURCE: Mechanical Engineering, September 1981. 336 lntemational Competitiveness in Electronics

Appendix 8A.Japanese and American Management Styles: A Comparison' Survey Results A primary objective was to gather information on A survey covering managers and oth .1 employees communications and decisionmaking styles. Sur- in four electronics companies provides the basis for vey questions were designed to indicate whether this comparison: American firms differed from Japanese in the ex- companyAl, an American consumer elec- tent to which decisionmaking and communications tronics firm operating only in the United could be described as hierarchical and formal (the States; hypothetical U.S. model) rather than informal and companyJ, a Japanese consumer electronics cooperative (the hypothesis for Japan). firm with operations both in the United States The results show that all the American-owned (J-A) and in Japan (J-J); and firmsA1, A2-J, and A3-Jrelied more heavily on companies /12-J and A31, the Japanese sub- written communications, both here and abroad. sidiaries of two American firms, one a manu- More surprisingly, firm J-Athe U.S. subsidiary of facturer of computers, the other of semicon- a Japanese companywas in many respects more ductors (not necessarily in that order). "Japanese" in decisionmaking and communica- All the firms were high performers in their respec- tions than the parent organization (J-J); the data tive portions of the electronics industry. show a greater proportion of upward communica- The data can be grouped in several ways. For in- tion and a lower proportion downward in the stance, a geographE; grouping gives: first, the two United States than in the same firm's home offices. organizations in the United Statesone American- Overall, however, the survey resultstable 8A -1 otvned (A1), and one Japanese-owned (J-A); and, showed much less variation in patterns of commtl- second, the three operations in Japanone Japa- nication among these firms than the pure Japanese nese-owned (J-J) and two American-owned (A2-J, and American models would predict. Additional A3-J). Alternatively, grouping the sample by own- survey questions indicated that the subsidiary A2-J ership yields a set of three American-owned firms is more "dependent" on its American parent, as (Al, A2-J, A3-J) which can be compared with the measured by written communications with head- Japanese-owned organizations (J-A and J-L. For quarters, than the subsidiary J-A was on its Japa- most purposes, the ownership distinction is more nese parent. - illuminating, probably because top managers who The survey results also shed light on hierarchy set the tone of an organization generally came from and formalization in the organizational structure the parent firm. In contrast, most of the middle- of each company in terms of the size /level ratio: the level managers had been recruited locally; thus in total number of employees in the organization di- organization J-A they were largely Americans. vided by the number of hierarchical levels. The The survey covered both middle and upper man- lower the size/level ratio, the more formal and hier- agers, utilizing interviews as well as written archical is the firm's structure. Again, the results responses. Nonmanagerial employees were also may seem somewhat surprising: the Japanese com- sampled via questionnaires to gather data on job pany was the most hierarchical, with its domestic satisfaction. The data must be interpreted with cau- and U.S. operations scoring the same-133 (J-J) and tion because of the small number of organizations. 134 (J-A). One of the American electronics firms At the same time, the survey results for electronics measured 150 in its Japanese organization (A3-J), come from a much larger body of data covering 10 little different from the Japanese-owned company. industries; differences across industries were few. The other two American-owned organizations had ratios of 284 (A2-J) and 533 (A1). In other words, 'This appendix is b Ised on "Personnel in the Electronics In- none of the American firms are particularly for- dustry: United States and Japan," prepared for OTA by M. A. malistic or hierarchical on this measure (which can Maguire under contract No. 033-1360. The report includes an be rather sensitive to differences in the overall size independent analysie of data collected for a project directed by R. T. Pascale. Pascale's own treatment. including discussion of of the companies compared). Another indicator, the Companies in other industries. can be found in R. T. Pascale extent to which they make use of written job de- and A. G. Athos. The Art of Japanese Management (New York: scriptions, found the American-owned companies Simon & Shuster, 1981). ranked higher in formalization. Ch. 8Human Resources: Education, Training, Management 337

Table 8A1.Responses of Middle and Upper Managers to Questions Dealing With Communications and Declsionmaking Styles

Companywide averages Al JA JJ A2.1 A3J Questions dealing with manager's own behavior: Number of telephone calls and facetoface contacts per day 81 69 72 24 55 Number of written communications per day 10 4 3 8 7 Hours in meetings per day 2 2.5 3 2.3 3 Percentage of calls to those higher in the organization 21%25% 23% 14%36% Percentage of calls to those lower in the organization 40%31% 31%56%37% Percentage of meetings with those higher in the organization 13 %'16% 41/4 8% 100/a Percentage of meetings with those lower in the organization 64%;i6% 84%88%80% Questions dealing with manager's evaluations of their supervisors' declsionmaking styles: Percentage of decisions Supervisor makes alone 36% 21%29%23%25% Percentage of decisions supervisor makes after factual Input from subord'nates 25%30%20%40%25% Percentage of decisions supervisor makes with participation by subor tes 43%49% 51%37%50% SOURCE: M. A. Maguire, "Personnel In the Electronics Industry: United States and. Japan," prepared for OTA under contract No. 033.1360, p. 8.

Responses to questions about characteristics in the wr, way,' and "sensitive to others who essential to managerial success revealed a greater work for h1." In the parent firm in Japan (J-J), the emphasis in the Japanese-owned firms on commu- typical manager "tries to achieve consensus" and nication within the organization both vertically and "permits broad latitude for subordinates," but is horizontally; this was true both in domestic (J-J) and also described as aggressive. American (J-A) operations. Managers in the Amer- While a reasonably uniform picture emerges for ican company Al would tend to "make as many de- the subsidiary J-A and its parent J-J, there was much cisions as possible at his/her level without bother- greater diversity among the characteristics of ing senior management," and "respect the chain managers within the American-owned firms. This of command, discuss ideas with immediate supe- was especially notable in company Al. Likewise in rior before discussing them with members of other company A2-J, respondents agreed on only one departments." In contrast, managers in the Amer- thing: that their superiors were aggressive. Coupled ican subsidiary of the Japanese company J-A with the similar characterization in the Japanese thought it important to "communicate extensively organizationJ-J,this suggests that, while ag- with managers in other departments;" managers in gressiveness has not always been viewed as cen- the parent firm (J-J) also stressed communication. tral to Japanese management, it may in fact be Within one of the American-owned subsidiaries in common in high-performing firms in both coun- Japan, A2-J, the responses indicated a feeling that tries. The survey results do paint a more hetero- each manager should make as many decisions as geneous picture of the managers in American- possible at his/her own level. Here the survey owned organizations. American companies oper- results do cor.--m a difference in management at- ating in japan exhibit some of the traits associated titudes between Japanese- and. American-owned with Japanese management, but it is the Japanese companies, with the American-owned electronics company which, as expected, has managers who firms exhibiting a greater degree of independent most strongly emphasize consensual decisionmak- decisionmaking even within their overseas subsid- ing and human relations. On this dimension, the iaries. composite managerial portraits indicate a clear difference in American and Japanese styles. Questions calling for a composite picture of the The human relations school stresses sensitivity manager immediately above the respondent elicited to subordinates. Table8A-2compares responses of several distinctions among the five organizations. nonmanagerial employees to questions related to On eight dimensions, those questioned were asked job satisfaction, together with data on rates of ab- to describe the actual characteristics of their senteeism and expenditures on employee pro- superiors (not the attributes they would like to see). grams. The Japanese-owned firms might be ex7 The managers in the U.S. subsidiary J-A were de- pbated to exhibit a greater degree of manager- scribed as: "readily accessible to subordinates ernployee jnteractionpresumably leadingto several echelons below," "permits broad latitude -greater satisfaction among the labor force. The for subordinates to work out solutions to problems results in tabb8A-2showthat very few workers 338 international Competitiveness in Electronics

Table 8A.2.Data Related to Employee Satisfaction Location of organization United States Japan Al J-A J.1 A2-JA3-J Percentage of workers rating themselves "very satisfied" with their jobs 20%28%2% 0 0 Percentage of workers rating themselves "satisfied" or "very satisfied" with their jobs 74%88%58%63%95 °/a Daily absenteeism 3% 1%<1 %< 1 % 1 °/a Social/recreational expenditures per worker $1.40$15 $33 $50 $38 SOURCE: M. A Maguire, "Personfml in the Electronics Industry: United States and Japan," prepared for OTA under contract No. 033-1360, pp. 19, 20. in Japan are willing to describe themselves as Matsushita's Purchase of Quasar highly satisfied with their jobs, but the picture changes considerablywith firms in Japan compar- What happens when a Japanese corporation takes ing more favorablyif "satisfied" responses are over an American firm? Changes in management included.z Japanese firms, known for their com- practices might offer insight into the Japanese ap- pany-as-family approach, might also be expected to proach. The purchase in 1974 by Matsushita Elec- spend more on social and recreational opportuni- tric of Motorola's Quasar divisionwhich pro- ties for employees. As table 8A-2 indicates, this is duced televisionsprovides a case in point, (Un- indeed true for organizations within Japan, regard- fortunately, conspicuous examples of a U.S. firm less of ownership. In any case, the results in table taking over a Japanese enterprise are lacking.) 8A-2 on job satisfaction should be interpreted with After Matsushita tc-ok control of Quasar, the new caution, as such questions typically yield high pro- owners reorganized the factory operations, located portions of positive responses. Moreover, clear-cut near Chicago, invested in new equipment, and telationships between expressions of job satisfac- began redesigning the product line, At the center tion and measured productivity levels are rarely of these efforts was the goal of improved product fou nd .3 quality. In contrast to the old production system, The differences observed between subsidiaries which relied on as many as seven quality control here and parent firms in Japan may result from con- inspectors per assembly line, Matsushita adopted scious decisions to downplay Japanese manage- a more integrated approach with responsibility for ment practices. The style that emerges is likely to quality spread broadly. By 1980, the defect rate on be a hybrid of American and Japanese practices. Quasar's assembly lines was about 2 defects per 100 In any event, this conclusion follows from the sets, compared to lh defect per 100 sets for Mat- survey data as a whole: there is no sharp contrast sushita's factories in Japan.4 These quality im- between the management approaches of American- provements were the result of systemwide changes. and Japanese-owned companies. Many of the pat- While resulting from a series of decisions made by terns observed are more closely associated with the Matsushita's management, they comprised far geographical location of the organization than with more than just matters of style or technique.. For ownership. Upper managers from the parent firm example, the company's extensive modernization tend to adopt many practices of the host country. .of the capital plant entailed expenditures of about On some dimensionse.g., accessibility of mana- 0 million for automated equipment, as well as an gers to lower level employeesthe Japanese-owned/ eritkrely new chassis factory in Mexico.5 Motorola firms do stand out. But in other cases, there are no offiaials stress that they knew just as well what had clear distinctions; only on measures of sensitivity to belone to make the Quasar facility more effi- to employee attitudes and participation are these cient, but had decided to allocate available re- consistent. sources to other parts of the company's business. If Quasar's gains in product quality and plant ef- 2Japanese workers also express relatively low rates of satisfac- ficiency came at considerable cost in new equip- tion with activities such as quality circles. Sec S. Takezawa. et al., Improvements in the Quality of Working Life in Three Japanese Industries. (Geneva: International Labour Offi-e,1982), 4J. Mihalasky and A.B.Mundell. "Quality and Reliability of pp. 77. 9898. Semiconductors and CTVs: United States v. Japan," Report No. S. E. Weed,T. R. Mitchell. and W.Moffitt, "Leadership Style, C972 prepared for OTA by Consultant Services Institute, Inc.. Subordinates' Personality and Task Type as Predictors of Per- under contract No. 033-1170.0, p. 77. formance and Satisfaction in Supervision." Journal of Applied ST. C. Hayes. ''The Japanese Way at Quasar," New York Times, Psychology. vol. 61, 1976. Oct. 16. 1981, p. Dl.

342. Ch. 8Human Resources: Education, Training, Management 339 ment investments, the emphasis on worker par- Conclusion ticipation and responsibility for quality is also significant. Quality control circles were not in- Both the survey results and the Matsushitaexam- troduced until recently, but Quasar employees have ple indicate that well-run organizations tend to be been encouraged to set their own production tar- open to new ideas and methods, including those gets and to meet in informal weekly discussions coming from the lower levels of the organization. about plant operations with foremen. Such prac- Distinctions between Japanese- and American- tices are hardly unique or exotic, but the atten- owned firms are fewer and less clear-cut than tiveness to all aspects of the manufacturing process sometimes claimed. While American employees stands out. Still, none of this has helped Quasarex- might resist some of the techniques associated with pand its market share L,ubstantially. Japanese management, worker participationeven Quasar, like other Japanese subsidiaries in the loyalty to the firmcan be fostered in a variety of United States, shows a flexible and adaptive man- ways. Some of these methods smack of paternalism, agemen` 3tyle, with manufacturing operations and but not all. As a number of American firms have quality control having a central place. Nonetheless, amply demonstrated, worker participation and at- if and when Japanese companies hire still larger tention to human relations can be a big help in 1 oportions of American managers, and adapt fur- building competitive organizations. ther to the U.S. environment, they may become more like wholly American organizations.6

A recent study by the Japan External Trade Organization nal Trade Organization. September 1981. Data on managerial tIETROJ on Japaneseowned manufacturing operations in the styles collected for the JETRO study confirm the trends described United States indicates that the number of Japanese nationals here: Japanese subsidiaries evolve styles that mix features com- transferred to subsidiaries tends to decrease over time. See mon to the Japanese model with other practices more charac- "Japanese Manufacturing Operations in the U.S.." Japan Exter- teristically American. CHAPTER 9 Employment Effects Contents

Page Overview 343 Impacts of Technical Change in Electronics on Employment 344 The Automation Debate of the 1950's and Since 344 Factors Affecting Employment Levels. -, 347 Employment Trends in the U.S. ElectrOnics ndustry 349 Consumer Electronics '353 Semiconductors ,. 356 Computers \ 357 Effects of Import Penetration and OffShore Assembly 359 Consumer.Electronics 1\ 360 Semiconductors -, 362 Foreign Investment in the United States \ 363 Projections of Employment Within the ElectronicsIn1dustry 364 Future Employment Patterns in Other Industries 366 Manufacturing 367 Services \ 370 Summary and Conclusions 372

List of Tables Tobit Nu. Page 74. Employment in Selected Portions of tile U.S. Electronics Industry 350 75. Predicted Growth Rates by Occupational Category Over the 1980's 365 76. Occupational Distributions in Electronics as of 1980 366

List of Figures Figure No. Page 55. Unemployment in Industrial Nations. 345 56. Labor Productivity and Employment by Sector of the U.S. Electronics Industry . (a) Consumer Electronics 351 (b) Semiconductors 352 (c) Computers 353 57. U.S. Employment in Television Manufacturing 354 58. U.S. Employment in Consumer Electronics 355 59. U.S. Employment in Semiconductors ar1:1 Related Devices 356 60. Effort Levels Associated With ProdUct and Process Design for Integrated Circuits 357 61. U.S. Ethployment in Vacuum Tube Manufacturing 358 62. U.S.' 'Employment in Cqmputer Manufacturing '359 Employment Effects

Overview

Shifts in the competitiveness ofan industry Within the industry, changes in competitive- like electronicsor for that matter technical ness have immediate consequences for employ- change alonehave both direct and indirect ment. If the U.S. electronics industry declines consequences for employment. In addition to in competitiveness, and sales fall in domestic changes in the labor force requirements both and/or foreign markets, employment will fol- of firms within the industry and firms thatsup- low. If rates of increase of sales drop, employ- ply it, the effects can spread broadlyacross the ment may also declinedepending on increases economy. jobt,pportunities within the United in labor productivity. Similarly, if. U.S. elec- States appear or disappear with changes in de- tronics firms expand their overseas production mand for electronics products, with shifts in activitiesfor re-import or for sales in foreign international competitive position, and with in- marketschanges in domestic employment creases in productivity. These forces interact normally follow. As competitive advantages in complex fashion. shift internationally, labor market dislocations Will continuing developments in electronics can occur even if the total number of jobs re- computers, office and factory automation, in- mains the same. Such dislocationscan include formation servicescause employment to in- geographical shifts in demand for workers, crease or decrease? Such questions have been along with changes in educational and skillre- debated for years, in the context of this and quirements; as computers and other electronic other industries. The conventionalresponse is systems have become more sophisticated, that technical change creates, in the aggregate, white- and grey-collar jobs have expanded more jobs than it destroys. While the kinds of much more rapidly than openings for unskilled jobs available will changeas terminalsappear or semiskilled workers. on more desks, opportunities for systems ana- Shifts in the international competitiveness of lysts (who plan and help operate dataprocess- American electronics firms also affect other ing installations) replace those for keypunch parts of the economy. Moreover, structural un- operators, for one instancenew technology employment can be created by changes in elec- creates new demand fast enough that total em- tronics technology that alter the ways goods ployment goes up. The conventionalresponse are designed and manufactured. Electronic assumes that such patterns will continue. But, typesetting has reduced the need for skilled just because in the past technical changecre- workers in newspaper publishing. Technologi- ated more jobs than it destroyed does notmean cal change may create new jobs for supervisory that this will be true it 'ile_tvture. Suchques- and maintenance workers, but it is hard to im- tions are broader than can be addressed here. agine that as many people will be employed in Too many forces affect levels of employment,' designing, manufacturing, and maintaining in- not to mention skill requirements. Analysis on dustrial robots as are displaced by them. Still, a detailed, disaggregated basis sufficient to iso- net effectsparticularly over extended periods late the influences of electronics (andupon it) of timecan seldom be disentangled from the would be extraordinarily difficult. This chapter other factors on which employment depends. has more limited aims: to summarize what is If aggregate economic growth is slow, andpro- presently known about employment in elec- ductivity risese.g., because of in-vestments in tronics, both past trends and future prospects. labor-saving equipment like robots,or com- 344 International Competitiveness in Electronics puter-integrated manufacturing more generally years. Next, data on employmenttrends in elec- jobs will be lost unless other sectors of the tronics are examined in the context of import economy,such as services, compensate. penetration, as well as offshore manufacturing by American firms. The chapter surveys em- The preceding chapter explored the educa- ployment forecasts for electronics, along with tion and training of American workers, as well case studies of impacts on othermanufactur- as management practices which determine ing and service industries. While there is no how effectively the talents of the labor force way of knowing how aggregateemployment are utilized and the possibility of shortages of will fare, technological changetogether with those with specialized skills. Chapter 8 in- shifts in the competitive positions of American cluded extensive comparisons between the electronics firmswill clearly have major im- United States and Japan. Here, the focus is pri- pacts on some industries and some job marily on the United States, beginning with a categories. review of the automation debates of earlier

Impacts of Technical Change in Electronics on Employment I' The Automation Debate of the The 1970's brought renewed concern; eco- 1950's and Since nomic growth slowed and unemployment rose. The trend was sometimes masked by the ups People have worried over technological and downs of the business cycle, but by the end change because of its impacts on employment of the decade, as figure 55 shows, it was clear and sometimes actively resisted new technol- that unemployment had been steadily rising in ogiesat least since the beginnings of the in- most of the industrialized West. Now the ques- dustrial revolution. The automation scare of tion has become: Will this trend persist? the 1950's focused on computers taking over Rather than mainframe computers as in the the workplacea fear that has resurfaced, 1950's, people now point to microprocessors more so in Europe than the United States. and microcomputers as the new technologies Twenty-five years ago, some commentators with thegreatest potential job-displacing predicted steadily rising unemployment due to effects.' As was the case 25 years ago, optimists automation; others were skeptical that comput- and pessimists view the consequences of such ers alone would have such grave consequences. developments quite differently. To the opti- Throughout the 1960's,' a number of interna- mists, labor-saving technology is nothing new. tional groups, including the Organization for Many more jobs will be created than lost, they Economic Cooperation and Development say. Moreover,,in the short term the impacts (OECD) and the International Labour Office, of microelectronics will not be that dramatic continued to study the effects of computers and because most investments in automated equip- automation on employment. As it happened, ment come during periods of economic the industrial nations experienced an upswing growth, when capital is available. As a result, in economic growth during the 1960's that put workers may be redeployed but only rarely will the automation debate temporarily to rest. Fall- lose their jobs. The optimists view structural ing levels of unemployment were sufficient indication to many that overall demand was the 'See, for example, C. Norman, Microelectronics atWork: Pro- key to jobs, with structural aspects decidedly ductivityand Jobs in the World Economy, Worldwatch I'apor Ad- cr; lnnti ae aaorPaatP ripmarirl prim/ 39 (Washington. D.C.: Worldwatch Institute, October 1980);__ _ Ch. 9Employment Effects 345

Figure 55.Unemployment in Industrial Nations'' 13

United, 10 Kingdom /

9 I United // 8 1 StatesiI Ii, /./ 1, / 'e / '7' --4'.'. E 6 "-- r..' r 0 / .... , ,r ...v.,, ,,, 1 E 5 // %, "c / I..."' France / , Italy --- 4 i , i '. -...._ / ^ ... '''''. ./*.---....-"*.--....of / ."\ t,...../11.-=...... %.%\I . I\ i i 1"- "11`...... ''s,...Z..... / \*.c..// 1 '' .,/ \ / 1 j% / ... a .,,, v. .... West . ,, "...... , I...... Germany 2 Japan 000 NO

0 yr 1960 1965 1970 1975 1980 1982b Year

o:,f,,,,,ahale,; the U S. concept Data for the United Kingdom exclude Northern Ireland b1982 data tor an c,nrnthes but ee, Unded Stoles and Italy based on fast 9 months.

SOURCE. Economic Report of the Pres,dent (Washington, D.C.: U.S. Government Printing Office, February 1983),p. 237. transformation as aprocess that creates jobs todial work, fast foods, selling insurance). Still, in newer sectors v.-1the economy: employment from the optimist's viewpoint., the expansion in manufacturing may shrink but opportunities of high-technology industries means more op- will increase in services; as the proportion of portunities for an educated labor force. Com- manual workers declines, the number of white- petition from low-wage, newly industrializing collar employees grows. Automation, further- countries (NICs) need not cause great concern; more, will free people from some of the worst so long as the world economy continues to jobs: dirty, boring, dangerous factory work; grow, industrialized iiations can concentrate sorting and filing; processing checks; perhaps on advanced products made by better paid and even delivering the mail. To the pessimists, of better trained workers, leaving the lower course. some of these lobs are not so hadand torilnnInov Qprtr,r c #1,. Mir'c 17. 346 International Competitiveness in Electronics

Others are less sanguine, their skepticism and capabilities of workers and the require- rooted in the belief that the world economy is ments of industry (ch. 8). They argue that em- now fundamentally different than in the 1950's ployment statistics for the United States al- and 1960's. "Structuralists" argue that perma- ready underplay the extent of real unemploy- nent shifts spelling chronic unemployment and ment, not to mention underemployment.3 underemployment have taken place. Funda- The debate between the optimists and pessi- mental to this view is the slow economic mists ranges far beyond the electronics indus- growth of the 1970's; to the pessimists, sudden try. But electronics technology has been a nat- rises in energy prices and other shocks to the ural locus of concern because it so clearly em- international economy are not enough to ex- bodies labor-saving advances by which ma- plain the slowing pace of growth. They argue chines perform tasks that people did in the that, at least in manufacturing, the expansion past. No wonder labor unions it?. the United in output needed to maintain cm-rent employ- States but particularly in Western Eu5ope ment levels has been increasingi.e., that out- have continued toraisequestions about put must grow more rapidly than in the past automation and electronics, 'and sometimes ac- in orthir to maintain a constant number of jobs. tively resisted new production methods. If true, and if this trend persists, it will become' more and more difficult toexpand employment The question: "How will electronics technol- by stimulating demand.2 Since labor produc- ogy affect employment?" is unanswerable. Pos- tivity in the manufacturing sectors of industrial ing the question more narrowly helps a little: nations has risen consistently faster than gross Will continued developments in electronics national productiGNPjahe pessimists empha- drastically reduce the number of workers size that compensating expansion in employ- needed in the manufacturing sectors of ad- ment must come from sectors other than manu- vanced economies? Will the effects be benefi- facturing. Many also argue that structural un- cial through elimination of burdensome tasks employment in advanced industrial nations while creating new and more interesting jobs? results from a permanent shift of labor-inten- These phrasings still cannot be treated with sive production to lesser developed countries any precision, but at least are more suggestive. and NICs, where wages are low. In the longer The problem is that no methods exist for deter- term, this might be a positive force; if interna- mining employment shifts caused exclusively tional specialization takes place, the more ad- by technical change. Too many other forces are vanced nations should be able to concentrate at work. A second analytical probiem relates on capital- and knowledge-intensiveindustries, to the type of employment impact. Advances and expand their employment in these sectors. in electronics may eliminate a job in one But in the short run it leads to severe disloca- plantbut a similar job may open in a nearby tions, already evident, for example, in consum- firm or in a distant city. Alternatively, a dis- er electronics or steel. placed worker might be able, to find employment only after retraining, or even reeducation. The same causes and effectstechnological change, productivity growth, shifts in ;nterna- tional comparative advantage, technology gaps 'One study has claimed that 80 percent of American workers by the optimists are "misemployed"i.e., are doing jobs for which they are ill- are thus viewed differently suited. See W. W. Harman, "Chronic Unemployment: An Emerg- and the pessimists. The latter see them as sig- ing Problem of Postindustrial Society," The Futurist, August nals of persisting unemployment. Unlike the 1978, p. 213. optimists, they emphasize obstacles to adjust- Leontief paints a grim picture 'of the effectsof technological change, mismatch, and misemployment: ment such as mismatches between the skills To argue that workers displaced by machines show\d necessarily be able to find employment in building these machipes does not make more sense than to expect that horses displaced b§ mechanical Ch. 9Employment Effects 347

From the perspective of the individual,geographical moves or retrainingcan aggravate what is already a severe blow on psychological as well as more tangible grounds.

Factors Affecting Employment Levels Directly or indirectly, the ability of American firms to compete internationally links Many of the forces that affect employment. Increasing sales here and abroad provide the foundation for a growing labor market, with aggregateex- pansion creating new job opportunities unless labor productivity goes up even faster. Conven- tional methods of forecasting labor market demand begin with output projections. Ina given sector, output and employment will depend in complex fashion on aggregate demand; ina period of economic downturn, job opportunities can still increase in some industries. While this has often been true in electronics, recessionary pressures during 1981 and 1982, as in 1974 and 1975, show that the semiconductor industry is far from immune from sales slumps and layoffs. For years, the interrelation between employment and inflation was pictured in terms of the Photo credit: PC well-known Phillips curve, which showed that Final adjustments during color TV assembly high rates of inflation tended to correspond to low rates of unemployment, and viceversa. But Employment isclosely linkedto labor by the end of the 1970's, the Americanecon- .-produ `ivitycommonly measured in terns of omy 'seemed prone to simultaneous inflation output per man-hour. If firms can pr duce and unemploymentanother gloomy portent more with the same amount of lab r, the to those on the pessimistic side of the struc- economy as a whole expands and sooes in- tural unemployment question. Onereason is dividual purchasing power. Growlin pur- wage and price rigidity. When demand falls, chasing power can create new dem,and which companies are reluctant to cut prices asa will in turn create new jobs; thus increases in means of expanding output, workers reluctant productivity do not of themselves/result inem- to accept pay cuts to reduce costs. Rather than ployment losses. But if the overalleconomy is greater output and employment at lower wage stagnant or growing only slowly, productivity and price levels, prices stay highaggravating growth in a given industry can well lead, not inflationsales drop, output must be cut, and only to decreasing job opportunities in that in- workers are laid off. Nonetheless, recentwage. dustry, but to net job losses within the econo- concessions in the steel and automobile indus- my. tries show that adjustment is possible if the As this implies, sectoral shifts must becon- slump is serious enough. sidered. Aworker rlignlaredby ricino .-

348 International Competitiveness in Electronics tivity and foreign competition in consumer labor productivity increased and costs were cut electronics finds little solace in growth else- to the point that vastly greater numbersof peo- where in-the economy. Similar patterns appear ple could afford to buy cars. Likewise, the in- at higher levels of aggregation. As pointed out troduction of color television cut into sales of in chapter 5 (see fig. 32), employment inboth black-and-white sets but expanded overall de- manufacturing and agriculture has shrunk rela- mand for TVs. Many examples could be drawn tive to services in the OECD nations. The serv- from the computer industry. ice sector makes an ever-growing contribz:tion to U.S. GNP, and the rate of job expansion The export competitiveness of domestic there has been high. What of productivity in firms, as well as market penetration by imports, services? Since productivity has grown less directly affect employment. Greater sales in ex- rapidly in services than in manufacturing (al- port markets mean more jobs athome. On the though productivity in many service sector cat- other hand, an influx of foreign goods may put egories is notoriously difficult to measure). Americans out of work. In recent years, con- cverall employment levels have been main- siderable attention has focused on jobs lost to tained in part by transfers of labor from manu- foreign low-wage industries making products facturing to lower productivity service sector such as TVs or textiles and apparel. Never- jobs. Of course, factory workers cannot always theless, competition with advanced nations can quickly move to service jobs, nor may they be equally importantevident in products want toparticularly if the jobsavailable are ranging from automobiles and machine tools low-paying or menial. The point is that sectoral to integrated circuits and aircraft. Asindustries shifts always irop;y some degree of dislocation. like electronics become more thoroughly inter-.- national in character, it is seldom easy to disen- The impacts of technological change take tangle the costs and benefits flowing from several forms. Automation, interpreted broadly shifts in competitive strength. Overseas pro- as extending to jobsoutside the traditional duction by American firms can be viewed as manufacturing sector, cuts into the need for a loss in domestic job opportunities;it can also labor. Computers eliminate jobs for file clerks; be seen, in at least some cases, as an entree into banking machines displace tellers; instead of new and expanding foreignmarkets (see app. three people in the cockpit, new commercial B on offshore manufacturing for an outlineof aircraft need two. Great Britain's telephone the complexities of such judgments). systerri provides a quantitative example: when electromechanical equipment was phased out Finally, employment levels always depend to in, favor of electronic switching during the some extent on the fit betweenthe demand for 1970's, employment dropped from over 90,000 manpower and the skills andcapabilities of the to 65,000.4 work force. Structural shifts affect not only the The effects of new technology depend in employment levels in various economic sec- large measure on the motives for its introduc- tors, but the kinds of peopleneeded. In the tion. Investments aimed at rationalizing the United States, the unemployed Youngstown production procer- by cutting costs, improv- steelworker may neither be qualified nor desire ing efficiency, or adjusting to newconditions to move into a Silicon Valleyelectronics com- tend to cause net declines in job opportunities. pany, especially since the pay isunlikely to be The British telephone system is a case in point. very high. In advanced economies,growth in On the other hand, technical change may ex- services has led to a variety of changes inlabor pand output or create new markets, resulting markets. In Sweden, for example, as the econ- in many more jobs. Henry Ford's moving as- omy has grown and theservice sector ex- , rrmin inhr,frrrn nnrtirinatinnamong older Ch. 9Employment Effects 349

lieve that none is available.5 At thesame time, grams to help attract industry, retraining has women have joined the labor forces of the in- never been approached systematically, in strik- dustrialized nations in greater numbers, taking contrast to nations such as West Germany; ing many of the service sector jobs. in addition to the vocational programsmen- The match between supply and demand in tioned in the preceding chapter, the German the labor marketnever perfectis thusan in- Labor Market Office matches unemployed trinsic part of the employment question. To workers with openings through a nationwide some extent, problems of skills and training are computer survey' There are no parallels in the those of response time; people's choicesmay United States. lag new opportunities, as may programs of This brief review illustrates the difficulty of study in educational institutions (ch. 8). Short- assessing the consequences of changes in tech- ages of entry-level electrical engineers in the nology or competitive positioneven in a single United States have reflected, not only r,..pid industry like electronics. First, many of the fac- growth in demand for the products of the elec- tors are interrelated. How can shifts in compet- tronics industry, but slow response within the itiveness be isolated from the effects of aggre- educational system to new labor market de- gate economic growth, which determines demand. This is one way in which employment mand for the industry's products? How directly is affected by public policies, at least to theex- must gains or losses of jobs elsewhere in the tent that schools and universities depend on economy be linked to changes within the elec- governments (including State and local) for re- tronics industry (e.g., new technologies) to sources. Government programs can also help justify an attribution to electronics? Should vir- men and women who find themselves unem- tual employment and unemploymentjobs that ployed or underemployed developnew skills would or would not exist in the absence of and find new jobs. Adjustment is butone of changes in electronics technologybe in- several avenues; during the 1930's, the Federal cluded? Finally, which impacts are most sig- Government instituted many programs toex- nificant? Those on individuals? On companies? pand employment. These massive public works On entire industries? Or are all three of com- efforts drew support from Keynesian theory, parable importance? What of regional disloca- which held that demand stimulation could help tions? There can be no easy answers to the ensure full employment. general question of whether continuing devel- Despite the experiences of the Depression, opments in electronics will have positive or and the many job programs since, the United negative consequences for employment in the States does not have a comprehensiveman- United States. power policy at the national level. Although The following sections look in more detail some States have set up worker training pro- first at changes within the electronics industry, then at effects on other sectors. °I1. Berg lind. "Unemployment and Redundancy in a %lost- Industrial' Labor Market," Work and Technology, M. R. Haug and J. Dofny (eds.), Sage Studies in International Sociology °L. Dobyna, "America Works When America Works," NBC (Beverly Hills: Sage Publications, 1977), p. 201. White Paper, Juno 25, 1981.

Employment Trends in the U.S. Electronics Industiy Changes in employment within any one in- declined from 26.0 to 21.8 percent of the titoefr,, fnien 'Anne, I.. I 1. i ir_ P .1 -350 international Competitiveness in Electronics sectoral basis; the number of jobs in the U.S. equipment, public address systems, and consumer electronics industry has declined, amplifiers nor musical instruments all fall while in computers and semiconductors ex- within SR, 36514 panding output has brought rising employ- 3674Semiconductors and Related De- ment. vices. (This category includes virtually all types of microelectronic components, Analysis of such trends depends on how the ranging to solar cells and bubble memo- industry is defined and subdivided. For in- lies, those manufactured by captive plants stance, data published by the Electronic Indus- as well as merchant firms.) tries Association (EIA) show 1.6 million work- 3671Radio and Television Receiving ers in the entire industry in 1982.8 EIA,how- Type Electron Tubes, Except Cathode Ray. ever, bases its tabulation on very broad SIC (Virtually all vacuum tubes are included (Standard Industrial Classification) categories. except for TV picture tubes and other cath- Among these is SIC 367, "electronic compo- ode ray tubes, and special purpose devices nents and accessories," which has nine sub- such as klystrons or X-ray tubes.) divisions. Only one - 3674,. "semiconductors 3573Electronic Computing Equipment. and related devices"is among the portions of (Processors and peripherals of all types fall the electronics industry that OTA has focused into SIC 3573.) on, otherse.g., "electronic coils, resistors,and capacitators"being less illuminating in terms In referring below to these S!C categories, of international competition. Therefore, discus- more inclusive namese.g., consumer elec- sion of employment in the rest of this chapter tronics for SIC 3651have been adopted. Both is limited to the following four SIC categories:9 semiconductors (3674) and the vacuum tubes they have largely replaced (3671) are examined, 3651Radio and Television Receiving so that growth in the first category can be com- Sets, Except Communication Types. (De- pared to contracticm in the second, spite the title, this SIC group includes more than just radios and TVs, extending to During the 1970's, employment grew in two nearly all home entertainment or consum- of these four SIC categories, as table 74based er electronic products; consumer audio on data gathered by the Bureau of Labor Statis- tics (BLS)shows. In microelectronics, em- ',Electronic Market Data Book 1983 (Washington, D.C.: Elec- tronic Industries Association, 1983), p. 144. This is the total of ployment has doubled, and in computers it has Labor Department employment figures for four Standard Indus- gone up even faster, while the consumer elec- ±: `al Classification categories: SIC 3651 (radio and TV receivers), tronics category has shrunk. (Most of the con- 366 (communications equipment), 367 (components), and 3573 (computers). Communications, with more than 550.000 employ- traction in vacuum tube production predates ees in 1982, makes up one-third of the total. 1972.) In 1982, the nearly 800,000 workers °Defined in Standard Industrial Class.ification Manual 193'2 covered by the SIC codes in table 74 totaled (Washington. D.C.: Office of Management and Budget, 1972), pp. 190 (SIC 3651), 193 (SIC 3674). 192 (SIC 3671). and 180 (SIC slightly more than 4 percent of the 19 million 3573). men and women in tho U.S. manufacturing

74.Employment in Selected Portions of the U.S. Electronics Industry

Number of employees and percentage of production workers (In parentheses) SIC category 1972 1980 1982° 3651, consumer electronics 114,500(74%) 85,900(70%) 74,400(67%) 3674, microelectronics 115,200(51%)226,900(44%)230,000(40%) 3671, vacuum tubes 46,400(70%) 42,600(52%) 43,400(61 Ch. 9Employment Effects 351 work force. making even this portion of the number of jobs doubled, with productivity ris- electronics industry larger than, say, steelmak- ing almost as fast until the mid-1970's. Past this ingwhich employs half a million.* point productivity growth has slowedbut, as Figure 56 compares trends in labor produc- pointed out in chapter 5, productivity trends tivity and employment (for production workers in terms of value can be misleading when tech- only) over the past decade for each of the cate- nical change is as rapid as it has been in the gories except vacuum tubes. In all three charts, data processing industry. Even so, value-added productivityisgiven as value-added per productivity in computer manufacturing has production-worker hour in real, Ltflation -ad- risen much more rapidly than for U.S. manu- justed terms. Productivity growth in consumer facturing as a whole. Many jobs have also been electronics, figure 56(a)where employment created in semiconductors, fig. 56(b), where declinedhas paralleled the all-manufacturing productivity gains were again substantially average, growing'slightly faster in earlier years. above the all-manufacturing average. The cy- In contrast, computer manufacturefig. 56(c) clical nature of employment in the semicon- ;;hows the most rapid rise in employment; the ductor industry distinguishes it from both consumer electronics and computers; the sensitivi- "BLS figures for the first 10 months of 1982 show 18.9 million ty of semiconductor production to recession workers in manufacturing-11.2 million in durable goods, 7.7 is magnified by the tendency of purchasers to in nondurabtes. quickly cut back on orders when their own out-

Figure 56.Labor Productivity and Employment by Sector of the U.S. Electronics Industry (a) Consumer Electronics (SIC 3651) 20 110,000 /14 I I 100,000

I I Productivity 90,000 all U.S. I manufacturing O

80,000 (S O

'C3 2 a. 70,000 .6

w Productivity-- E consumer electrcinics 60,000

50,000 352 International Competitiveness in Electronics

(b) Semiconductors (SIC 3674)

2,0 90,000

80,000

20 0 O Productivity ' 70,000 all U.S. 0 manufacturing 0

2 - / , i i / 60,000 6.-0 --- I I / 10 ''' r II / 7 1 / 1 /rI A / 50,000 % Employment Productivity , % i produclion semiconductors ti/ workers

40 000 0 1965 1970 1975 1980 Year put drops, sometimes to double-order in up- drop while productivity climbs, particularly if swings for fear of shortages. coupled with rapid technical change and high investment. But as the examples from electron- As the plots in figure 56 demonstrate, the por- ics in figure 56 illustrate, there can be a great tions of the electronics industry that showed the deal of variation across sectors: productivity highest rates of productivity growth also ex- rises at different rates; sometimes employment perienced the highest rates of employment goes up, sometimes down.Still, over time, tech- growth. Increases in productivity were associ- nologically progressive U.S. industries have ated with the creation of jobs, not their elimina- generally experiencednot only above-average tion. The reason is simple: output in computers productivity gains, decreasing real prices, and and semiconductors grew at very high rates, increases in salesbut relative increases in em- spurred by exports as well as domestic sales. ployment as well." While an increase in em- The domestic market for radios and TVs grew "Denison and others have studied the contributions of more slowly, exports were small,and import technological change to economic expansionfor example. E. penetration has been severe. Denison.Accounting for United States Economic Growth (Washington, D.C.: Brookings Institution, 1074). For an analysis As the cases of computers and microelec- of trends in electronics, see W. Kendrick.--Impacts of Rapid frnnirc when rates ofchance in tech- Technological Change in the U.S. Business Economy and in the Ch. 9Employment Effects d 353

it 130,.200 35

120,000

u 30 110.000 Productivity 0 computers 3 10D,0002 25 0 /11\ 90,000 a Employment 0 A / production // / workers // \ // 80 000 6 / \ fal / ,/i ----- 70,000 / . '''s ooe. ..% / ----- ' ' 151 // V ...-'' / .....-.- Productivity- all U.S. 60,000 1.. el...... 1.-----'°.....".... manufacturing le i 50,000 1967 1970 1975 1980 Year

NOTE Census Bureau ',owes for employment may not agree exactly with BLS tipures cited elsewhere in this chapter. BLS figuresare collected at the plant level and may include some workers from t the r SIC categories if a plant makes products that far under several categories Census figures are used Inthese charts for consistency with productivity data from the Census Bureau Val-radded figures have been converted to constant dollars using the implicitprice deflator for consumer durables.

SOURCES 1965-71 1977 Census of Manufactures. 1978-80-1980 Annual Survey of Manufactures.

ployment is not inviolably associated with the which was especially precipitous over the early development of new technologies and produc- 1970's (as noted on the plot, the data cover TVs tivity growth, the pattern is not an uncommon only, not consumer electronics as a whole). one. That employment goes up does not, of Jobs for production workers dropped by half course, mean that adjustment problems disap- between 1971 and 1981. Over these years, a pmrbut it can provide leeway to deal with number of U.S. manufacturers either merged them. The next sections examine employment with Japanese or European producers or left by sector in more detail. the business. On the other hand, the industry now includes more than 10 foreign companieS Consumer Electronics with assembly operations in this country that contribute to the employment totals in the Trends in Employment figure. 354 International Competitiveness in Electronics

Figure 57.U.S. Employment in Television Manufacturing

60,000

50,000

40,000

30,000

20,000

All employees ProducWn workers 10,000

0 1980 1966 1970 1975 Year aMonochrome production in the United States had dropped to low levels by 1975.

SOURCES: 1966-70 Television Receivers and Certain Parts Thereol (Washington, D.C.: U.S. Tariff Commission Publication 436,November 1971), p. 1971-75 Television Receivers. Color and Monochrome. Assembled or Not Assembled. Finished or Not Finished. andSubassemblies Thereol (Washington, D.C.: U.S. International Trade Commission Publication 808, March 1977), p. A-117. 1976, 1977Color TelevisionR..ceivers: U.S. Production, Shipments. Inventories. Imports. Employment. Man-Hours. and Prices. Fourth Calendar Quarter 1977 (Washington. D.C.; U.S. loternationa: TradeCommis- sion Publication 866, March 1978), table 5. 1978, 1979Color Telet,ision Receivers: U,S. Production,Shipments, inventories, Imports, Employment, Man. Hours, and Prices, Fourth Calendar Quarter 1979 (Washington. D.C.; U.3. International Trade Commission Publication 1036. February,1980). p. A-7. 1980, 1981Color Television Receivers: U.S Production, Shipments, Inventories, Exports, Employment, Man-Hours, and Price:, FirstCalendar Quarter 1982 (Washington, D.C.: U.S. International Trade Commission Publication 1245, May 1982), table 5.

ago; the causes of theemployment declines in tronics market (ch. ,4, table 8), and rathbi less figure 57 extend well beyond import competi- in terms of jobs. TotalT employment in SIC tion or offshore assembly, with technological 3651which covers many other consumer change a major force. Although the contribu- electronics productsis considerably greater, tions of the various factors cannot bequantified as shown in figure 58. Still,the number of with any precision, the spread of solid-state workers here has been in decline since 1973, chassis designs and associated manufi, for similar reasons. methods dramatically reduced employment requirements in the industry. Productivity Figure 57 includes only those people in- As domestic output of TVs grew overthe vnlved inTV manufacturing. Television ac- years covered by figure 57 (seech. 4, table 9), Ct. 9 Employment Effects 355

Figure 58.U.S. Employment in Consumer Electronics (SIC 3651) 140,000

120,000

(al 100,000 cu O " N. 80,0007 \ O \ iv --- .... 60,000 E 7

40,000t All employees ---- Production workers

20,000

0 /V11j___ 1960 1965 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 Year SOURCES: 19601965-1977 Cerius of MOnufactures! 1972.82Bureau of Labor Statistics. output divided by the number of production in ch. 4 illustrates the rise in imports of in- workersjumped from 150 sets per worker in complete sets and subassemblies over the lat- 1971 to560 in 1981. Interms of value-added ter part of the 1970's) simply dividing the total per production worker, productivity was up by output of TVs by the number of employees con- about .40 percent during the decade7-.a trend siderably overstatesproductivity gains. not far different from that for the broader con- However, the value-added'productivity meas- sumer electronics category seen in figure ures adjust for this. 56(a).* During this period, the proportion of domestic value-added dropped as American Thus, there is no question that; ductivity manufacturers shifted labor-intensive opera- increased considerably during th'1970's, the tions to developing countries; whether made result of design changes and autos ation driven by American- or foreign-owned companies, by competitive pressures (ch. 6). As manufac- TVs produced in the United States now include turers moved from monochrom' `o color pro- more imported components and subassem- duction, they shifted to more highly automated blies. manufacturing facilities. Somewhat later, re- Because of these trends (table13 designed solid-state chassis cut the number of parts, hence the labor content; only 6 percent 'In terms of .:;onstrit 1972 dollars, annual value-added per production worker in TV manufacturing went from $22,200 in of the color TVs made in the United States 1971 to $31,600 in 1977, falling to $27.300 in 1981. See 1977Cen- were solid-state models in 1970, but by 1976 susof Nfanufactures: CommunicationEquipment, Including essentially all had been redesigned around Radio andTV. MC77 -I -36D (Washington, D.C.: Department of Cominerce, jutie'1980). p. 36D-5 and1982 U.S. Industrial Outlook transistors." A good part of the productivity (Washington,D.C.: Department of Commerce. January 1982), p. growth over the1970'sresulted from changes 343. C:onversions to 1972 dollars were made using the implicit price deflator for consumer durablesEconomicReport of the 356 International Corbpetitiveness in Electronics in chassis design and associated manufactur- to low-wage offshore locations during the ing methods. 1970's. While there are no precise figures on foreign workers employed in these plants, the Although productivity gains in consumer Department of Labor believes that the number electronics have contributed to declining em- may be over 30,000more than employed in ployment; the composition of the work force domestic TV operations.12 These people substi- has not changed greatly. As table 74 and figure tute quite directly for American workers. 58 both illustrate, the ratio of production workers to nonproduction workers has decreased Semiconductors relatively slowly. In TV manufacture rather than consumer electronics as a whole, the shift Since the mid-1950's, employment in semi- has been greater, mostly taking place by the conductor manufacture has grown rapidly, mid-1970's (fig. 57). The semiconductor indus- from a few thousand when production of semi- try, for one 'example, has seen more rapid conductor devices was just getting underway, changes in skill mix (table 74). to well over 200,000figure 59. These totals include captive manufacturing. During two Imports' and Offshore 'Manufacture periods-1969-72 and 1974-76employment Earlier chapters described the inroads made dropped sharply as a result of recession. by imported TVs, both monochrome and col- As figure 59 also shows, the proportion of or. Few black-and-white sets are now manu- production workers in the domestic industry factured here. Orderly Marketing Agreements has declinedfrom 66 percent of the total work (OMAs) restricted imports of color sets during force in 1963 to 40 percent in 1982. Major the period 1977 to mid-1982, but figure 57 causes include the transfer-of production oper- shows that the quotas did not arrest employ- ations offshore and advancing technology. ment declines. Still, jobs would have been lost More complex manufacturing methodsin- even faster without OMAs. cluding automationhave increased the rela- American consumer electronics firms relo- tive need for technicians and other nonproduc- cated many of their manufacturing operations ulnformation from Department of Labor.

Figure 59.U.S. Employment in Semiconductors and Related Devices (SIC 3674)

250.000 All employees ----Production workers 200,000 a) a) 0 E 150.000 a) 0 05 E 100.000 z

50,000

tt 0 19631970 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 Ch 9Employment Effects 357 tion workers. High levels of research and devel- ably small impacts on domestic employment, opment have contributed to expansion in non- offshore investments driven by lower wages production ranks; the number of man-hours directly displace American workers, just as in devoted to integrated circuit design has been consumer electronics. Offshore manufacturing increasing exponentiallyfigure 60. Techno- also contributes to the declining proportion of logical advance in microelectronics has thus production employees in the United States. Un- been paralleled by a decrease in semiskilled skilled assembly labor accounts for most of the and unskilled employees relative to skilled jobs' overseas; U.S. firms employ about three- workers and professionals in U.S.7based man- quarters as many people in their foreign plants ufacturing. The result has been an "upskilling" as they do here: around 180,000, of which more of the domestic labor force. Employment op- than 80 percentas many as 150,000are pro- portunities for technical personnelengineers, duction workers.14 Among U.S. merchant semi- scientists, technicianshave grown rapidly. As conductor firms, perhaps 90 percent of all these trends continue, the proportion of pro- assembly work is performed overseas.15 duction workers in domestic semiconductor Many U.S. companies make semiconductors operations will fall even more. solely for internal use, but no disaggregation American semiconductor firms transferred of employment data is available for these cap- "back-end" operations overseas at a rapid pace tive facilities. While most produce specialized during the 1960's, with more than 50 foreign devices in relatively low volumes, with con- manufacturing plants established during the sidera'o)e variation in month-to-month levels, decade." While point-of-sale plants have argu- IBM is a large producer and large employer. "A Report on the U.S. Semiconductor Industry (Washington. Because some of the overhead and administra- D.C.: Department of Commerce. September 1979), p. 84. tive tasks associated with captivcfnanufactur- ing may be performed elsewhere in the firm, Figure 60.Effort Levels Associated With Product the proportion of production workers is prob- and Process Design for Integrated Circuits ably higher than in merchant manufacturing. 300 As semiconductor production grew, the vacuum tube industry (excluding cathode ray tubes, hence TV picture tubes) declinedfigure 61. While tubes still find specialty applications, by the early 1970's, substitution of semiconductors had caused domestic employment to drop 200 by one-third from the peak level of 1966. Although jobs in tube manufacturing have been E lost to technical change, far more people are ca.' now employed in making semiconductors than were ever employed in making vacuum tubes. Computers 100 Computer manufacturing, like microelectronics, has seen rapid employment growth with simultaneous productivity improvement although, as emphasized in chapter 5, productivity measures can be misleading where

0 "Summary of Trade and Tariff Information: Semiconductors 1960 1965 1970 1975 1980 (U.S. International Trade Commission Publication 841, Control 358 International Competitiveness in Electronics

Figure 61.U.S. Employment in Vacuum Tube Manufacturing (SIC 3671) 80,000

70,000

60.000

50.000 40.000 ' 30.000

\ ;rowam=

20,000 All employees ----Production workers

10.000

1962 1964 1966 ..,1968 1970 1972 1974 1976 1978 1980 1982 SOURCE: Bureau of Labor Statistics. Year the product changes so much. Regardless, ad generally served foreign markets. As in semi- vances in computer systems have created vast conductors, some of this foreign production numbers of jobsnot all in computer manufac- may substitute for exports from the United turing. Many of these new jobs have originated States, but overseas sales are often tied to local in the user community, and in softWare pro- production, limiting the extent to which point- duction. Figure 62 illustrates job growth in the of-sale plants displace domestic jobs. industry itself, including peripherals. Even more so than in microelectronics, the trend has The summary above of employment trends been away from production employees and by sector in the domestic electronics industry toward skillcci workers and white-collar profes- shows that the number of jobs has inzreased, sionals. but not everywhere or uniformly. Increases in Unlike either semiconductors or consumer semiconductors and computers have more electronics, employment in computers and pe- than offsetin magnitudethe declines in con- ripherals has not been greatly affected by im- sumer electronics and vacuum tubes. The com- port penetration or offshore production. Many position of the work force has changed; em- American computer firms have invested over- ployment gains have been greatest for nonpro- Ch. 9Employment Effects 359

Figure 62.U.S. Employment in Computer (and Peripheral Equipment) Manufacturing (SIC 3573)

450.000

400.000 All employees Production workers

350,000

300,000 0

..._ 250,000 0

200,000

150,000

100,000

ea.

. 50,000

0 1955 1960 19651967 1968 1970 1972 1974 1976 1978 1980 1982 Yea, SOURCES: 1955-85-1977 Census of Manufactures. 1968.82 Bureau of Labor Statistics.

Effects of Import Penetration and Offshore Assembly

An increase in imports or a transfer of man- The employment consequences of import ufacturing operations offshore can cut into penetration and offshore assembly are felt in domestic job opportunities. The United States a context of global shifts inmarket structure, is importing more manufactured goods of all implying long-term changes as well as imme- types, not only consumer electronics and semi- diate impacts on people, firms, and industries. conductors, making the import penetration The dynamics are important on both time question especially timely. Moreover, to labor scales. In expanding markets, firms that can unions, offshore production amounts to the ex- respond quickly to new opportunities any- port of jobs. For policymakers, both phenom- where in the world may be able to increase ex- enabut especially imports--have been a ports and consolidate their positions, aided by growing concern. products that take advantage of new technol- 360 InternationalCompetitiveness in Electronics/ ogles. This happened during the 1970's,/when gross domestic product. But what if firms can American semiconductor firms capitalized on only lower their costs and maintain or expand the shift toward metal oxide semiconductor in- their markets by moving offshorewhether to tegrated circuits ahead of their overss rivals. take advantage of low-cost labor and be better Today, Japan's avowed goal of capturing more able to compete with imports, or simply to than 30 percent of the world competer market manufacture their products nearer the ultimate by 1990 (along with 18 percent of the U.S. mar- market? Firms weigh a variety of such factors ket) reflects a belief that longstanding patterns in deciding whether to invest overseas, al- can be disrupted when growth is rapid. though ultimate decisions generally turn on cost savings. From the standpoint of the Na- This section looks more closely at the effects tion as a whole, rather than a particular com- of imports and offshore production on employ- pany, the costs and benefits may be quite dif- ment in consumer electronics and semiconduc- ferent. Appendix B discusses the impacts of tors (neither is important at the moment in offshore manufacturing on the aggregate econ- computers). As pointed out in chapter 5, indus- omy and outlines the range of effects compared tries 'do not rise or decline in competitiveness with alternatives available to the firm. This ap- simultaneously; looking at employment on a pendix includes a case study drawn from the sectoral basis gives only part of the picture, and uxperience of an American company which in then an equivocal one: Still, the sectoral ap- vested in a subsidiary in Taiwan. Briefly, the proach is'a valid starting pOint, for reasons that conclusion of the case illustration is that the are discussed. in some detail in appendix B. offshore plantestablished to assemble auto- The first.question is: What are the causes of mobile radioshelped maintain competitive- import penetration? Imports may rise because ness vis a vis Japanese manufacturers and demand exceeds domestic capacity or consum- prevented even more U.S. :-:bs from eventual- er preference shifts to foreign:made goods. Jap- ly being lost. As this suggests, in consumer anese penetration of U.S.. markets for dynamic electronics the movement offshore by Ameri- random access memories (RAMs) is an exam- can nroducers can be viewed as a defensive ple of the first case, TV imports at least in part reaction to imports. In contrast, the motivation the second (imported automobiles are a more for overseas manufacturing in the semiconduc- obvious example). In the first case, jobs may tor industry has been cost reduction and mar- not be lost because of imports, but the rate of ket expansion driven by domestic competition. increase in domestic job opportunities may The consequences for employment have been slow. In the second case, immediate decreases much different. in employment are likely. The full consequences of import penetration Consumer Electronics depend on the industry. Declining output in Almost half the consumer electronics market some induStriesa prominent recent instance in the United States has been taken by imports; again being automobilescan have major spill- in addition, many products assembled here de- over effects elsewhere in the economy. As sales pend heavily on imported components and of domestic cars lagged, jobs were lost in firms subassemblies. Penetration of consumer elec- making steel, tires, and components. Some- tronics markets has coincided with employ- times companies can limit impacts on individ- ment decline, as shown in figures 57 and 58. uals by allowing employment todecline Imports of black-and-white TVs rosé from one- through attrition rather than layoffs; even so, quarter to three-quarters of U.S. sales over the the overall pool of job opportunities shrinks. period 1967-77. Color TV imports peaked in The effects of offshore production are no 1976 at a level nearly tenfold greater than in morn strniehtforward. On the nne hand. all 1967. then dropped because of OMAs. A thira Ch. 9Employment Effects 361

. Today, all U.S. TV manufacturers operate of eventsimport competition must indeed be foreign production facilities. In addition to the counted as the primary cause of job losses in attraction of low-wage labor, Items 806.30 and consumer electronics.'" 807.90 of the U.S. tariff schedules encourage But this is not the whole story: Is it possible offshore assembly (ch. 11). During thelast half that the ready availability of Item 807.00 re- of the 1970's, 30 to 45 percent of all color TV duced incentives for American managers to cut imports'entered under Item 807, although final costs and improve labor productivity at home? assembly remains concentrated herein part Might U.S. firms have avoided offshore pro- because of foreign investments to avoid the duction by adopting more capital-intensive 0A-i mposed quotas' automated manufacturing processes here? Jobs Despite limits on imports, employment in TV still might have been lost, but the costs and manufacturing did not recover. In testimony benefits would have shifted. The behavior of before the U.S. International Trade Commis- American executives is often contrasted with sion, the International Brotherhood of Elec- that of their Japanese counterparts, who recent- trical Workers reported that 20,000 workers ly have faced similar difficultiesi.e., competi- had lost their jobs in the TV industry due to tion frocountries with much lower labor imports.lo To what extent are imports to blame, costs. Some observers have claimed that given that domestic productivity improvements Japanese consumer electronics firms have in- and offshore investments by U.S. firms have vested more rapidly and more boldly in, mech- also contributed to employment decline? anized production technologies such as automatic component insertion (see ch. 6 for a fur- It is oversimple to argue that the total num-. ther discussion of rates of adoption of automa- ber of foreign workers engaged in.production tion). for shipment to the United Stateswhether employed by U.S. or foreign firmsrepresents do- As with many such questions, the truth prob- mestic employment loss. In most cases, U.S. ably lies somewhere between. The availability consumer electronics firms had little choice of Item 807 reduces the pressure to find cheap- concerning offshore production. Movement er manufacturing methods at home. It is also abroad was a defensive reaction, not a strategy true that the Japanese, when themselves con- eimed at expanding markets and improving fronted with the rather sudden emergence of profitability. To assume that jobs overseas competition from other Far Eastern countries, substitute directly for U.S. employment is tan- transferred some of their production to lower tamount to assuming a stable competitive en- wage sites. In part, such transfers were miso vironmentwhich was not the case. Rather, caused by the 1977 OMAwhich, bye limiting employment declines followed losses in com- shipments from Japan, created incentives for iitiveness; American firms had higher costs Japanese firms to move to export platforms than their rivals, and little scope for develop- but Japanese managers exhibit little rehictance ing strategies that would preserve domestic to take whatever steps seem necessary for jobs. They pursued the. ,obvious route: in- preserving hard-earned market positions. creased .automation. to raise productivity at-Along -with__ developing new __production_ hoMe. combined with transfers__,. of labor-inten- methodsan uncertain businessJapanese sive operations offshore. Only some companies firms would probably have shifted labor- survived; the other were purchased by more intensive production abroad in any case, sim- successful manufacturt-,s or left the industry. ply for insurance. In this respect. Japanese In this senseas part of a more complex chain managers have behaved much like Americans.

mFestimony lietbre the U.S. International Trade Commission r.11 P...... it,treIT A:1(11-1C11.'Intornatinnal RrnthPrhnnd of Elpr- "For a generally contrary view, see. A. 0. Krueger. "Restruc- ' Competitiveness in Electronics

rapidly. Furthermore, American semiconductor manufacturers have exported much mole actively than consumer electronics firms. Figure 59 showed the steadily growing employment in the U.S. semiconductor industry. Domestic jobs more than doubled during the 1970's; offshore employment probably expanded even faster. The question again is: Do imports, or foreign workers employed in the overseas operations of U.S. firms, stand for job opportunities lost to Americans? Imports and offshore manufacturing are more closely coupled for semiconductors than for consumer electronics. Nearly 40 percent of U.S. semiconductor sales are classed as imports, but more than three-quarters of these are re-im- Photo credit: Control Automation ports by American firms under Items 806 and 807 of the tariff schedules. Offshore produc- assembling electronic components tion is central' to the U.S. industry. Still, shipments from Japan have also risen swiftly nployment would have dropped,, over the last5years. ithout OMAs, the lesson-7-re- The offshore facilities of U.S. semiconductor . industriesis that controlling manufacturers concentrate on the labor-inten- is to provide companies a res- sive steps in the production process primarily ich they can take measures to assembly. In the mid-1970's, Finan estimated ompetitiveness is unlikely to im- that manufacturing costs for integrated circuits lent prospects. Indeed, jus:: the could be cutlin half through offshore assem- to be true, as manufacturers bly.18Cost/p9ce competition has thus been the -. sts by improving labor produc- primary motive for foreign °investments; Amer- :tent\ that they succeed, employ - ican semiconductor firms moved offshore to will even in situations reduce costs rid expand markets. Moreover, ;rowth in output takes place. As the competiti9n has been largely among domes- notection seldom functions as tic firms; investments predate Japanese com- assistance to displaced workers. petition by a decade and more. If in the case of consumer electronics, offshore manufacturing was a reaction to import competition; in etniconductors semiconductors the primary motivations were offensive. Capital investment requirements of semiconductor devices have have.been_one_olthe forces_at_work. In order . lily, exceeding imports by 1982 to keep up with demand, semiconductor firms -1 earlier years, the United States have been unde\r, continnal pressure to add new more semiconductors than it capacity (ch. 7). Offshore assembly offered flex- uch trends portend job losses?

oint, with Japanese manufac- 1 taw. F. Finan, "The International Transfer of Semiconductor haif or more of the burgeoning Technology Through U.S.-Based Firms," Working Paper No. 118, :et, will employment in this por- National Bureau of Economic Research, December 1975, p.60. The savings are greater for simpler integrated circuits and dis- ronics industry suffer as in con- . . . . Ch. 9Employment Effects 363

ibility; firms could avoid the risks of investment. Japanese investment in the United States ments in automated equipment that might soon has grown rapidly, from a cumulative $152 mil- be outdated, expanding capacity without tak- lion in 1973 to $4.2 billion by 1980.19 The desire ing funds from capital-intensive wafer fabrica- to open new markets and to ease trade frictions tion and testing equipment. are among the forces behind this influx. Japa- nese-owned firms now assemble nearly 4 mil- What are the implications for job opportunities? As the case study in appendix B illus- lion color TVs here each year. North American trates, these depend in part on the time hori- Philips adds well over a million. zons. Given rising foreign competitiveness in As this suggests, most of the past investments microelectronids, offshore production now in electronics have been limited to consumer helps meet international as well as intranation- products. Japanese interests seem bound to al competition.If overseas manufacturing widen, however, with plants for assembling in- helps U.S. firms maintain their Competitive- tegrated circuits the next step. In typical ness, the net impact on domestic employment foreign-owned manufacturing plants, only a might he positive over the longer term. Further- few ur--;er management slots are reserved for more, point-of-sale plants are sometimes able exec from headquarters. Viewed strict- to sell in markets to which the U.S. parent ly froz:1 an employment perspective, therefore, would have difficulty in exporting because of onshore manufacturing has positive conse- trade barriers. In some instances at least, quences for the United States. Viewed more American firms may thus be able to strengthen broadly, the picture becomes mixed: many of their long-term competitive -.position by in-. theskilled and professional jobs remain vesting overseas, enlarging domestic as well overseas. as foreign employment. Still, in the short term, Generalizations about employment that offshore investments, cut the number of job op- would apply to all parts of the electronics in- portunities for Americans. In this respect, dustry are impossible. In the case of consumer questions of the impact of Items 806 anc1,807 electronics, import penetration is closely asso- of the tariff schedules are similar to the more ciated with job loss. In contrast, employment general problemisolating the consequences has grown steadily in both domestic and for- of foreign direct investment of any type on eign operations of U.S. semiconductor firms; employment. Such matters have been investi- overseas investments have helped cut costs, ex- gated extensively over the years. The most pand markets, and increase competitiveness. common conclusion is that direct investment Simply in terms of numbers of jobs, expansion by American corporations has increased net in semiconductors and computers has more employment in the United States: nevertheless, than offset declines iconsumer electronics. the opposite result is sometimes reached, again This does not mean, of course, that such trends depending on the particulars. In the im- will persist indefinitBly. Nor is it any consola- plicationsboth short and long termcan only tion to people who find themselves out of work. be evaluated on a case-by-case basis, anc, de- The rest of the chapter looks to the future. : Tend on assumptions concerning the future _ .competitive environment for American firms. Foreign Investment in the United States

Foreign investments here bring yet another iglapanese Manufacturing Operations in the United States." dimension to questions of domestic employ- Japan External Trade Organization. September 1981. 364 International Competitiveness in Electronics _ . Projections of Employment Within the Electronics Industry

Before examining impacts on other parts of percent annually (the "low trend") to 3.8 per- the economy, this section treats the industry cent (the "high trend"). These assumptions itself,in the context of Bureau of Labor compare with a 197379 average of 2.8 percent Statistics (13LS) employment projections to per year. BLS has assumed growth in labor pro- 1990. The perspective is much broader than the ductivity to stabilize at,the rather low levels of discussion of possible shortages of engineers recent years.21 .and skilled workers in the previous chapter. On this basis, BLS predicts that aggregate Ideally, projections of future trends would growth in U.S. employment will range from 1.6 be based, not only on a model for aggregate to 2.0 percent annually over the decade of the economic expansion, but also on sector-spe- 1980's, considerably below the 2.7 percent c:ific variablesgrowth in particular product yearly rise for 1975-79. Women will get two- markets, demand for workers with certain thirds of the new jobs. The durable goods por- kinds of skills, levels of imports and exports... tion of manufacturing is expected to grow Unfortunately, this much detail is seldom at- faster than the all-industries average, non- tempted. BLS projections, virtually the only urables slower. analyses available with industry-specific output, are based on an econometric modela Output increases in computers and related limited tool, although representative of the state equipment should lead all other manufacturing industries; employment in the computer inof the art.20 dustry will grow from about 420,000 in 1982 BLS began making econometrically based to perhaps 600,000 by the end of the decade. employment ()injections two decades ago, in- If these projections prove realistic, employ- troducing a macroeconomic demand model in ment in the computer and peripherals sector 1975. Their current procedure includes five will comprise as much as 3.1 percent of the basic steps: 1) pr,7!;ctions for the economy in total manufacturing work force by 1990, com- the aggregate; .2) disaggregation of GNP by de- pared to 1.6 percent at the end of the last mand categories; 3) distribution of demand by decade. Employment in the electronic compo- categories to producing industries; 4) output nents sector (SIC 367) is expected to grow at projections by industry sector based on an about 2.2 percent per year in both low- and input-output table; and 5) forecasts of labor pro- high-growth scenarios, well above projections ductivity, total labor hours, and.number of peo- for manufacturing as a whole. In the low- ple employed at the sectoral or industry level. growth scenario, 33 of the 150 industries ex- A critical input in terms of employment is the amined show employment drops. One of these estimated gross demand for the products of an is radio and TV manufacturing, with an an- .gross.. output is divided.. by an ticipated decline_averaging 1.A_percent per year estimated productivity level (output per em- over the period 1979-90. Thus, if BLS projec- ployee-hour) to yield the labor hour projection tions prove realistic, past employment' trends for the industry, and thus employment. The in electronics will persist: there will be con- model as a whale is sensitive to a wide range tinuing decline in consumer electronics, rapid of assumptions, most fundamentally those for growth in computer manufacturing, and con- GNP growth. BLS's recent projections have been based on GNP increases ranging from 2.4 "For more detail, see V. A. Personick. The Outlook for In- Ch. 9Employment Effects 365 siderable expansion in components. Together, these three portions of electronics might, under the most favorable circumstances, account for more than 7 percent of U.S. manufacturing employment in 1990. The projections are all con- ditional, needless to say, and BLS's approach shares the principal limitation of virtually all forecasting techniques: current trends are expected to continue, breaks with the past seldom anticipated. BLS also estimates employment by occupational category across industries; inall scenarios, white-collar jobs grow faster than total employment, blue-collar jobs slower. White-collar workers will make up slightly more than half the 1990 labor forcethe frac- Photo credit: Western Electric Co. tion is slightly less nowwith notable increases Semiconductor wafers being loaded into furnace in the professional and technical category.22 0 Table 75 lists occupations in electronics for As shown earlier, the electronics industry ex- which BLS predicts the greatest percentage in- perienced a more-or-less gradual shift toward crease during the 1980's. All are grey- or white- fewer production workers and more white- cellar jobs. The nonelectronics categories are collar workers during the 1970's, with the big- included for comparison; 5 of the 10 fastest gest change in semiconductor manufaocturing growing occupations in the complete BLS (fig. 59). Table 76 gives occupational break- listing are electronics-related. Despite the hit.,h downs in consumer electronics,components, growth rates, categories starting from a modest and computers according to BLS data for 1980. base will not account for large numbers of new While BLS expects some further upskilling dur- jobs. ing the 1980's, the projections (not shown) ==M. L. Carey; "Occupational Employment Growth Through which may or may not be well-foundedindi- 1990," Monthly Labor Review. August 1981, p. 45. cate these to be mostly matters .of a percentage point or two.Note that the SIC categories in table 76 are broader than used earlier; the Table 75.Predicted Growth Rates by consumer electronics data cover SIC 365,rath- Occupational Category Over the 1980's er than the "home entertainment"subdivision, Predicted increase 3651; electronic components, SIC 367, includes in employment all types of components, not just microelectron- Occupationa (1980-90) ics; and the computer category referred to earli- Paralegal. (9/3°/D er, 3573. is a subdivision of SIC 357.Moving Data processing machine mechanic Computer operator 72 the boundaries of these categories outward Computer systems analyst 68 probably makes little difference for consumer Business machine se,iice technician 60 electronics and computers, but components as Computer programer 49 Employment interviewer 47 a whole are not nearly a,skill-intensive as Computer peripheral operator 44 microelectronics; thus the iportions of tech- Psychiatric aide 40 nical professionals in table, areconsiderable allonInclusive.stest growing occupations in electronics are listed together with selectea occt.;:ations outside of electronics (in italics) for comparison. underrepresentations for semiconductor firms 366 International Competitiveness in Electronics

Table 76.-Occupational Distributions in Electronics as of 1980

Consumer Electronic electronics components Computers (SIC 365) (SIC 367) (SIC 357) White. and grepcoilar workers 27.9% 32.0% 59.0% Professional alid technical: Engineers (and scientists) 3.6 6.2 11.3 Engineering technicians 2.8 6.1 9.3 Computer specialists 0.6 0.6 6.2 Other 3.0 2.8 5.7 Managers 4.7 5.4 9.4 Salesworkers 0.7 0.7 1.0 Clerical 12.5 10.2 16.1 Bluecollar workers 62.1% 63.6% 38.2% Craft 17.3 14.8 12.1 Assemblers and machine operators 44.8 48.8 26.1 Service workers and others 10.0% 4.3% 2.9% SOURCE Bureau of Labor Statistics. in the ranks of white-collar workers (not all the number of professional and technical with high levels of educationor training) or workers will grow from 87,700 in 1980 to over skilled, grey-collar technicians-in contrast to 117,000 in 1990. In the more mature consumer consumer electronics and components, where electronics industry, the absolute number of these jobs make up less than a third of the total. blue-collar workers is likely to decline, as well Employment expansion in computers willcon- as the proportion. Taken together, the trends tinue to be most rapid in skilled categories indicate a continued shift toward more highly (table 75); numbers of service and repair techni- skilled jobs in electronics. Computer manufac- cians and systems operators will increase, turing, in particular, will be a leader in employ- while jobs for keypunch operators-whose ment growth and in demand for new.skills over skills are becoming obsolete-will dWindle,as the next decade; the picture for this industry will work for those without special training. foreshadows trends expected elsewhere in the Likewise, in components, BLS estimates that U.S. economy.

Future Employment Patterns inOther Industries

If analysis of past trends in electronics is fined occupational categories. This isexpen- problematic, looking ahead to the impacts of sive and time-consuming, demandinga sophis- electronics on other industries isa still more ticated appreciation of how industryuses tech- tenuous exercise. Yet it is a vital one, for future nology; in consequence, such studiesare sel- developments in electronics have far-reaching dom attempted. implications for the entire economy. Useful Uncertainties abound. First. past trends- policy guidance could flow froman under- including examples of technical change in in- standing of how technological change affects dustries other than electronics-can offer only employment patterns. Public and private train- a general guide; there are no guarantees that ing and retraining programs would benefit if current employment patterns-outcomes of vulnerable job categories, as wellas those for large numbers of incremental and evolutionary which demand will rise, could bemore reliably chances -will cersist. Second. many imnats Ch. 9Employment Effects 367 lug to take one examplemay increase em- change is slow and past trends supply prece- ployment in firms designing and building the dents, the unexpected consequences of new equipment used, decrease employment in the technologies escape forecasting methodologies apparel industry, but perhaps have positive im- virtually by definition. Examples from the past pacts on employment at the retail level (one illustrate little beyond the seemingrandomness reason might be that custom design would be- of the impacts of technological change. This come cheaper, with Smaller runs of styles and in itself is an important,lesson, but means that sizes sold in specialty shops). Attempting to the state of the art is such that even well- trace such second and third level effects in- documented historical case studies can seldom volves the interplay of business decisions, eco- provide direct policy guidance. nomic and product cycles, imports and ex- The basic problem is that, even if it were pOrtsnot to mention the unpredictable mature possible to predict how technical change in of consumer demand. The following sections electronics would affect some other industry, do not attempt to answer the!" question of there is no necessary relationship between whether electronicS technologies will have net these findings and the consequences for the positive or negative impacts on U.S. job oppor- economy as a whole. Building up the picture tunities, but simply illustrate some of the forces on a detailed, sector-by-sector basis would be at work. a vast undertaking. MostsTof the past attempts European governments, sensitive to the po- whether dealing with manufacturing or serv- tentially negative employment consequences ices or bothhave been more limited, falling of electronics and automation, have commis- into one of two categories: 1) elaborate but sioned numerous reports on the subject, with abstractanalytical frameworks, typically_ uniformly disappointing results. Micro-level econometric; or 2) case studies outlining im- analyses exploring impacts on a particular craft pacts on particular sectors. The first,ex- or industry are difficult to integrate with emplified by the BLS analyses. discussed macro-level studies and aggregate economic earlier, have seldom been very illuminating in forecasts. Yet this couplingthe complex and terms of real-world experience. The second evolving interplay among technical advance, often yield insights that are useful but limited utilization within various economic sectors, to relatively narrow segments of the labor and the response of the labor marketis critical forcebank tellers, coal miners, postal work- on both supply and demand sides. For exam- ersas illustrated by the case examples that ple, companies typically install labor-saving follow. equipment in periods of economic expansibn, when workers can he transferred to other jobs Manufacturing rather ths,i laid off. Over the longer term, then, a given firm can often use normal attrition to Many of the studies addressing manufactur- help manage the size of its work force. Where ing begin by distinguishing between product this is the case, direct attribution of decreases and process applications. These overlap in the or increases in employment opportunities to sense that computer-based process control sys, technology can be difficult to defend.23 terns, to take one example, can be viewed in eithei light. As a "product," they are developed While forecasting methods do a reasonable and sold by firms in the capital goods industry. job of predicting employment within either ag- In the alternate view, automated process con- gregate or disaggregate categories as long as trol is one aspect of an ongoing transformation "The authors of a British study write: -Microelectronics tech- of production in many industries. Employment nnlnuy will affect manufacturing industry in so many ways that rtnnfre-siirsvA, ini brall xyiraIA,Cnitbr mob hrrti- 366. International Competitiveness in Electronics

by market.24 In theory, the greatest gains come equipment; within the system,fev,eastaers, where new products are introduced into new service persOnn'el, and operators a needed. markets. Pocket calculators and video games Further reductions may be in stcre, with fuily are c\amples. While they may replace other elect -onic equipmentexpected around\1990 goodselectromechanical calculators, for ex- cutting the work force to as littleas oneAnth ampleto the extent that new products expand its former size.25 markets or create new ones, employment will rise. Evi Sting products introduced into new Printing markets have parallel effects. The "personal" computer is not a new technology or a new Computerized typesetting provides a second product so much as an ad 'ption of microproc- example of the introduction of new products essor-based data processing systems to the into old markets. High-speed photo-typesetting needs of individual households and small busi- equipment, along with typesetting computers, nesses. Low-end minicomputers of the early have transformed the printing industry. The 19711'5, such as the PDP-8, were similar in many equipment is much less labor-intensive than the respects to current personal computers, but the hot-metal typesetters that have been replaced, PDP -8 was never marketed as such. In contrast, and productivity has jumped. With electronic the introduction of new or replacement tech- typesetting, an operator selects type size and nologies into old markets often cuts into job stte, column width, spacing, and other layout opportunities. Recent and well-publicized illus- specifications on a video screen, composing an trations include electronic switching in tele7 entire page at a time; the older linotype ma- communicationsprincipally telephone sys- chines, stemming from the end of the 19th cen- tems=and electronic typesetting'in the print- tury, prrduced one line of type at a time. After ing industry. In essence, these technologies electronic photocomposition had been intro- caused step changes in labor productivity, with duced at the New York Times, the Sunday subsequent employment declines. In such cla-,.sified section could be completed in 20 cases, output may expand, but not rapidly minutes rather than 3 days. Over the mid- enough, to compensate. In between the ex- 1970's, the staff in the .s.nmposing room de- tremes of the examples above fall many which clined from 830 to 685 employees, and would have mOre moderate impacts on employment. have dropped much further except for the ability of the printer's union to maintain many jobs Several case studies are outlined below, in- that were in fact redundant.28 (Of course, if one cluding those of telecommunications and type- looks at media as a whole, electronics has setting, to illustrate typical impacts of elec- created vast numbers of jobs.) tronics-related technologies on employment patterns. Unfortunately, while productivity is now much higher, demand for books and .news- The British Telecommunications Industry papers has not changed much. Between the mid-1960'swhen only about 2 percent of all The introduction of electronic switching in typesetting in the United States was performed the British telephone system exemplifies the replitcement case. Employment dropped from "M. Wilkinson, ''System X: The Need to Shake-up the :hone- 90,000 in 1973 to 65,000 by the end of the makers.' "Financial Times,Oct. 18, 1078. "The union a2 theTimeswas more successful than most at decade, Jobs. were lost both in manufacturing holding on to jobs for its members. For a detailed treatment of and among those employed running the sys- this case, see "The Impacts of Robotics on the Workplace and tem. Declining export sales contributed to job Workforce," Carnegie-Mellon University, Schou] of Urban and Public Affairs, June 14, 1981, pp. 35fl, Othar examples of applica- loss in the manufacture of telecommunications tions of electronics technologies in printing' den be found in J R. Werner, "The Role of Electronics in The Modern News-. . FelirnvIng NI.' ynti H. Rush, "The Impact of Micro- paper," and J. L. Boyd, R. E. Robey, and J. S. Richards, "Auto- electronics on the U.K.: A Suggested Classification and Il- mating Newspaper Production," sess. 21, The Role of Electronics lustrative Case Studies." Occasional Papers Series, No. 7, Science in the Graphic Arts, 1979 Electro Professional Program, New' Policy Research Unit, Universin, of Sussex, June 1978. . York, Apr. 24.26, 1979.

C. Ch. 9Employment Effects 369 by the new machinesand the end of the Computer-Aided Manufacturing and Design 1970's, penetration rose until about 90 percent of all newspapers were composed using com- Continuing integration of computer technol- puterized equipment..The impacts on printers ogy into manufacturing operationscomputer- as craft workers have been severe. Not only are aided manufacturing (CAM)will eventually fewer people needed for photocompoSition, but have major consequencesfer erciployrnent (ch. they must have different skills. Few printers 6). Nonetheless, such developrneitsincluding have found jobs as computer programers or robots and software-programmable a 1..fiornated service personnel. Unions have been less con- equipment of many kindsshould generally be cerned with the total number of job opportun- viewed as evolutionary steps in the .automation ities than with protecting individuals. Work of the workplace, continuing down paths orig- forces have been reduced through attrition; the inating many years ago. Much the same is true pension system created incentives for early of computer-aided engineering design (CAD), retirement. Printers, proud of their traditional which consists in p-art of automating tasks-=' craft skills, were not very receptive to retrain- ranging from drafting to numerical analysis ing, although this had always been a central formerly done manually. In addition, both part of the uni,o's philosophy. While the strat- CAM and CAD make possible work that could egies adopted h'y organized labor when con= not be performed at all in earlier years. Ex- fronted with such t c oblems have varied, the amples include machining parts without the example of the printing industry is not untyp- aid of drawings,-continuous balancing of rotors ical of instances where replacement technol- with material removed by lasers, or finite-ele- ogies have been introduced into existing mar- ment analyses of stresses and deflections. kets; labor-management relations tend to be As computers spread through manufactur- critical factors in coping with job-displi,, rnt ing, impacts on employment will be, at least effects. at first, incremental and random-si eming. In the longer run, productivity will be greatly im- Electronic Watches proved; labor-intensity will drop, and large In the watchmaking industry, an example numbers of manufacturing jobs will disappear, from consumer goods manufacturing, elec- particularly those with lower skill levels. In this tronically based products took more than half sense, the long-term effects will in fact be revo- the total market within the space of a decade. lutionary. The work force will face continuing In Switzerland alone, 20 to 30 percent of ex- structural shifts, and labor-management rela- isting assembly labor was displaced.27 Skill re- tions will be under strain as accommodations quirements for assembling electrOnic watchs are sought:Changes in employment patterns are negligible. Along with deskilling of the pro- inagiven industry will depend on the duction work force, international shifts oc- characteristicproduction processeshow curred as firms in the Far East took over mar- susceptible they are to automationas well as kets for lower priced watches; most of the rel- growth in markets and shifts in competitive. atively simpleintegratedcircuits needed ness. Computers will have their greatest irn: are also made in Asia. Managements of Swiss pacts when accompanied by large -scale reor- watchmakers reluctant to switch to the new ganization of the work place,, as happened in technology found their firms rapidly losing continuous process industries with the in- ground, with effects on employment that were troduction of computerized process control. even more devastating than among manufacturers choosing to embrace electronics. Numerically Controlled Machine. Tools The diffusion of numerically controlled (NC) (ch. 6) machine tools illustrates the results of ,'Technical Change and Frnignyment. ,op. cit., p. 136. incremental improvement in manufacturing

3 7e) 370 international Competitiveness in Electronics

technology. A survey of 24 American firms Are the Case Studies Typical? revealed comparatively limited impacts on employment.28 NC machines were generally None of these examples can be taken as rep- purchased when business was good and out- resentative. They are anecdotal accounts of put expanding; the new equipment helped events that have followed the introduction of firms produce more without hiring extra electronics-relatedtechnologies. Technical workers. Nor were many employees displaced; change generally proceeds in piecemeal most moved on to other production jobs, fashion, with pace and impact that vary from although skilled craftsmen sometimes found case to case; given hindsight, of course, such the transition to NC machines difficult. Man- seemingly random and incremental events may agement also had to learn to operate in a new show patterns invisible at the time. environment. Overall employment remained In the examples recounted, the jobs created more-or-less static, but the skill mix changed generally called for different skills. Typical new and some individuals were faced with entire- openings were for computer operators, or serv- ly new jobs. If the impacts of NC machine tools ice and repair personnel trained to work on the have been mild, it would be misleading to latest generation of equipment. While patterns generalize this to future developments in of job loss and job creation vary across indus- CAD/CAM. NC machining is a major step in tries, production jobsunskilled, semiskilled, metal cutting, but a much more modest devel- or skilleddisappeared in all cases except NC opment from the viewpoint of manufacturing machining. Future employment impacts will technology as a whole; the next two or three be influenced, not only by the technology itself, decades of advances in CAD/CAM will bring but by the general state of the economy at the more radical change to the factory floor. time new technologies are introduced, by the attitudes of workers and unions to automation, Pet Foods and by the choices of corporate managers. In In an example of automated process control, some cases, job losses will be mitigated by ex- a British firm with a large share of the pet food panding markets, particularly if workers are market invested in a computer- controlled pro- retrained. Overall, however, a shrinking work duction system.29 Instituted with the goal of ra- force in manufacturing points to continuing tionalizing the production process, the system displacement and adjustment problems. was expe,cted.to cut employment by three-quar- teis over a 5-year period. The proportion of unskilled production workers dropped precip- Services itously, while more management and engineer- The service sector has been growing more ing personnel were needed. An absence of rapidly than manufacturing. Can the U.S. unions, combined with an extensiveimpaign eamoniy continue to generate new jobs in serve to convince workers that the new c!uipment ices ai a high rate? Office work has been a ma- woulrl. eliminate the least desirable jobs, appear jor source of past expansion. With the elec- to h,: been critical factors in the acceptance tronic office on the horizon, will this source of the new equipment. The small group of dry up? If office automation begins to cut deep- workers selected by Management to run this ly into employment opportunities, the ability 'equipment expressed considerable satisfaction of the service sector to compensate for losses with their greater responsibilities. The rest of in manufacturing will be seriously impaired. the production work force lost their jobs. 2 R. T. Lund, et al., "Numerically Controlled Machine Tools Office Automation and Group Technology: A Study of U.S. Experiences:" Keport CPA 78-2, Center for Policy Alternatives, Massachusetts Institute Fortunately, this seems unlikelyat least in of Tethnologk, Jan. 13, 1978. the near term. Office work, breeding ground K . Dickson, "Petfoods by Computer: A Case Study of Automa- tion," ni TheMicroelectronicsRevolution. T. Forester (ed.) (Cam- or Parkinson's Laws, will probably continue bridge, Mass.: MIT Press, 1981). to expand. At least some white-collar jobs seem Ch. 9Employment Effects 371

relatively impervious to automation in the sense that people can find other things to occupy their time. This is partly a consequence of the lack of c: :.-put indicators or other measures of white - collar productivity.Nevertheless, beyond office wort:, electronics may v: du:.:e job opportunities (or the rate of job creation) in sectors like !nsportation, retaling, banking, and the postal system. Less is known about the effects of automation in services than in manufacturing. Over the past two decades, the xerographic copier Photo crsclit: Wang Laboratories has probably had greater impacts on office Word processing: one of the early steps work than any other piece of technology, yet in office automation these seem hardly to have been studied. Has the office copier created jobs? What have been the effects on organizational efficiency? No one Examining occupational categories makes it seems to know. It has saved so much drudgery, clear that the mode of utilization of the new however, that few are likely to care. technology is just as important as the speed of adoption, again as in manufacturing. If new More concretely, studies of the application technology is instituted primarily as a substi- of electronics to services generally findnot tute for narrowly defined functions such as in- revolutionary change, but gradual evolution ventory recording, bank telling, or filing, best viewed as an extension of computer ap- employment is likely to drop over the longer plications already in place. Such studies em- term unless jobs expand in other areas (such phasize the extent to which workers such as as sales). Where computers facilitate more ef- typists or clerks whose job skills may become fective and extensive information processing, obSolete can be redeployed, seeing, for exam- new jobs may be generated. Where demand for ple, word processing as a straightforward ex- new types of services is created,employment tension of typing. will rise. The central features of the electronic office Consider the proliferation of word process- expanded applications of data processing ing equipment, which affects the tasks now equipment, including communications and performed by a well-defined group of employ- word processinghave thus far been intro- ees. Based on the results of work-measurement, duced into existing or conventional office en- tests conducted in organizations that have vironments. In this respect, the analogy with switched from typewriters to word processors, NC machines and industrial robots is close. productivity often more than doubles. While While office automation promises to reduce this might suggest that half the typing work staffing needs in conventional jobs, new tasks force faces unemployment, in practice nothing are at the same time created in operating and like this has happened. Indeed, some firms maintaining the systems, as well as using them. have invested in word processors in response Since office work is seldom very efficient or a shortage of typists. In other. instar..-es, well-organized, computerization is likely to . -here typists have been made redundant they. have its first effects at the margins of these have moved to other parts of the organization. 1,e ople-centered activities, rather than leading In many cases, people just write more words, to sudden and major shiftsWholesale reorga- nizations of the workplace will be slower than In fact, since word processors make it easier in manufacturing. to produce multiple revisions of the same docu- 372 International Competitiveness in Electronics

tricot, productivity cannot be measured simply ployed in banking in the United States grew in terms of words or drafts typed. The charac- by half, confounding predictions of employ- teristics of the technology lead directly to an ment losses.° Two interpretations are possi- increase lie number of slightly modified ver- ble: the first is that growth in job opportunities sions prk..,to a final copy, whether this is a slows under these circumstances; the second, short letter or a report hundreds of pages in that electronics allows banks to expand their length. While the benefits of this may be ques- functions in ways that would otherwise be pre- tioned: the point is that oversimple estimates cluded. These interpretations are not mutual- of productivity gainsachieved or potFmtial ly exclusive; both have some validity. Clearly, overstate the probable employment conse- electronics technology has modified and ex- quences. tended banking functionsan obvious exam- Eventually, offices will be structured in sub- ple is the automated 24-hcur teller. Nonethe- stantially different ways. Some jobs will be less, in Europe, employment growth in bank- eliminated,others modified.Interactions ing and insurance has already begun to slow; among people, individually and in groups, will a well-known report to the French Government change. Matters of timing and approach to the predicts that one-third of all jobs in banking inf!allatiOn of new office equipment will, as in ad insurance might he eliminated over the the case of factory work, affect employee sup- uecade ahead.31 While perhaps overly dra- port or resistance, thus the effectiveness with matic, such predictions point to the concern which the equipment is utilized, and people's these issues have 'aroused, particularly in satisfaction with their work. Western Europe. Like electronic funds transfer, electronic Other Services mail has been viewed with some apprehension. Service sector jobs outside the office include The U.S. Postal Service has made notable health tare, retailing and selling of all types, strides in productivity over the past decade, banking, transportation, and postal services even reduced its labor forcebut seldom as a (more broadly, communications). In principle, result of electronics. In the future, electronic electronics could alter many of these, but mail may cut deeply into job opportunities for where and when -or whether is another mat- postal workers; employment could drop by 20 ter. to 25 percent over the next 20 years.32 In banking, computer processing of mag- netically encoded checks has made it possible 3°See j. Ilenize, -Evaluating the Employment Impact of infor the sanie number of employees to handle formation Tachnology," Technological Forecasting and Social an ever-growing volume of transactions. Change, vol. 20, August 1981, p. 41. tronic funds transfer remains costly, and thus ,'Microelectronics at Work: Productivity and lobs in the World Economy, cp. ci'., pp. 36-37. The French report is S. Nora and far has seen only limited use in retail banking; A. Minc, The Computerization of Society (Cambridge, Mass: applications to interbank transactions have MIT Press, 1980). been much more prominent. Such '2Implications of Electronic Mad and Message Systems for. U.S. Postal .Airvice (Washington, D.C.: U.S. Congress, Off ice of menu have not led to work force redu0'.; as; Technology Assessment, OTACIT-183, August 1982), ch. 6. Thu during the 1970's, the number of people em- postal service employed nearly 700,000 people in 1980. Suriimary and Conclusions Examples from both manufacturing and impacts resulting from technical change in services can be interpreted either optimistically electronics on the economy as a whole cannot or pessimistically; in the absence of more sys- be answered. But regardless of the view one tematic studies, the question cf employment takes, unprecedented adjustments lie ahead for Ch. 9Employment Effects 373

both individuals and firms. Should aggregate that large numbers of Americans find them- employment increase. the introduction of new selves without work liecause they lack the technology will alter the jobs that people do capabilities that this and other industries de- and change their interactions with one another pend on. Should total employment increase only slow- Like all technical change, advances in elec- ly compared to growth in the labor pool, or tronics will bring a mix of positive and negative decrease, the adjUstment problems will be ex- effects; at present, there is little factual basis traordinarily severe, more so in a country like for either an optimistic or a pessimistic view the United. States which has little experience of the longer run impacts. Firms manufactur- with manpower policies, and where many peo- ing:electronics products will, for some years, ple have come to view adjustment assistance continue to create substantial numbers of new as a fallure. jobs. In U.S. manufacturing as a whole, the rate In recent years, the number of new job op- of growth of job opportunities has already portunities generated by the U.S. economy has slowed, and jobs may gc down in absolute slo-ved. A good deal of the future expansion terms. A major source of declines will be com- will he in computer-related fields,,nly those puter-assisted automation. Will job growth else- with appropriate training and skills be in where compensate? Anticipating events in the position toke advantage of these oppor- service sector, where productivity growth has ;unities. Upskilling in the {:omputer industry been low, is more problematic than in manu- hat; been going on for as indicated by facturing. While there may be only a few cases theincreasing' proport!A.;.qof white-collar of employment impacts as severe as in news-. employees, compared to production workers. paper printing, there will be a multitude of ad- In fact, the white collar-blue collar distinction justment problems for individuals; these are no longer carries much meaning; the labor likely to accelerate as electronics technology force is becoming increasingly stratified. Dis- continues to permeate both manufacturing and tinguishing those with specialized skills from serees. In the end, much t. ill depend on over- those without is only a starting place for ex- all rates of economic growth. amining the many new gradations. job opportunities also depend on comp.r.ti- -A common notion, for example, is that com- tiveness. Employment, typically falls when in- puters will bring "user friendliness" to many dustries lose ground in either domestic or L.- jobs so that unskilled workers can perform ternational markets. Even if aggregate eco- them. This is potentially misleading. User- nornie growth brings greater demand, only the endliness permits people with good skills to more efficient companies can take full advan- ork with complex and sophisticated systems tage. Generally speaking, firms and industries that otherwise would demand highly special- that make effective u-se of new technologies expertise. User-friendliness also tends to will generate new jobs, or if jobs are lost, this :hangi. the abil required in the labor force. will come more slowly; indeed, companies. Efficient utilization of a word prOcessor de- seldom have any choice but to adopt new tech- pends en different skills than manual typing. nologies if they wish to remain competitive. Mistake .free entry is not So imp,,rtant, but tak- Those that move quickly (but not too quickly) ing advantage of the full range of capabilities can often gain an edge over their competitors of the system requires a certain grasp of its via new products or productivity improve- logiclind capabilitiesmental skills, not man- ments in existing lines of business, Ultimate- ual (and different from the spelling and gram- ly, the greatest numbers of jobs inay disappear rrna now, learned in school). Proanctivity in where firms, industries, or nations do not keep many types of jobs increasingly depend on pace with technological advance. such abilities; it would he doubly unfortunate Firms or industries whose competitive pusi- .if the i.S. electronics indu:;:.ry were to.suffer tion is already in decline may he forced to shortages of trairied at the same time automate or purgl.fie other routes..to lower costs 373 374 = International Competitiveness in Electronics and higher productivity simply to survive. In Only where the market is expanding rapid- some cases, then, declining employment is ly can employment growth parallel,productivi- associated with attempts to revive competitive ty advances. This has been tile case in the advantage. particularly when an industry or semiconductor and computer industries, but firm is threatened with competition from low- not in consumer electronics. To the extent that wage countries. But it would be a mistake to the American economy continues to grow only attribute the accompanying job losses solely to slowly, many of the productivity gains flowing imports. In consumer electronics, U.S. cor- from applications of electronics and computers porations have automated their production fa- will have negative first-order effects on employ- cilities and mover' oft,hot e: this costs v. 5. jobs ment. Still, few practicable alternatives exist; in the short term, but may expi,. d or help once robots or other automated technologies maintain the total market for American prod- become cost effective, the pressures to use ucts over the longer term. Moreover, as the them become virtually irresistible. More jobs electronics industry becomes more and. more could ultimately be lost through failure to adopt international-with American firms procL,-:ing such technologies than by pursuing them. goods overseas for foreign markets as well as re-importation, and foreign firms setting up The implication is straightforward: some assembly plants hereit becomes increasing- people, companies, industries, and regions will ly difficult to evaluate impacts on the American lose competitiveness and lose jobs. Tile rela- labor force in isolation. tionships between technical change, employment, and international competition may be Most fundamentally, only by using labor effi- complex, but from the standpoint of public cientlywhich often means investments in policy, the negatives are wholly predictable. automationcan U.S. firms ma' 'ain thci_r in- They cannot be avoided, but the country could ternational competitiveness. Improvements in prepare for them, both to ease the inevitable productivitya pathway to increased compet- adjustments and to help maintain U.S. com- itivenesscan have serious employment im- petitiveness. Because changes in industrial pacts on particular groups of workers, geg- structure bring new job requirements, policy graphical regions, and industrial sectors. 'I lie measures aimed at encouraging both public 1-3sential question is: How can the negative im and privately funded education and training .ts rnemployment be minimized while cap- are central to effective adjustment policy. italizing on the potentials of new technology? CHAPTER 10 ustrial Policies Contents

377

383

37(,)

.., . ;mu]

, ..- ...... -120 CHAPTER 10 Na.:onal Industrial Policies

Overview

Government policies directly and indirectly tangibles. When counting the yen does not suf- affect the international competitiveness of in- fice, how does the United States countervail dustrial sectors. The impacts can be positive subsidies? or negative, tangible or intangible; they may fall on domestic firms or foreign enterprises. This ch,pter treats industrial policy in com- The American electronics industry has claimed parative fashion, with special attention to in- that the policies of the U.S. Governmentsome- stitutional context and the evolution of in- times damage its competitiveness, while for- dustrial policymaking, as well as the place of eign industrial policiesparticularly those of electronics in strategies for economic develop- Japanalso place it at a disadvantage. This is ment. Policies in the United States are covered a familiar argument: many U.S. business lead- only briefly; the next chapter treats U.S. trade ers assert, on the one hand, that U.S. policies policies in greater detail, while chapter 12 ex- are -counterproductive and That they would be amines policy alternatives for this country. better off without Government interference, Industrial policy means different things to and on the opposite hand, that in othercoun- different people. To some, the term brings to tries government polices, far from being coon mind government programs for supporting and terproductive, give their competitors powerful promoting targeted industries, typified by the advantages in international trade. Suchques- French "plans" or Japan's government-funded tions turn on the general tenor of relations research and development projectssector- among government; business, and other in- specific attempts to assist industry. Beyond sec- terest groups (consumers, organized labor) as toral measures, a vast array of public poli- well as the details of policy. ciesdealing with taxation. trade, human re- As the importance of electronics became Ob- sources, science and tecisnology, antitrust, vious anti competition intensified, foreign gov- labor markets and economic adjustment, governments sought policies that would promote ernment procurementalso influence the de- the growth and development of their own in- velopment and viability of industries like elec- dustries. These trends seem bound to continue, tronics. OTA prefers to view industrial policy not only in industrialized nations like Japan but broadly, as encompassing both sectoral target- in developing economies. Questions of central ing and the many policy measures withag- concern for American policymakers include: gregate rather than sector-specific aims that How do, industrial policies differ among na- often have less direct effect's on private firms) tions? To what extent can the effectiveness of International competitiveness, at root, de- these policies be evaluated? Do actions taken pends on the efforts of private firmsthis is by forer_m governments give the electronicsin- as true in countries like Japan with relatively dustries of these countries significant com- comprehensive and well-developed industrial petitive advantages? Can industrial policies policies as in the United Statesbut public. "create" comparative advantage? These are policies help shape the environment within hard questions. The monetary value of :;ubsi- which corporations .)perate and managers dies can seldom be approximated accurately. Even where this is possible, it does not tell whether the money was well spent or wasted. . _ More important, the industrial policies of coun- U.S. Industrial Competitiveness: A Comparison of Steel, AMP- anti Electronics (Washington 1).C.: U.S. Congress. 01 tries like japan work in large part through in- Ike of Technology As56,17,ment, 07.1-15C-135, July 11P31),p. 151.

377 99-11 1O 53 - 25 378 Internarional CommThriveness in Electronics inhke decisions. The decisions of envernment answer, but historical and institutional factors officials are important t...)o. Natioici have apas wcii as stages of economic development and proached industrial policy differently; govern- the exigencies of day-to-day politics play a part. ment i:Ltervention is more common and viewed more positively in France or Tapan than in Where government has for years promoted West Germany or the United Sometimes industrial development -France rather than public policies are clearly defined end con- Britainpublic sector involvement in the econ- sciously developed, sometimes they evolve fr emy is more widely accepted as legitimate. In ad hoc fashion-, the traditional pattern here. such countries, policies directed at a single industry such as electronics have usually re- Still, such gei.: :alities can mislead: in all flected overall economic objectives. Institu- major industrial nations, policies toward the tional mechanisms that facilitate coordinated electronics industry have changed over time; policymakingcentral banks or development in several, debates over new approaches are banks, respected planning councils, centraliza- underway. Furthermore, policies often differ tion of responsibility within one or a few across an industry. More than 20 years ago. bureaucratic ministriesenhance the ability of Japan began a series of programs intended to government officials to implement industrial foster the 1.,;r,)N.vth of an indigenous computer policies. These features are lacking in the iminstrybut the government did little by com- United States. During the greater part of the parison to directly promote consumer elec- postwar pet iodwhen American industries tronics. In the United States, public policies such:3 electronics and aircraft were clear toward the automoli:inclush'y have cer' fired leaders world competitionpublic policies on regulations, while trade issues have been here were directed, not at economic develop-, stre,sed in the context of steel. U.S. agricultural policy has been much more highly de- ment, but at regulation. veloped than policies toward manufacturing. As th's chapter demonstrates,. industrial Why then have some nationsFrance, Japan. policies will be a prominent feature of the in- Taiwan, for examplesmoved toward well- ternational competitive environment for the defined and rather comprehensive policies foreseeable future. While other countries are directed at electronics, while countries like the busy developing them, the United States is still United States have not'? There is no simple groping for a response.

The Context for Industrial Policy

Public policies directed at electronics should receded to less than 5 percent of GDP. From be viewed in light of a nation's overall eco- 1976 to 1980 alone, Lie share of Japan's exports nomic developmentstrategy. Table 77 gives in accounted for by elect;tics went from 9 to 14 summary form a number of indicators of eco- perccmt. 2 nomic position and induct; )o!'hy for five Such shifts, the results of complex-economic--- countries. Electronics and other h_gh-technol- currents, form part of the policy context. Major ogy inde tries grow more important as manu- changes have also been occurring within elec- facturing and services displace agriculture. In tronics. Continuing the example of Japan, con- japan, agriculture accounted for more than 20 no.ht of the gross domestic product (GDP) in 1955, when the electronics industry was in- ,Trends in the Electronics Industry in 1980 (Tokyo: Electronic significant by intern standards; by the Industries A.Isoci,lion of Japan, 1981), p "Industrial Review end of the 1970's. ihe agricultural sector had of Japan-1981.- Japan Economic lourn,11, p. 33. Table 77,Economic and industrial Policy Indicators

United Stales Japan West Germany United Kingdom Taiwan

1. seNices as a proportion of gross domestic product (GDP).... 63% (1980 55% (1979) 49% (1979) 53% (1980) 38% (1979) 2. P&D.erenditures as a percentage of GDP 2.39% (1981) 1.97% (1979) 2.32% (1980) 211% (1978) 0.65% (1981) 3 Government as a source of R&D funds 47.2% (1982) 27,4% (1980) 49% (19791 55% (1979) 43%a (1980)

(rather than mii.tary) R&D as,3 ,..'.,rcentage of total R&D. 70,0% (19S1) 97.4% (1977) 92.4% (1979) 69.7% (1978) NA 5 Electronics R&D as a percentage of toia: R&L. (1975) , 22°A 28.4% (1979) 30% 26% NA 6A. Government R&D spendingon electronics as a percentage of total government R&D ;pending (1975) 30% 32% 31% 34% NA 6B industry R8,D. spendingon electronics as a percentage of toe! industry R&D spending (1975' 21% 26n'o 30' 21% NA 7. Percentage 1978 government P.D funds going to: Economic development 22% 13% NA Defense 12% 52% NA 8. Organizatm of policymaking syFiem Fragmented Centralized Decentralized Fragmented Centralized 9. Gov rnment.business relations Adversarial Cooperative Structured Semiadversarial Cooperative representation of business and labor views

10 Patents granted 119811 65,770 50,904 13;429 22,924 NA Balance of trade in electronics with the U'tited States(1981\ (millions of dollars) S4,235' $5,8780 $1,592 ±$1,696 -$1,635' 12, Overall policy and sirdtegy Ad hoc Leapilog: Adaptive: Adaptive: Catchup ind.genous stresses stresses technology technology commercial development development appocations NA r.01 qxcludes expenditures '01 mil,jary Un[ted States yvith all nations

CNeptIve 57 lento exports to tee United Stalesexcerio,i,yj imports from the Unwed Stales.

Ten rear Economic Development Plan for TaiNan, Republic, of China," TaYoan Councilfor Economic Planning and Development, March 1181

Technical Change and FJ:'onomic Policy (Paris Organization 'or Economic Cooperationand Development, 19801, p. 31 Denser San gyo no KeiuSdiN3 no Hoho to sono Elcyo ni Kansuru Crime Hokoku (Survey Report, on Trends in the Internationall:.alion of the Electronics Industry and Their Influence, PartII ol East and Southeast Asial (Tokyo: Nihon DenSht Hikai Kogyckal (Electronic IndustriesAssociation of Japan), March 1681, p 121.

K. Scnalt, Industrial Innovation in the United Kingdom, Canada and the United States (LondonContemprinl, July 19811, p. 9 J Bionson and H. B. Malmigren,"Technology and Trade Policy' Issues and An Agenda for Action," report prepared for Department of Labor and OYce of the US Trade Represenlafroe, October 1981,p 158 S,,;7',ey of R&D Activities in the Yea 1980, Pepublic of China," National ScienceCouncil, Republic of China, 1981 ante 1980 Ohl ashingr.'1, D.C.: National Science Board, National Science Foundation, 1981),pp. 214214 Ee,-,stronics. Jan. 13, 1982,

C J Mcsbacher, "Will R&D Funds Be More Than SP Billion in '82," Industrial Research& Development, January 1982 p 106 Nat,onal Ri!..e.!15 of Science and Technology Resources-1982 iWashmgi311. DC' National Science FOundaoon. 19821, p 33. Devc;.:irent Report 1931 (Nevi York, Oxford University Press, 19821,p 114.

DOLT), 'or Science and TechnoIly, Nation' Research Council (San Francisco WH Freeman,. 19132i, p 519. Elovforec Marne' Data 1982 (Weleplon, DC.. Electronics Industries k,.ociation, 19821.7

from LIS Patent CPice, Embassy of Japan, Science Divinn, Coovalion Counci American Affairs. Republic of China

3 3 380 Inte. at;cnal Competitiveness in Electronics surer products declined from two-thir is of_ of persistent economic problems. Moreover, at country's excort.. of electronics in 1971 to aggressive industrial policies in one country about half by the end of the decade, while simi- breed responses elsewhere. The turn toweed conductor and computer exports increased. industrial policies. particularey in nations lack- Many of Japan's consumer electronics ship- ing a tradition of government involvement in ments to the United States have been displaced economic affairs, is partly a reaction to these by prc'lucts from other Asian nations, partly new circumstances; in other countries, in- a resut of rapid industriaLzation in countries dustrial policy is nothing new, just a continua like Hong Kong and Koree.3 tion of past practices under a different ;lame These changes in the composition of japans exports reflect shifty in the international divi- Policy Orientations sion of labordeveloping economies are now As partofa nation's overall development producing more consumer electronic goods, strategy, industrial policies can be directed at while advanced nations concentrate on high- catching up, leapfrogging, or staying ahead in er technology products. Industrial policies can worldwide competition (table 77). Absence of be viewed as responses to such structural a clearly defined industrial policy may indicate changes: they may attempt to modify or resist general satisfaction with the situation, the case them, to smooth adaptation to change, to com- in the United States until recently; lack of a plement or even induce it. "Success" is most well-defined industrial policy could also reflect likely when policies work to accommodate or a belief that it is improper for government to reinforce rather than impede changes in in- concern itself with such issuesa widespread dustrial structureprovided the policies are attitude here. In contrast, during the 1960's the based on sound judgments concerning the French and JapaneSe began supporting and de- strengths and weaknesses of a country's in- fending their computer industries against what dustries, both domestically and in the interna- they viewed as an American challenge. tional marketplace. This is no easy task: still, policies toward electronics in othkcountries, In many countries and at many times, defen- if not the United States, should be v' wed sive industrial pclicies have been devised-- in these termsas components of natio._in eco- intended to preserve existing economic struc nomic strategies based at least in part on per- tuis, maintain employment, and protect be- ceptions and projections of structural shifts in leagured firms and industries.4 Often defended the world economy. as temporary (ch. 1.1), protective measures fre- quenny turn out to be persistent if not perma- International economic conditions now favor nent. American corporations less than in the earlier postwar years; this is one reason for the grow- Adaptive inrlus: jai policies seek to ening interest in industrial policy for the United courage structural change by facilitating shifts States. This country, along with the rest of the of resources to growing and productive indus- industrialized 'est, has experienced low rates triesthose in the process of becoming more of econom:c griwt1Lrising inflation, and high competitive. Tn contrast to the defensive ap- unemployment over the past decade. Competi- proach, adaptive industrial policies begin with tion has intensified among firms here and the assumption that some sectors will eventual- abroad, all seeking to maintain or enhance their ly decline in size and importance. in practice positions in markets that may be growing only the boundaries 1;etv.reen various sorts of in- Under the-se conditions, governments dustrial policties are vague; for instance, sub-; have turned to industrial policy as a way out sidles or protectien for a given sector may be rationalized as a means of encouraging adap- ?White Paper en Infernal lona! Trade-1980(Tokyn: Ministry Internationat *I'rede and Industry, September 1080), p. 32. As See W. Diebold, Jr..Indesirial Policy a5 an IntervalimalIFS1.10 discussed elsewhere. Orderly Mari- -Hug Agreements have, also (New York: McG:pw-11 1880), op. 7-8. for an outline of typos contributed to this shift. of industrial policies, Ch. 10National Industrial Policies 381

tation, while in practice they function as tract research undertaken by smaller enter- defenses against decline. prises, along with cooperative R&D in in- More ambitious than adaptive policies are dustrial research associations. The Very High- those that attempt to induce change. This im- Speed Integrated Circuit program of the U.S. plies moving beyond a response to economic Department of Defense is aimed at integrated forceshere government takes the lead in ini- circuits (ICs) for military applications, but will tiating industrial change, with the object of im- have commercial spinoffs. The Economic proving the competitiveness of some sectors Recovery Tax Act of 1981 included a tax credit of the economy. Both this approach and the for R&D spending, as well as accelerated de- adaptive strategy tend to tre associated with no- preciation of equipment used in research. Ja- tions of dynamic comparative advantage and pan also offers tax credits to firms that increase the belief that governments can anticipate and their spending for R&D over past levels. Be- plan for shifts in the structure of advantage. yond this, the Japanese Government directly supports projects aimed at commercial micro- As pointed out in chapter 5, the competitive- electronics and computer technologies. ness of all sectors of an economy cannot improve at once. To pursue a positive develop- Many countries assist regions, small busi- ment strategy, .a nation must begin with at least nesses, perhaps entire industries through in- the implicit acknowledgment that some of its vestment grants and subsidies. The United industries will likely decline. Common ground Kingdom's National Enterprise Board provided concerning the prospects for industry is easier 50 million pounds to capitalize the semicon- to find in economies with simple structures. ductor firm Inmos. In the United States, the Nations that are still attempting to catch up Small Business Administration loans money at have an easier time in formulating policy; they favorable interest rates and with lengthy repay- face fewer choices, fewer possibilities. ment periods. Regional development loans have stimulated investment by American and The Tools of Industrial Policy Japanese semiconductor firms in Ireland and Scotland. National banks, particularly in- In market economies, governments bring a dustrial development banks, have been impor- more or >,eT,!,-; standard set of policies to bear on tant vehi-Aes in many countries for channel- inductr,ai development measures used for ing funds to particular sectors. purposes ranging from improving competitiveness to encouraging regional development or Government procurement is widely used to strengthening the national defense. Regardless support national firms. Military procurement of whether a country is attempting to pursue has been much more important in the United an integrated policy, a wide variety of govern- States, France, and Great Britain than in coun- ment actions will inevitably affect the in- tries like West Germany. The "Buy Japanese" dustrial portion of its economy. policies of public corporations such as Nippon Telegraph and Telephone (NTT) were for years In the case of electronics, many countries an integral part of Japan's policies in elec- have instituted policies affecting costs and sup- tronics. In 1980, NTTwhich purchases siz- plies of capitalfor R&D as well as for invest- able amounts of communications and elec- ment in plant and equipment. R&D supports tronics productsagreed, after lengthy negotia- can take the form of low interest loans, direct tions, to open some procurements to foreign subsidies, or government contracts. In West- bidders. American firms have made only lim- Germany, government funding supports basic ited progress in selling to NTT, but the atten- research as well as projects aimed at commer- tion given the case indicates that government cialization carried out by the laboratories of the procurement is becoming more subject to in- Fraunhofer "Gesellschaft; the German Ministry ternational negotiation, perhaps less usable as of Science and Technology also subsidizes con- a tool for the promotion of domestic industries ), 385 382 International Competitiveness in Electronics

(nontariff barriers to trade, of which this is an tions to the Foreign. Corrupt Practices Act example, are discussed more extensively in the like,ise intended to support U.S. firms in next chapter). foreign markets passed the Senate but not the House of Representatives in 1981. Still another category of policy measure includes those bearing on the regulation of in- Policy measures of the types outlined above dustrial structure. Nations can influence the have been deployed by governments every- structure of their industries by encouraging or where in their attempts to influence the discouraging mergers, not to mention national- development of industry and improve competi- izing firms or industries as the Mitterrand gov- tiveness. Generally speaking, tariff barriers, ernment in France has done. American com- controls on foreign investment, and competi- petition policy has emphasized the regulatory tion policies were the tools of firit choice dur- sidei.e., antitrust enforcementwhile in ing earlier postwar years; since the late 1960's, France and the United Kingdom, governments as trade liberalization gained momentum and have steered companies into mergers (e.g., the direct trade barriers were dismantled, R&D computer manufacturers CII-Honeywell Bull policies and investment stimuli have come to in France and ICL in Britain) intended to create the fore. In the wake of intensified competition "national champions." Encouraging mergers, in a wide range of industries, trade negotia- often through financial incentivessometimes tionsboth bilateral and nv.4.11ateralhave in- referred to in Europe as structural policyhas creasingly centered on szt..,5.tkiies and indirect been a common feature of policies toward elec- barriers. tronics in most developed nation. While the typical rmx_ hdustrial policy Some countries use foreign investment con- measures has shifted over titn,-.the group of trols-to restrict inward flows of capital, and policy tools from which they ft re chosen has thus preserve domestic markets for local firms. not changed veryl--...tch. The policies In years past, such regulations, as well as re- of various nations draw on isrie sa:'.4'8 basic in- strictions on imports and technology from gredientsR&D supports, invez::-,:ent grants abroad, played a central role in the industrial and subsidies, public sector pocuzement, policies of Japan; several examples in elec- merger policy, controls on foreign investment, trouics were outlined in chapter 5.5 tariffs and other trade policies. Naiions combine these depending on their assessments of Finally, tariffs and other varieties of trade the strengths and weaknesses of their own in- policy are an ever-present force in international dustries and the objectives of their economic competition. Countries erect tariff walls to pro- development programs. tect new Or old industries; the European Eco- nomic Community, for instance, maintains a The key to effective national policies has lain, tariff of 17 percent on ICs to discourage im- not in the individual policy tools but in their ports and stimulate domestic production. The combinationin the extent to which the poli- United States negotiated import quotas on col- cies chosen complement one another and work or televisions with Japan, Taiwan, and South toward a more or less consistent set of objec- Korea during the 1970's in an attempt to deal tives. The timing of policy initiatives and the with the problems of this industry (as discussed receptivity of private firms to government pro- in ch. 11). Trading nations all maintain export grams are also important, but the success- or promotion measures intended to help local failure of industrial policies is determined to firms sell in the world market. In the United a large extent by the ability of policymakers to States; the Export Trading Company Act develop and implement a consistent frame- (Public Law 97-290) passed in the fall of 1982 work and approach, one appropriate to that na- is one of the most recent examples; modification's position in the international economy. The remainder of the chapter reviews industrial policy in a number of countries, with par- 'See also R. S. Ozaki. The Control of Imports and Foreign Cap- - ital in Japan (New York: Praeger. 1972). ticular attention to electronics. (1110National Industrial Policies 383 Industrial P6liciets 'Obrrtpared

The failures of in po,licy are. rnuali ,,dgr .ir- ,Insive'aldustrial structure, Much as in more evident than th- does one. planners, also reconsidering Weigh the contribuii .1f ricy,errinnent their l'-.adil.i;:nal.approach, have become more to economic develer,;;it--eiitlrre.1 on a generci!- open teehnology exchanges and business or a sectoral basis v yQ 11.:,:g.kw.i1ry has bee a ventures in.../,':;) 7ing foreign firms. While the inin the -take-off- stave,... .forces dustries i7 -cou'atries like Taiwan and South Mg more or less in co.re-.:0-1 spir?2,:lindustrid,',- Korea 13..avz. 2:1mady become major producers ization? This was the :: japanes of ni:xlcileq.17-,.? products like color TVs, steel, shipbuilding, indi5 simpler rnic..Az.,:,-.tctronic devices, and computer tries in earlier years. a *Labor fore:: 2c.,:ripherals, i!s far from certain that such na- and rapidly expandin:nart:,at5.',4(ere aided by tiVOS can suf:,c:ed in advanced electronics tech- the government's push, at is iK)1,1; the case in riotogies. Wi%.7.7power limitations are the most other nations that ha-se booxi,l. experiene.e Ire re c st-.-a FA. rapid economic groo:th-. 1.zioitith Koref. Develop;Ag Covro.!TfAls 11/1.g.0.1'1;.'0?.. Government has consistently The past rise in vapiriiclustrialization; the public sec - the electronics inc v,-,v:i.c.is c:1 el 1:4:amber of new- to has perhaps been more pervasive ly industrializing {NICs), most any of the other NICs. Policy in- them in Asia. Mar.:.; ,.'.:.a-:ie7ilationsTaiwan, Vtr-1:21r.tints hale ranged from money to gui- South Korea, Brazilhave chAen paths of goir-. reliatz-1..- it '2,,adirect taxes, raw materials ernment-guided economic davelopn'iont, albeit ;.:A11,5 to exporters, target figures with many gradations in the (rdent of goVerA- for R&D. Korea's export fi- ment involvement. With the exception n..:13e.j,ni:. have also been unusually China, which has emphasized "self-suffi- to other NICs.6 ciency," the Asian nations have relied heavily For ma- the Korean economy- ex- on imported technology while capitalizing on panded at rate, with annual increases cheap labor. In countries like Singapore, Hong in gross national product (GNP) averaging 10 Kong, and Taiwan, economic development pol- percent over the period from the early 1960'S icies have relied more heavily on encouraging into the mid-1970's. Labor-intensive manufac- diversified exports of manufactured goods than tured goods provided the foundation for this protecting local industries against import com- growth;-exports have become much more im- petition. The typical attitude toward foreign portant to South Korea's economy over the past- electronics firms has been pragmatic, with decade, growing from 12 percent of GNP to 35 American and Japanese involvements tolerated percent.' Electronics has been an export leader, or encouraged because of benefits in technol- the most rapidly growing sector. Korea's elec- ogy transfer and infrastructural development. tronics industry is still small compared to In years past, the electronics industries in Japan's, but it accounts for more than 10 per- most NICs centered pn relatively simple 'con- - cent of Korean exports. sumer productsradios and black-and-white More recently, South Korea's economic mira-. TVs, pocket calculators, electronic watches, cle has fallen on the same hard times that have toys and games. Now, policy pronouncements from these countries are calling for shifts '"Korea's Eximbank Provides Incentives To Diversify Export toward more sophisticated goods. In Taiwan, Mix. Destination,-IMFSurvey. Nov. 26, 1979. p. 366. 7P. Hasan and D. C. Rao, Korea: Policy Issues for Long-Term which has perhaps the most ambitious govern- Development (Baltimore: johns I lopk;.ns University Press, 1979), ment programs, the stated aim is a more knowl- p. 20.

381 3134 International Competitiveness in Electionics

afflicted the rest of the world. The slump was Table 78.Korean Etectron:cs Production sudden: whereas Korea's output of electronic Output (millions of dollars) products grew at the astounding rate of 40 per- 1981 19a6a cent per year during the 1970's, production ac- Consumer $1,600 $5,800 tually fell in 1980, although rebounding strong- Industrialb 490 2,700 ly in 1981.8 South Korea's Government con- Components 1,710 4,800 siders continued-growth in electronics neces- $3,800 $13,300 sary for recovery, and the industry remains a Total electronics exports $2,200 $7,000 aProjected. focal pr-t of development strategy. Korea's corricutels and le!ecommunications equipment. fourth economic plan (1977-81) concluded that SOUF.a Spaeth, "Korea's Electronics Industry Making Rapi;.: GainsShit! to High.Technology Products:' Asian Wall Street Journal Week, f, Dec. long-term export viability would depend on 20. 19132. p. 1. The projections came from South Korea's Ministry of structural changes in manufacturing. The plan Commerce and Industry. called for rapid increases in exports of electronic products." Korea's Government assumes of the Korean industry. Such a reorientation that other developing economies will provide will entail shifts in R&D emphasis, with in-, stiff competition in sectors like textiles and ap- creases in funding for both product and proc- parel, where Korean industry has in the past ess technologies. To this end, the Ministry of been strong; thus, the country needs to con- Commerce and Industry has begun channel- tinue moving into durable manufactures for ex- ing funds to Korean electronics firms for de- port. The government also intends to deem- velopments in semiconductors and comput- phasize petrochemicals and heavy industries ers." like steelsectors that helped lead Korean eco- To help focus research efforts, the Korean In- nomic growth in past years. The fifth and latest stitute of Electronics Technologyestablished plan released by South Korea's Economic Plan- with government support in Gumi, the coun- . ning Board proposes dramatic cuts in in- try's Silicon Valleyis to be built into a center- vestments in these portions of the economy, piece for research in electroi-21'cs. The institute with expenditures on electronics boosted has been installing production lines for very substantially." Table 78 summarizes proiec- large-scale ICs; the equipment will be used for tions by the Korean Government; electronics commercial production as well as engineering exports are expected to climb to $14.5 billion development." While the staff of the $62 in 1991. The most rapid growth is projected in million institute remains small, planners hope industrial electro- ,cs products, including com- that it will eventually house more than a thou- puters and communications equipment, with saud research workers." a heavy emphasis onmicroelectronics. The share of total electronics output accounted for In addition to R&D assistance, the South Ko- by consumer products is expected fo begin rean Government has provided investment shrinking by the latter part of the decade, with funds to electronics firms and supported them a pronounced move away fromthe less sophis- through procurements. For instance, Gold Star ticated components that are currently a staple Semiconductora joint Korean-U.S. venture will receive a loan of more than $40 million °Denshi Sangyo no Kokusaika no Hoko to sono Eikyo ni Kan- from both foreign and domestic sources, in- suru Chose Hokoku (Survey Report on Trends in the Interna- cluding the Korea Development Bank, to man- tionalization of the Electron,cs Industry and Their Influence, ufacture telephone switching equipment which Part II on East and Southeast Asia) (Tokyo: Nihon Denshi Kikai Kogyokai (Electronic Industries Association of Japan). March 1981). p. 103; A. Spaeth, "Korea's Electronics Industry Making "One report states that $800 million has already been invested Rapid Gains in Shift to High-Technology Products," Asian Wall by the government"Fourth Five-Year Plan," Electronics 14'etik- Street Journal Weekly, Dec. 20, 1982, p. 1. ly. Apr. 25, 1979, p. 19. Denshi Sangyo no Kokusaika no Hoko to sono Eikyo ni Kan- 12-Korea's Electronics Industry Making Rapid Gains in Shift suru Chose Hokoku, op. cit., p. 56. to High-Technology Products," op. cit. Eventually, the institute "N. Thorpe, "South Korea's Economic Program Reduces Ex- expects to sell the production facility to a private firm. pansion of Several Major Industries," Wall Street Journal, July ""South KoPea Seeks Electronics Rebound," New York Times, 24, 1981, p. 24. Mar. 24, 1981, p. D5.

38° Ch. 10Natio:la, Industrial Policies 385 will be purchased by the Ministry of Commu- onstrated their ability to compete in the color nications.t4 A second major Korean electronics TV market, but if they cannot get foreign tech- firmSamsung, also partly U.S.-ownedis in- nology in other areas their progress in elec- volved in the project as well. When the govern- tronics will be slowed. ment decided to begin color TV broadcasting In view of these obstacles, does South Ko- in 1980, Samsung won loans to aid in the pro- rea's development strategy seem feasible? duction of color receivers. Foreign firms have There is no question that 'Korean firms are well also benefited from investment incentives, al, placed to expand their shipments of products though South Korea's electronics industry has like color TVs, passive components, discrete been less dependent on overseas capital than transistors, and small-scale ICs to more ad- most others in Asia. Foreign-owned companies vanced countries. Korea is already the world's are exempt from Korean income, property, and bigge=:3 producer of black-and-white TVs, and corporate taxes during the first 5 years of Koet.:n firm:: have been among the leaders as operation.15 Asian nations have taken over mLch of the Government generosity has not prevented world's i'.,roduction of consumer electronics bottlenecks such as rising labor costs and skill productstable79.But developing the capabil- shortages among the 180,000 employees of ity for designing and developing new products Asia's third largest electronics industry. The based or domestic technology and resources recent push toward indigenous technological is a more ambitious and less certain undertak- capability implies heavy R&D commitments, ing thin manufacturing commodity-like prod- but most South Korean firms have only limited ucts using standardized, well-understood tech- human. and financial resources to devote to niques. these ends. Furthermore, other countries are likely to be cautious in transferring electronics Taiwan technology to Korea now that the country's competitiveness is apparent. Japanese firJis The Taiwanes electronics industry runs a have refused repeated requests for licenses close second in sales to Korea (ch. 4), and covering video cassette recorder (VCR) tech- employs more people. Both governments have nology.18Korean producers have already dem- followed the Japanese pattern in emphasizing electronics. At the center of Taiwan's current ""Cold Star Semiconductor Raising Loan for Move Into Ad- 10-year econorni,-. plan (1980-89) is the develop- vanced Electronics." Asian Wall Street journal Weekly, Apr. 13. 1981. p. 8. The company is owned 44 percent by Western Elec- ment of the machinery, electronics, and 'nfor- tric and 56 percent by the Korean Lucky Croup. mation industriesfavored because of high 15C. Webb, "South Korea," Electronics We^kly. Apr. 25.1979. value-added, modest demands for energy, and p. 19. "M. Inaba, **Koreans Press Japan To Share Video Cassette comparatively high technology content. Tai- Profits." Elgctrorric News, Nov. 30,1980. p. F. Nonetheless, sev- wan has the best trained corps of engineers and eral Korean firms already produce VCRs of their own design. scientists in the Far East outside of japan, mak-

Table 79.Market Shares in Consumer Electronics for Japan and Other Asian Nations

Share of total world market, 1979 Japan All other Asian nations Total Asian share Videocassette recorders 0 93.2% Color TVs 3.8 31.5 Monochrome TVs 49.6 65.9 Radios 71.8 77.0 Audio tape recorders 52.8 91.0 Auto radios and tape players 18.7 67.3 Other home audio equipment, stereos 12.1 52.2 SOURCE: Denshi Sangyo no Kokusalks no Hoko to sono Eikyo nl Kansuru Chosa Hoko .0 (Survey Report on Trends In the internathmallzation of the ElectronttAIndustry and Their Influence, Part II on Efailt and Southeast Asia) (Tokyo: Nihc7 Kikal Kogyokai (Electronic Industries Association of Japanj, Mn ch 1981), p. 2.

3 .386 4 Interrztiorral Competitiveness in Electronics ing the more technology-intensive sectors nat- the country's efforts are concentrated on ural targets. The country's development plans special-purpose machines compatible with the encompass ICs, computers and peripherals, Chinese language, along with minicomputers, and high-end consumer products such as peripherals, and software.19 Examples of the VCRs. The Taiwanese, like the South Koreans, initiatives being discussed include join'. ven- are not satisfied with their image as manufactures with Western firms in which government- turers and assemblers of components, pro- sponsored training efforts would nrovide ducers of cheap TVs and consumer goods. Ac- skil!.;,1 workersf.. r`,.are development.20 cording to the current government plan, elec- Along with other 2-:Isian electronics indus- tronics output is to double over the decade." tries, Taiwan depends heavily on exports (table For some time, the Government of Taiwan 19, ch. 4; Taiwan exported 80 percent of its has been encouraging shifts from labor- to electronics production in 1979, South Korea 70 knowledge-intensive industries. One vehicle percent), with the bulk of these shipments has been the Electronics Research and Service going to U.S. markets. Taiwanese firms such Organization (ERSO), which gets about 40 per- as Tatung and Sampo have already set up color cent of funding from public s:_.urces. ERSO, TV production facilities in the United States. established in 1974, is one of four divisions of With en economy that has Teen growing at an the Industry Technology Research Institute annual rate of about 8 percent, unemployment (ITRI); projects have included computerized in- at less than 2 percent, and a persistent trade dustrial control systerris, Chinese language surplus with the United States, Taiwan's elec- computers, and semiconductor development. tronics industry is well positioned for further The organization also negotiated the technol- expansion. But Taiwan faces many of the same ogy transfer agreenient with RCA that helped problems policymakers in Korea are grappling Taiwanese firms produce c-MOS '-Cs for the with. The country will need grever numbers country's watch industry.18 ERSO is engaged of well-trained technicians and engineers, in manufacturing as well as R&D, and has higher levels of spending on R&D, and contin- helped introduce improvadquality control pro- ued improvements in labor productivitythe cedures in Taiwan's electronics industry. latter of growing significance as wages rise. Wage increases have rendered Taiwan's As for South Korea, Taiwan may not have labor-intensive industries increasingly vulner- the financial and human resources needed for able to competition from other developing rapid development in electronics based on in- cOuntries,.an important motive for the govern- digenous technology. And again, foreign pat- ment's stress on knowledge-intensive sectors ent holders fearful of new competitors appear reluctant to negotiate agreements with Tai- and another parallel with Korea. Policy pro- wanese companies, particularly in more ad- nouncements call for greater use of computer- mr leaders based autorn liontuincrease productivity and vanced products such r export competitiveness. ITRI leaders hope that within the Japanese e have Taiwan will be able to independently develop urged "accornmodai, pug Asian small computers and the associated software economiesmeaning that Japan should con- fcr both domestic and export markets. Govern- centrate on leading-edge technologies while ment planners believe ti.at Taiwan will have importing less sophisticated goods from else- the best chance of successif,instead of where in Asia; but if Taiwan's government is attempting to challenge IBM or the Japanese,

Ving. "Tai I is Counting on Its Computer ineustry to Economic Development Plan for Taiwan. Repub- ly 'he Finnorny,- Asian Wan Street lour- iii Tai wen Council for Econdmic Planning aidDi 'y,grile or Perish: Electronics VfA(.!;'' March 19812, p. 39. M. message."iattar Winds, October 1980. p. 11. "K. 11. laiwan Pushes High Technology." Electron "Noig,%is Softwiire Center ir 1.'",'Iwart.- Asian May3. 1980, p. 1GO. jour- Mar. 15, 1982, p. 4-

3;iti Ch. 117NatiotJat Industrial Policies 387 serious about its commitment to high technol- Commission, a central agency for policymak- ogy. such an accommo6ation would probably ing and implementation, is a further indication not be acceptable' 21 f othe government's new direction. Ten factorises in China now produce comput- China ers,: ranging from microcomputers to machines More strongly committed to self-sufficiency similar to PDP-1.15 and IBM 360s.24 The State than other industrializing economies, Chines .Administration of Computer Industry tSACI) progress in electronics and other industries has has programs underway to utilize the nation's been unevenin part because of longstanding existing computing capability more efficient- conflicts between the development of science ly, and intends to move tc ward smaller ma- and technology and the quest for revolutionary chines rather than relying on large main- social change. China's desire for :self-sufficien- frames. As one route to such objectives, SACI cy has also created obstacles to efiiment mass is establishing joint ventures with foreign con- production; as a case in point, components are. cerns. In 1981, the China Technical Services still soldered into circuit boards by hand, while Corp. and the Japanese firm NEC (Nippon in the West wave soldering has been employed Electric Co.) signed an agreement for a com- for more than 20 years. This is not to say that puter center in Beijing. NEC will provide a the country's electronics industry is unrelieved- medium-sized machine free of charge, and an ly primitive: the People's Republic of China annual 4 -month training course for 30 to 40 (PRC) builds mainframe computers as well as Chinese software specialists. The Chinese will ICs roughly comparable to mid-1970's U.S. supply other facilitiesi. along ro,rith the center's products. Nonetheless. until recently most of staff, including interpreters. A similar agree- the computers were one-of-a-kind machines, ment has been signed with Sperry Univac, 'asking even transportable software.22 while negotiations have taken place with other The picture has changed in the last half- U.S. firms, including Wang Laboratories and dozen years as a new consensus on the impor- .Honeywel1.25 Both Japanese and Western firms tance of science and technologyone of the hope to establish themselves, in the potential- "four modernizations" advocated at the Fourth ly lucrative PRC market. National People's Congress in 1975emerged AS such ventures indicate China is putting among China's leaders.23 in the National: Plan a good deal of effort into training computer for Development of Science and Technology, specialists as a basis for more effective utiliza- announced at a nationwide science conference tion of information processing technologies. in 1978, eight technical areas were singled out Electronics and computer technicians will for special emphasis, among them computers. study-at an Information Processing and Train- In calling fOr the development of China's capa- ing Center, established in '1979 wit. Wlity in a wide range of electronics technolo- from the United Natti. As. Among the plans 'for e,es, including large -scale ICs, microcomput- the center, to be equipped with as Burroughs ers, peripherals, software, and computer net- mainframe, as well as five Hewlett-Packard werks, the plan termed computer science and 3000 series minicomputers, are development technology "a conspicuous hallmark of the of a world patent index, collection of inforrna- level of modernization of a country."' The tion on food supplies, a data base on power reestablishment of the Science and Technology generrtirm and distrfi.Johoti :Tor the Electric Power studies of urban traffic flows, and "Taae YoecoP( .iii modeling.26 Soch endeav- ''See the summary of the Electronic incinstry Associatton of Japan's report en Asian electronics in A-; in Wit11 Street four- "D. Burstein, "Chinese' Fomerit Another Revolution," Elec- Weekl;% gene 8. 1981. tronics, Jan. 13, 1983, p. 115. 72K. Berney, "Computer Sales to China." China Business "K. Berney, "China's Computer Revolution," China Business Reviow, September-October 1980, p. 25. Review, November-December 1981. p. 14. "R. P. Sulitmeier, Science, Technology nod China's Drive for ""U,N. Aid for China's Computer Modernization.- China lies;. Modernization (Stanford, Calif.: Hoover Institution Press, 1980). ness Review, September-October 1980. pp. 33.34.

331 385 Jn.ternational Cornpet;tive-7ess'in Electronics ors imply that China--at least at presentmay The discussion above does no more than be more interested in applications than in sketch in a few of the outlines of industrial building production capacity, not only in com- policy toward electronics in developing Asian puter ea uipment but in electronics more gen- economies (Japan is treated in some detail erally-. below). Outside Asia, governments in countries like Brazil and Mexico have also nurtured Other Nics rapidly expanding electronicsindustries. Brazil, for instance, has used access to its Rapid growth in the Asian electronics in- rapidly growing market as the carrot for ac- dustry extends well beyond Taiwan and South quiring U.S. minicomputer technology.29 In all Korea. Hong Kong's companies, which have these countries, foreign investments by Ameri- been basically assembly-oriented suppliers of can and/or Japanese firms have been one of the products like watches and calculators, ac- starting points for indigenous development. counted for 13 percent of the colony's exports Today, these nations are aggressively attempt- in 1980.27 In. Singapore, which has also been ing to strengthen their own capability and a major assembly site. the government has in- reduce their dependence on more advanced troduced policies intended to encourage semi- countries. None of the policies employedthe conductor manufacturing as well as product establishment of government- supported R&D ticn of computer hardware and software; the facilities, tax breaks and financial subsidies for government, for example, owns 25 percent of local firms, preferential procurement, govern- Tata Elxsi, a joint venture involving U.S. and ment encouragement 0: or participation in Indian interests formed to make mainframe joint ventures with foreign firmsare unique processors.28 Government policy in Hong Kong or even unusual. Such measures are part of the has been less intrusive than in Singapore, but standard list. Still, government planners in the electronics industry there has also been NICs have often pursued them more consist- moving toward high technology. ently and forcefullySouth Korea is especial- Clearly the Asian NICs are all, in one way ly striking in this regardthan have developed or another, attempting to learn fromand economies. This is partly because the paths are emulate Japan's approach to industrial policy. well marked for NICs in comparison to ad- In the earlier postwar yearS, Japanese com- vanced nations with complex industrial Simi' panies imported technology, while government tures. The explicit focus on stre ,gt*,01 decisions favored heavy industries; newly in- mestie tecl' nie '.1 how- -0, shift in dustrializing nations in Asia have already aban- emphasis has led to increaseJ demands for enned this approach for one more like Japan's transfers of technology as a condition for sales current industrial policy. The question is or investment by foreign firms. Countries mak- whether South Korea, Taiwan, and other NICs ing such demands rIT alternatively, offering have, at this juncture, the resources to supper' incentives to ".tract .;hnology inflowssee technological self-sufficiency. Rot even if ;heir lift 144 as a pro,equil, for building their own progress _in developing how technol- capabilities. Soffit.; multinational electronics ogies proves slow, these countries will be in- firms have accepted these conditionswhich creasingly competitive in world markets for at times have been a prerequisite for market less sophisticated electronics products, well entry, a tactic that Japan employed in years able to challenge manufacturers anywhere that pastmore readily than others. The draft fail to maintain a technological edge. UNCTAD (United Nations Council on Trade and Development) code on technology transfer

l'"Says Electronics Could Lead as Hong Kong Export Earner." Baranson u nd H. B. Malmgren, "Technology and Trade Electronic News, Oct. 26, 1981. p. FF. Policy:. 1Aues antia Agenda for Action," report prepared for 15"See CPU. Software Mfg. Leading Singapore's Ftiture."Elec- Department of Labwr arat Office of the U.S. Trade Representa- tronic News. Dec. 7, 1981, p. (2 tive. October 1981, ; 125-126.

3 9 Ch. 10Naiional Industrial Policies 339

illustrates the strong desire among developing ment has paid more attention to some indus- nations everywhere for technology acquisition tries than others, this has most often been a on more favorable terms. result of political pressures, as in the case of A problem that the developing Asian econ- textiles, or national security considerations. omies all sharesL,me more so than others And not even in the Department of Defense is expanding their pools of engineers and tech- could one find anything like an "electronics in- nicians. Countries like India, Taiwan, and dustry policy." Following World War H, U.S. Korea have labor forces containing substantial foreign economic policy centered on an ambi- numbers of engineers and scientists, many of tious recovery program in Western Europe and them educated in the West. Nevertheless, while Japanthe Marshall Plan. But despite this em- some of these nations have managed to mobi- brace of economic planning for other parts of lize their human resources more effectively the world, domestic economic policies have re- than others, none of the NICs have enough volved around macroeconomics and regula- skilled people t move rapidly into high-tech- tion. The United States has avoided promotion, nology electronics production: They do have planning, and targetingthe common tools in one advantage: their engineers are not paid other countries. nearly as well as in the advanced countries. In electronics, microlevel involvements, leav- With salaries perhaps one-quarter those in the ing aside national defense, have generally had United States, the industrializing Asian econ- regulatory thrustswitness the lengthy anti- omies are striving to capitalize on lower R&D trust prosecution of IBM. One reason the U.S. costs as they earlier did with unskilled labor." Government has been willing to endorse eco- While it is unlikely that any of these countries nomic planning overseas but not at home lies will quickly bridge the commercial and tech- in the unrivaled strength of American corpora- nological gaps separating them from Japan and tions _:wring most of the postwar period. In the West, and while theirnrroarbes to in- light of the success of American firms such as dustrial policy differ in ' ,uvernment in- Boeing, IBM, or General Electric in world mar- tervention and relianc: 1 market mecha- kets, the focus of policymakers here on free- nisms, all seem committee: to some variety of market competition is quite understandable; coordinated industrial policy as a means o5 for the Federal Government to consider poli- supporting local electronics manufacturers in cies that would promote "national champi- both domestic and world mark ons"as the French didwhen these champions already existed, would have seemed super- United States fluous if not counterproductive. The U.S. Government has not developed a Public policies have, nonetheless, exerted consistent, systematic set of policies directed considerable influence on the American elec- at industrya task that, even if judged desir- tronics industry. Military procurements stim- able, would be much more difficult for the ulated developments in computers and semi- world's most complex economy than one that conductors. Since the 1960's, trade policy has was still industrializing. It has become a com- been a persistent concern in consumer elec- monplace to note that, while numerous public tronics. Taxation, regulations of many kinds policies exert direct or ]tidirect effects on firms (particularly in the f-Aecommunications sector), and industries, the American approach is ad patents, protection of computer softwareall hut_ In this sense, then, U.S. industrial policy have been debated in various contexts. But in also differs from that in 1;:ipari or many of the total, the Federal Government's policies have European nations. While the Federal Govern- been a patchwork, often based on objectives 3°A. Spaeth. "Asian 'NICs' Rely on Cheap Brairraower To Plan quite different from those motivating the in- Output of More Advanced Gr)ds."Wall Street Journal. Jan.5, dustrial policies of other countries. Antitrust 1983. p. 25. enforcement stands out especially.

393 390 International Competitiveness in Electronics

Antitrust tronics are few. One case aro-ie at the end of the 1970's. when GE and Hitachi proposed a Where competition policies in other coun- joint venture to manufacture TVs in the United tries have been vehicles for mergers, joint ven- States. The two companies suspended their ne- tures, and consolidation, notably in the com- gotiations after the Justice Department threat- puter industrythe rationale being to create ened to sue under provisions of the Clayton companies big enough to compete effectively Act.33 More frequently, the possibilityeven if antitrust enforcement in the United States has remoteof costly and protracted litigation aimed it breaking up large enterprises.31 De- seems to have caused American firms to steer spite the common association of bigness with c!ear of cooperation in R&D.34 While much of badness. American law does not prohibit oli- the complaining by the business community gopoly (industries dominated by a small num- over antitrust reflects no more than theusual ber of firms), but limits predatory or exclu- antagonism toward Government regulation, it sionary tactics. Therefore, antitrust violations does appear that companies have been little intend to be difficult to prove, cases lengthy and clined to explore the bounds of the permissible, expensive.32 simply because the risks have been seen as far How has antitrust enforcement influenced greater than the rewards. Largely as aresult the international competitiveness of American of repeated expressions of concern, the Depart- electronics firms? As has been the case so often ment of Justice issued a set of writtenguide- with U.S. industrial policy, the side effects may lines covering joint R&D ventures, but a good have had the greatest impactin this instance, deal of ambiguity nevertheless persists.35 Even uncertainty over the intentions of the Depart- where no single project has great import, a ment of Justice and the Federzl Trade Commis- general discouragement of joint R&D efforts sion. Business and industry in the United could eventually have a large cumulative im- States claimperhaps, with justification that pact. Moreover, if joint international research antitrust enforcement is ambiguous and threat- projects proliferate, American antitrust lawin ening, that Government officials, knowing the the absence of more concrete guidancemay line to be vague, try to keep companies far present an obstacle to participation by U.S. back. Instances in which enforcement inten- firms.36 tions have been known toactually stop mergers, joint ventures, or acquisitions in elec- Trade and Foreign Economic Policies If antitrust has recently been at the forefront "On U.S. antitrust law and enforcement. see U.S. Industrial of U.S. industrial policies as they have affected Competitiveness: A ComParison of Steel. Electronics. and Auto- mobiles. op. cit., pp. 184-185. Also ch. 12 of the present report; J. W. McKie, "Government Intervention in the Economy of the "J. Crudele and J. Hataye, "Fear for TV Jobs as Justice Blocks United States." Government Intervention in the Developed Econ- GE-Hitachi Venture," Electronic News, Dec. 4, 1978, sec. I. p. 1. omy. P. Maunder (ed.) (London: Groom Helm, 1979), p. 75; and 34s..-ae, for example, D. H. Ginsburg, "Antitrust, Uncertainty, M. Keller, "Regulation of Large Enterprise: The United States and Technological Innovation." Antitrust Bulletin. winter 1979. Experience in Comparative Perspective," Managerial Hierar- p. 635. chies: Comparative Perspectives on the Rise of the Modern In- "Antitrust Guide Concerning Research Joint Ventures (Wash- dustrial Enterprise. A. D. Chandler, Jr., and H. Daems (eds.) ington, D.C.: Department of Justice. November 1980). At the end (Cambridge, Mass.: Harvard University Press. 1980), p. 161. of 1982, the Justice Department announced thai it woidd not seek "The Department of Justice initiated its suit against IBM in to bar the formation of Microelectronics & Computer Technol- 1969, with the trial beginning in 1975. A decisionwhich would ogy Corp., the joint venture involving a dozen U.S. firms in- certainly have been appealed regartass of the verdictwas still tended to pursue R&D in advanced electronics technologies (ch. well in the future when the case was dropped by the Govern- 5). ment in January 1982. At the same time, the Justice Department "For a proposal that foreign enterprises be allowed to partic- resolved a 7-year antitrust suit asking that AT&T divest itself ipate equally in the government-sponsored R&D efforts of all na- of Western Electric, the communications company's manufactur- tions, see Report of the U.S.-Japan Economic Relations Group, ing arm. On the settlements, see ''Statement of William F. Bax- January 1981, p. 80. Japan has recently agreed to open its ter, Assistant Attmt.,,y General, Antitrust Division, on Recent fifth-generation computer project, and others like it,,to Japanese Actions of the Department of Justice in U.S. v. AT&T and U.S. subsidiaries of U.S. companies. See U. C. Lehner. "U.S.. Japan v. IBM, Before the Committee on the Judiciary, U.S.Senate," Pact Would Bolster Joint Research," Wall Street Journal, Nov. Jan. 1982. 1, 1982, p. 35.

39,1 Ch. 10National li,dustrial Policies.= 391 the computer industryvia the IBM case Because the military market is now so small trade policies concerned with dumping and compared to commercial sales, specialized other unfair practices have been central in con- contractors do much of the work on devices sumer electronics. Trade policies and their ef- for defense systems. Largely in response to the fects ale treated in detail in chapter 11; the slow rate of introduction of advanced micro- point here is simply to note their significance electronics technologies into military hard- as part of U.S. industrial policy. After years of ware, the Department of Defense initiated an litigation, the competitive battle in color TV is R&D program directed at very high-speed in- still proceeding in the courtroom as well as the tegrated circuits (VHSIC) beginning in fiscal marketplace. A legalistic thrust analogous to 1979. With an initial 6-year budget of more than that in computers has dominated public policy $200 millionsince expanded substantially impacts. the VHSIC program is intended to speed the development of ICs that meet military needs. To take a somewhat broader perspective, as Involving all three services, VHSIC has been world competition in electronics has increased, structured around bidding by firms and groups U.S. policymakers have renewed their attempts of firms for contracts covering a variety of well- lo reduce overseas trade barriers. Nontariff and defined R&D tasks. Although the ICs them- indirectbarriersrestricting the entry of selves will be tailored to military applications, American products into foreign markets have research results in areas such as processing been particular targets. A new flurry of activity technology, computer-aided circuit design, and came in 1982; the many bills introduced in system architectures will find their way into Congress that could be loosely grouped as deal- the commercial efforts of U.S. merchant firms. ing with trade reciprocity illustrate the depth While most of the VHSIC contracts are closer of concern. Progress on such questions will be to development than basic reseai--.;h, the De- slow; since most countries view subsidies and fense Department has also initiated a program other tools of industrial policy as internal mat- entitled Ultrasmall Electronics Research inters, they are difficult to address via interna- tended to support R&D that will pay off 10 or tienal negotiations. 20 years in the future.38 Procurement and R&D Even with the increases stemming from the VHSIC program, Federal support of R&D in In contrast to the antitrust and trade orien- semiconductor-related technologies remains a tations visible in computers and consumer elec- much smaller fraction of total U.S. semicon- tronics, American semiconductor firms have ductor R&D than in tit- 1960's. While the com- seldom, since the 1960's, been directly affected parisons are less thal:. -..:taightforward because iy public policies. Through the 1950's and allocations of spending to R&D categories tend 1960's, the Federal Government stimulated de- to be rather arbitrary, and disaggregated data velopments 'n microelectronics by purchasing seldom available, an idea of the current sig- semiconductors for military and space pro- nificance of Federal funding can be pieced grams, as well as by supporting R&D (much the together. same was true for the computer industry in its early years). During this period, the Govern- For 1980, the latest year for which data is ment purchased a large fraction of U.S. semi- available, total U.S. R&D spending by the "elec- conductor outpute.g., for the Minuteman II tronic components" sectorwhich is con- missile. In 1965 the Department of Defense ac- siderably larger than microelectronics alone counted for about 70 percent of U.S. IC sales, has been put at $1.354 billion.3° For the same while by the end of the 1970's, the figure had °The 5-Year Outlook on Science arid Technology 1981 (Wash- dropped to around 7 percent.37 ington. D.C.; National Science Foundation NSF 81-10. 1981). p. 33. "Electronic Markel Data Book 1982 (Washington, D.C.: Elec- "An Assessment of the Impact of the Department of Defense tronic Industries AssoCiation. 1982). p. 121. The figure. from data Very-Nigh-Speed Integrated Circuit Program, National Materials collected by the National Science Foundation, is for SIC category Advisory Hoard Report NMAB-382 (Washington. D.C.: National 367. which has nine subdivisions. Of these, semiconductors (SIC Research Council. January 1982), p. 6. Also see ch. 4. 3674', is 4-ertainly the largest performer of R&D. 395 392 International Competttivenesc tn Electronics year, tabulations of R&D spending by U.S. mer- Evidently, then, the Federal Government con- chant semiconductor firms from sources such tributes something between 5 and 10 percent as annual reports give totals in the range of of the total. This estimate illustrates the con- $800 million. It is more difficult to determine tinuing decline in the Federal presence; over spending on microelectronics by captives, the period 1958-76, Government spending ac- which seldom- re-port such data separately. counted for about 15 percent of all U.S. semi- Allocation of software development costs also conductor R&D." leads to ambiguity; as microelectronic devices Indeed, it appears that even in the early become more complex and more like complete years, Government purchases were a greater systems, softare becomes a major part of the spur to the industry than R&D contracts.42 By research, design, and development effort. providing a guaranteed market, Government In any case, given that IBMlargest of the procurementmostly for military purposes captive producersno doubt spends several stimulated the growth of the industry at a hundred million dollars annually on microelec- critical stage in its development. At the time, tronics, total U.S. R&D expenditures on semi- semiconductor manufacturing was a far differ- conductor-related technologies in 1980 must ent business than today; it was part of the have been well over $1 billion. How much of defense sector of the economy, whereas sales this did the Federal Government provide? For to the Government are now dwarfed by sales fiscal year 1980, Government expenditures for to computer manufacturers and other nonde- R&D related to ICs have been reported as $61 fense customers. million, rising to $71 million in fiscal 1981.4° Taxation '°An Assessment of the Impact of the Department of Defense l-ery-High-Speed Integrated Circuit Program, op. cit., pp. 20-22. The Economic Recovery Tax Act of 1981 Fur purposes of this rough comparison. R&D related to ICs can (ERTA) was supposed to speed economic be taken as equivalent to R&D related to microelectronics. The Federal contribution includes work performed in Government growth and build U.S. competitiveness by in- laboratories. but this accounts for less than 10 percent of the creasing incentives for saving and investment.

totalthe rest being contracts and grants to industry. universities, and independent research laboratories. About 10 percent of the Government money comes from_the National Science Foundatioh (NSF), most of the rcst from the Department of De- fense. NSF's share of basic research support is closer to 30 percent. In fiscal 1980. the VHSIC program accounted for 40 percent of the Government's total spending on microelectronics R&D. As table 77 indicates, overall R&D spending by the.U.S. Government is heavily skewed toward military needs compared to countries like Japan or Germany. IBM spends well over $1 billion annually on R&D: the company's R&D spending on very large-scale ICs has been reported to total about $1 billion over the period 1977-80G. Gregory. "The U.S. Wages Micro -War," Far Eastern Economic Review, Mar. 16, 1979, p. 124. "A Report on the U.S. Sem4...mductor Industiy (Washington.

*Is D.C.: Department of Commerce. September 19N9). p. 8. The estimates are those of the Semiconductor Industry Association. In 1038. Department of Defense contracts and grants accounted for nearly a quarter of the industry's R&D spendingN. J. Asher and L. D. Strom, "The Role of the Department of Defense in the Development of Integrated Circuits," Institute for Defense Anal- yses paper P-1271, May 1977, p. 3. The percentage has thus been falling more or less steadily for many yea!!;. "Asher-and Strom, op. cit.; J. M. Utterback and A. E. Mur- ray, "The Influence of Defense Procurement and Sponsorship Phob croon:Pr:el Corp. of Research and Development on the Development of the Civilian Electronics Industry," report CPA-77.5, Center for Policy Alter- iy 16-bitmicroprocessor natives: Massachusetts Institute of Technology, June 30, 1977. Ch. 10National Industrial Policies 393

The wholesale changes in U.S. tax policy em- macroeconomic policy have taken priority bodied in ERTA have affected all industries; competitiveness and economic efficiency have perhaps most significant are the altered depre- Seldombeen at'the forefront, or even in view. ciation schedules discussed in chapters 7 and Trade policy complaints in consumer elec- 12. ERTA also extends a tax creditamounting tronics have come from domestic firms and to 25 percent of any increase in R&D spending their employees, with Federal agenciesill- over a base figureas part of a package of in- equipped to take an independent viewreact- centives for research. Although young high- ing to these pressures. Short-term response to technology companies may not have profits to political pressures has in fact been the com- set against the tax credit, at least some elec- mon denominator of U.S. industrial policy. tronics companies will benefit from the R&D Yet as competition has intensifiedin com- provisions more than from accelerated depre- puters and microelectronics, jet aircraft and ciation. telecommunications systemsboth Congress As pointed out elsewhere, the ERTA package and the executive branch have begun to debate has thus far had little perceptible effect on in- the question of a more explicit industrial policy vestment. Moreover, it appears that, in com- for the United States.'" In addition, the Depart- parison with other U.S. industries, the relative ment of Defense through VHSIC and other pro- attractions of investments In electronics may grams, the National Science Foundation, and have been diminished. The telling point in the the Department of Commerce have all stud- context of U.S. industrial policymaking is this: iedeven attempted to design, often under such outcomes have been neither intended nor rubrics such as innovationpolicies that would anticipated, instead resulting from the unex- stimulate basic as well as applied research, and amined give-and-take ofthe political process.

Industrial Policymaking 4,To give only a few examples, and leaving aside such related topics as innovation or productivity, late in 1980 the Subcom- As many of the examples above indicate mittee on Trade of the House Ways and Means Committee issued from antitrust through taxation (many others the United States-Japan Trade Report (Sept: 5, 1980), calling for could be adduced)public policies influencing overall improvement of the economy rather than trade protection as a response to Japan's growing challenge in high-tech- the American electronics industry have lacked nology industries. A report by J. GresserHigh Technology and a framework and sense of direction. The very Japanese Industrial Policy: A Strategy for U.S. Policymakers (Oct. notion of objectives or "goals" for policy, in 1, 1980)recommending a more focused U.S. response was published soon thereafter under the auspices of the same commit- any but the most immediate sense, has been tee. The Subcommittee on Trade's Report on Trade Mission to anathema for policymakers here. In contrast, the Far East (Dec. 21. 1981) reiterates many of the same themes. other countries have pursued economic devel- More recently, the Joint Economic Committee his released a study by M. Borrus, J. Millstein, and J. Zysman entitled interna- opment quite consistently, making use of nu- tional Competition in Advanced Industrial Sectors: Trade and merous policy tools. While in the United States Development in the Semiconductor Industry (Feb. 18, 1982) there has beey no one agency to serve as a focal ,which stresses the importance of electronics for overall economic development. point for Industrial policies, other nations have Dozens of hearings in Congress over the past several years developed policymaking approaches involving have covered such issues, two examples being Industrial Poli- more or less permanent industrial advisory cy, hearing, Joint Economic Committee, Congress of the United States, May 18, 1982; and U.S. Industrial Strategy, hearing, Sub- councils, ministries accountable for well- committeeon Economic Stabilization, Committee on Banking, defined policy areas, mechanisms for coordina- Finance and Urban Affairs, House of Representatives, Sept. 22, tion. Here, many agencies participate in policy 1982. Trade and tax policy debates inside the Reagan adMinistra- developmentsometimes on a regular basis, lion have dealt at least peripherally with electronics, as has an sometimes infrequently. interagency study on high-technology tradesee An Assessment of U.S. Competitiveness in High-Technology Industries (Wash- More often than not, policies affecting elec- ington, D.C.: Department of Commerce, February 1983). The tronics have been formulated with little con- Commerce Department has begun work on an inventory of industrial policy measures employed by other countries, while he sideration of possible impacts on international Department of Labor has a long-standing interest in industrial competitiveness. National security, antitrust, policies, particularly as they deal with adjustment.

99-111 0 - 83 - 26. 3 9 394 International Competitiveness in Electronics encourage productive investments by the pri- and CGE (Compagnie Generale d'Electricite) vate sector. There is no dearth of concern over have joined the roster of national enterprises, policies affecting industries like electronics, together with a number of banks. The goal has but little consensus as yet on the direction that been not only to increase the financial re- 'policy initiatives should take." sources and market power of French corpora- Chapter 12 addresses policy alternatives for tions, but to create prestigious flagships that the United States in some detail; here the point can' ead the economy. Saving jobs in threat- is thatin contrast to ongoing debates over in- ened industriese.g., steelhas also been an dustrial policy in other countries, which tend important motive; furthermore, the govern- to focus on review and redirection of measures ment's plan for the electronics industry prom- already in placethere is still no consensus in ises to create 80,000 new jobs over the 5-year this country on the need for a more coherent period 1982-86.45 "National champion" firms industrial policy, much less on the form it were a capstone of French industrial policy might take. In a sense, the United States is start- during the 1960's, when France became the ing off behind in the race to develop effective first nation to mount a direct challenge to IBM. industrial policies simply because U.S. in- Aircraft, nuclear power, and telecommunica- dustries like electronics led the competitive tions have been other government .favorites. The idea of national champions never really race for so many years. died, and has simply been revived in slightly France different form under Mitterrand; electronics computers, semiconductors, consumer prod- Perhaps more than any other advanced West- ucts, communications, office automationis to ern nation, France has centralized and coor- be at the core of France's ftture industrial dinated its industrial policymaking as one of policy. the primary ingredients in an interventionist approach to economic policy. While the tools Policymaking mechanisms in Francecen- and tactics have shifted over time, the policies tered on the ministries of Industry and of adopted by the current Socialist Government Economy and Financediffer greatly from trace their origins to the planning process those in the United States, as might be expected adopted by France in the aftermath of World in a country where the idea that the state can War II. The continuity of the French system, and should play a role in industrial develop- like that of Japan, is one of its salient charac- ment has been widely affirmed. In the policymaking system that has evolved, the Ministry teristics. of Economy and Finance takes the lead in The Setting channeling funds to favored sectors (ch. 7), while the Ministry of Industry is more heavily. The French have accepted government in- involved in day-to-day matters, as well as tech- volvement in the economy as legitimate and nology and microlevel planning. Within the necessary. Industrial policies are part of a con- Ministry of Industry, the Directorate of the text that includes extensive public ownership Electronics Industries and Data Processing is of both manufacturing organizations and finan- responsible for efforts such as the Government cial institutions; under Mitterrand, the elec- Program. for Development of Electronics, an- tronics firms CII-Honeywell Bull, Thomson, nounced late in 1982. Since the Socialists took 440TA's comparison of U.S. competitiveness in three industries power in 1980, the Ministry for Research and led to the suggestion of a "macroindustrial" policy. The intent Technology has taken a larger role in industrial would be to provide infrastructuTal support for American indus- policynot only the, design of policies for tries, rather than moving toward, explicit Government decisions favoring some sectors over othes. Examples would be policies high-technology industries like electronics, but directed at labor markets, technological development, human resources. taxation, and economic adjustment. See U.S. Indus- " "Government Funding for Electronics Industry Discussed." trial Competitiveness: A Comparison of Steel, Electronics, and West Europe Report. Science and Technology, No. 118, Joint AUtomobiles, op. cit., pp. 157-165. Publications Research Service JPRS 81678, Aug. 31, 1982, p. 3.

396 Ch. 10National Industrial Policies 395 also restructuring elsewhere in the French government and industry. Business has been economy. Broad 5-year economic plans con- able to operate within a fairly predictable tinue to be part of the policymaking process, context. although they have receded into the background compared to 30 years ago. Finance The staffs of French ministries tend to share The French financial system, like the plan- similar educational backgrounds, typically the ning mechanism, enhances government influ- prestigious grandes ecoles. Not only the public ence over economic development. Capital al- sector, but large industrial enterprises as well, locationssee chapter 7are controlled to con- are managed by a small and homogeneous siderable extent by administrative fiat rather elite; the Socialists placed at the head of na- than market forces. A rather small number of tionalized firms by the Mitterrand Government financial institutionscloselytiedtothe are in most respects indistinguishable from the bureaucracy whether or not actually national- men they replaced. The closed nature of this izedlink government policymakers and com- system helps the French bureaucracy wield panies seeking funds. The Treasury determines authority more like that granted public officials interest rates on bonds; through the Ministry in Japan than in the United States. Distinguish- of Economy and Finance, as'well as a variety ing features of French industrial policyan of semipublic lenders and the banks, the emphasis on sectoralmeasures, perhaps government can exert considerable leverage stronger even than in Japan, and the encour- over credit decisions. Specialized institutions agement of corporate. consolidationreflect such as the Institut de Developpement In- not only the power of the bureaucracy but the dustriel (IDI), funded from both public and community of interest binding industry and the private sources, provide risk capital to medium-. state. sized firms; IDI has also made equity investments in the computer firm Compagnie Inter-. Planning nationale pour l'Informatique (CII). Even in light of the French Government's traditional Much has been written on indicative plan- use of finaritial channels, Mitterrand's invest- ning in France, which can be traced to the im- ment plans for electronics are extraordinarily mediate postwar period and the Monnet Plan ambitious. The industrywhich is now rough- for rc,.onstruction. As the term "indicative" ly half nationalizedis to invest $20 billion suggests, the country's 5-year economic plans over the period 1982-86, with the government have not been imperative, but based on con- providing about 40 percent of the total." It is certed actions mutually agreed on. In earlier not clear where the money will come from. years especially, officials in the Planning Com- mission played key roles in bringing together leaders in government and industry. Le Plan Calcul A major function of the planning process has French policies as they have affected elec- simply been to gather information about past tronics have been shaped, as elsewhere in progress and future prospects by sectors in the Europe, by historical circumstancei.e., the economy. The hard decisions have been made relative weakness of French industry com- elsewhere, with the role of the Planning Com- pared to American corporations. The result mission largely facilitory. While the rapid during the 1960's was a concerted thrust in postwar recovery of the French economy can- computers known as Le Plan Calculnot not be attributed solely to planning and in- unlike what the French are now undertaking_ dustrial policy, the planning exercise has in electronics as a whole. helped crystallize perceptions among the bureaucracy, as well as decisionmakers in "Ibid. The 5-year investment plan calls for the government private industry, creating a shared referent for to provide 55 billion francs of the 140 billion total.

393 396 International Competitiveness in Electronics

By the early 1960's, the enormous strength against foreign competitors, and guaranteeing of IBMcombined with the comparative weak- domestic procurements for the company's ness of European firmswas perceived as a productsto enable the firm to challenge IBM. serious threat to the viability of the French elec- A related series of measures over the late 196e's tronics industry. It was the "American chal- and early 1970's comprising Le Plan des Com- lenge"viewed as a technological lead, but in posants (Plan for Components) was to help reality just as much commercial superiority with the development of semiconductor de- that stimulated an ambitious effort by the vices, primarily for computer applications. French.'" The well-known Plan Calcul came in Despite the support provided CII, the firm 1966, on the heels of serious difficulties for the never approached its targets. American com- French computer industry. In the "Affaire panies continued to dominate sales in France, Bull" of 19Ril, the American firm General Elec- and CII's chief market turned out to be the tric had purchased Machines Bull, a faltering government. By 1975 Le Plan Calcul had effec- French computer manufacturer. At about the tively been abandoned, as a variety of factors same time, the U.S. Department of State re- combined to defeat the best efforts of French fused to grant export licenses for two of Con- policymakers (who now insist that their efforts trol Data's largest processors. These were to at least prevented further erosion of the na- have been used in the development of fusion tion's indigenous capabilities). Some critics em- weapons; the refusal helped convince French phasize the contradictions inherent in a pro- policymakers of the, need for an independent tective strategy within a highly competitive in- computer industry. Since then, if not before, dustry, and a policy designed and implemented the French military, although taking some care by technocrats with little experience of com- to stay. in the background, has had a major say mercial realities." Hindsight shows the effort in industrial policy .decisions affecting elec- to have been overambitious, an attempt to con- tronics and telecommunications. front American firms across a broad line of Le Plan Calcul was intended to build an in- products rather than in selected niches. In this dustry capable of challenging IBM; to do this, sense, national goals took precedence over ,bureaucracy engineered the merger of two sound business strategy. Finally, the money existing manufacturers, forming a new public that the French Government pumped into CIL,- corporation CII. The government provided perhaps $350 million between 1966 and 1976 capital to the fledging champion, but as the looks rather insignificant next to; say, IBM's product of a union between two firms which resources." together held no more than 7 percent of the Into the 1970's, then, French policy toward French computer market, the new company the electronics industry centered on one com- had a long way to go. panyCII. By 1975, when the failure of Le Plan CII's efforts were directed first and foremost 'Calcul was clear, the government encouraged at medium to large mainframesthe rise of the CII to merge with Honeywell Bullthe de- minicomputer was just beginning and had not scendant of Machines Bull that emerged from been widely recognized. CII was to be an ex- the sale of General Electric's computer busi- port leader, as well as providing for France's ness to Honeywell. The new company, CII- own needs, of which national security was at Honeywell Bull (CII-HB), was majority French- the forefront. Although CII was attacking IBM owned; it quickly received further government at the latter's point of greatest strength, French __planners hopedby providing export and_other 44J Zysman, Political Strategies for Industrial Order (BerkoleIV-- subsidies, encouraging shipments to the Soviet Calif.: University of California Press, 1977), p. 99. bloc and developing countries, protecting CII 4°The $350 million estimate is from Technical Change and In- dustrial Policy: The Electronic Industry (Paris: Organization for "This interpretation of the "challenge" is elaborated by N. Economic Cooperation and Development, 1980), p. 46. Jequier Jequier, "Computers," Big Business and the State. R. Vernon (op. cit., p. 217) gives a figure of $120 million between 1966 and (ed.) (Cambridge, Mass.: Harvard University Press, 1974), p. 193. 1970. ,Ch. 10National Industrial Policies 397

assistance totaling perhaps $700 million.5° In Frenth partners. France would get technology, 1982, after prolonged discussions, the Mitter- the U.S. participants access to the FrenCh mar- rand government effectively nationalized CI l- ketparticularly the lucrative telecommunica- R, which will become a subsidiary of a gov- tions sector, well protected by the PTT, These ernment-controlleafriolding company taking up joint ventures, in which the French partners the old name Machiiies Bull." Machines Bull held controlling interests, tied Thomson to will be the centerpiece of Mitterrand's com- MotorOla, Saint-Gobain to National Semicon- puter thrust, discussed below, with CII-HB ductor (in a firm named Eurotechnique), and responsible for mainframes. Matra to Harris. In addition, the plan sup ported two more firms: Radiotechnique, a Le Plan des Composants Philips subsidiary in France, and EFCIS, orig- Recognition that CII -HB, needed infusions of inally owned by the French atomic energy au- semiconductor technology led to a new 5-year thority (Thomson purchased a majority interest, components plan in 1977. Military require- in EFCIS in 1977). ments were also a strong motive. Since the Le Plan des Composants was developed at 1950's, France had maintained a small but a time when France's Government was redis- high-quality semiconductor research effort. covering market forces. Attempting to learn However, this had never been translated into from Le Plan Calcul, French planners decided, a commercially viable merchant industry. By to support a number of firms. Rather than fun- the early to mid-1970's, perhaps a hundred. nel the money set aside for the program to a French engineers and scientists were engaged single champion, the five companies would in R&D on advanced microelectronic devices; compete with one other. Although an element the country did not have the capability for of competition was thus built in, each partici- mass-producing ICs. French engineers had lit- pant was assigned certain technologies in tle background in microprocessors, nor in which it was to take the lead. Matra-Harris, for MOS (metal oxide semiconductor) ICs, most of example, would specialize in c-MOS since this their expertise being in bipolar devices for com- was Harris' strength; later the joint venture munications and consumer products. Le Plan negotiated a further agreement with Intel, des Composantsdesigned by the ministries largely to gain the latter's n-MOS technology. of Industry and Defense, plus the PTT (respon- ICs were new technologies for both Matra and sible for postal services and telecommunica- Saint- GobaLi, which were picked for the pro- tions)was intended to rectify these deficien- gram in part because of their success in other cies. In contrast to Le Plan Calcul, France did fieldsin the case of Matra, its high-technol- not attempt to keep foreign participants out, ogy experience in aerospace was a particular but sought to build on American technology. attraction. Saint-Gobain-Pont-a-Mousson, a ma- In this way, French planners hoped to move jor producer of glass and chemicals, had de- toward self-sufficiency in microelectronics, cided of its own accord to diversify into elec- with the eventual goal of a major share of the tronics; in addition to participating in Le Plan European market. des Composants, the company purchased substantial interests in CII-HB and the Italian com- The vehiCles included three joint ventures puter and office equipment firm Olivetti dur- linking American semiconductor firms with ing this period. More recently, the French have decided that, if one national champion is too "Technical Change and Industrial Policy: The Electronic In- few, five are too many; since the Mitterrand dustry, op, cit., p. 46. Other estimates have ranged as high as $1 billion. Prior to' the merger with Honeywell Bull, C11 had been government came to power, extensive-discus- a participant with Philips and Siemens in the European con- sions aimed at consolidation have been under- sortium. Unidata. The consortium did not prove workable. way., Three centers of excellence in microelec- 51:C11-Honeywell Bull Announces Restructuring in Line With French Plans for Computer Firms." Wall Street Journal, Dec. tronics seem likely to emerge. Both Matra and 21, 1982. p. 30. Honeywell's interest has been reduced to about Saint-Gobain have been nationalized, with 20 percent. Saint-Gobain evidently forced out of elec- 398 International CompetitIven&: mics tronics.52 Eurotechnique has been sold to such as videotext7that French planners em- Thomson, which purchased the shares of both barked on in the .rnid-1970's; the PTT's am- Saint-Gobain and National Semiconductor, bitious projections 'envision 25 million ter- while retaining a technology exchange agree- minals in French homes by 1990, pointing ment with the American firm. toward a rapidly growing market fOr semiconductors. As part of its telecommunications The R&D portion of!.9Plan des Composants, policy, the government has forced the sale oof known as Le Plan Circuits Integres, channeled two foreign-owned companies (subsidiaries of aboUt $150 million in government funds as di- ITT and Ericsson) to the Thomson group. rect 'grants and loans to the five companies. The nioney supported work on very large-scale Recent Developments ICs, ranging from circuit design to the development of 'processing equipment. Research cen- French industrial policy has been in some- ters, inClUding the Electronics and Informatics thing of a turmoil since Mitterrand's election. Teclincloi.Laboratory of the French Atomic The outlines of the Socialist Government's pro- Energy Colnission, were strengthened, while gram remain murky, although the intent is to Le Plan Circuits Integres also supports micro- emphasize electronics. Initiatives in semicon- processor ap\ilications through a new Informa- ductors, computers, communications, and con- tion Agency.53 sumer products are likely to be even more tightly coordinated than in the past. And, while the It is 107-) earl, to judge the success of Le Plan themes of nationalization and merger policy dos C. .-Josants in building a viable commer- predate Mitterrand, the Socialists have carried cial industry, b6t in terms of technology French this aspect of French industrial policy still semiconductor\ firms have mane great prog- further. ress. Eurotechnique manufactured its first ICs at the end of 1980 and has since expanded out- Even beforeMitterrand came to power, the put at a high rate. EFCIS's production of ad- Eleventh Five-year Plan (1981-85) had targeted vanced devices 1?egan about the same time. De- electronics for special support. The plan sin- spite rapid incrOses in production,_ however, gled out six fields for massive government the French entrants remain small on a world assistance, with electronicsranking third in scale (see ch. 4, table 32), suffering from thin French exports of manufactures, after machine product lines aind limited distribution net- tools and chemicalsviewed as a critical works. Still, the technical know-how they have sector." Under the plan, total R&D expendi- acquired from Atnerican firms places them in tures in France are scheduled to increase to 2.5 Elthiantageous positions compared to other percent of GNP by 1985. Currently, France is European semiconductor manufacturers. making a more concerted effort than any, other European country to strengthen its technolog- In recent years, the French bureaucracy has ical base and promote high-technology in- also given a gobd deal of attention to minicom- dustries, with considerable attention to train- puters and peripherals through Le Plan Peri- ing greater* numbers of engineers and tech- nformatique. Moreover, the components pro- nicians. Le Plan des Composants indicated that vam has been linked to a major push into tele- the French had learned from the mistakes of communicationsincluding developments Le Plan Calculand also from the commercial 52See "Possible Strategies fir Executing Microelectronics failings of the Concordewith French indus- Plan," West Europe Report, Science and Technology. No. 112. trial policy as it affects electronics and other Joint Publications Research Service JPRS 81340, July 22, 1982, high-technology sectors passing into a new Absorbs-Eurotechnique-Financial stage ;nne-marked by-a-more suphisticated-un Times, Jan. 21, 1983. p. 14. . ""190 Million Francs in Next Five Years for VLSI Research," West Europe Report, Science and Technology, No. 89, Joint Pub- "Rapport de la Commission Industrie, Commissariat General lications Research Service JPRS 80022, Feb. 3, 1982, p. 7; "Le du Plan, July 1980. p. 48. According to this report, electronics Developpement des Applications de L'informatique," Lettre 101, has received about 10 percent of all direct sector-specific aid Oct. 7, 1980. to French industry in recent years (p. 113).

el 0 Ch. 10National Industrial Policies. 399

derstanding of international competition in larger effort aimed at strengthening the entire commercial Products and technical develop- French electronics industry. ments. This shift began during the 1978-80 While the overall outlines remain vague, ttie period. tinder Prime Minister Barre, the gov- government is promising that investments in ernment 'claimed to be "decontrolling" mar- electronicsfrom both public and private kets, fo: instance cutting back on price con- sources, and including investments by foreign- trols. With a deemphasis on planning and na- owned firms (IBM, Texas Instruments, and tional champions, came more stressat least Motorola are among the American electronics in official rhetoricon market forces. Is Mit- companies with a major presence in France) terrand likely to reverse this trend? will total $20 billion over the 5-year period A fundamental plank in the Socialist plat- 1982-86. The Government Program for Devel- form had been nationalization of companies opment of Electronicspresented in Sep- like CII-HB. As Mitterrand himself explains the tember 1982 after an extensive study by an Socialist strategy, the object is first to win back Electronics Industry Task Force.is .to be coor- the domestic market in key industries such as dinated by an Interininisterial Committee for steel, machine tools, semiconductors, and Electronics, with representatives from the min- small computers.55 In conjunction with further istries of Industry and Defense, the Plan, and nationalization in the financial sector, the an- the PTT. nounced philosophy was "flexible" national- A primary vehicle will be 9 "national proj- izationwith the government providing con-. ects," chosen from 14 originally recommended siderable support while promising to eschew by the Task Force. These national projects, extensive involvement in business affairs or economic planning at the micro-level. In ap- which will get extensive government support, pearance, this is not a sharp turn from the past; are intended to link private and nationalized despite, the lengthy history of planning in firms, as well as the labor and user Comm,. France, nationalized firmsso long as they ties. The nine projects have the follou have performed adequatelyhave operated rel- titles:57 atively free of direct intervention by the bu- consumer electronics; reaucracy. information displays; local networks; In R&D and technology development the new cable TV net, government has also moved ahead in bold if very large-sc,lie ICs (fabrication as well as seemingly disorganU.od fashion. Research sup- design); port has been increased under the current central processing units for small comput- 5-year plan, and is to include a new microelec- ers; tronics project as a fellow -on to Le Plan Cir- computer-assisted education; cuits Integres, plus more money for computers computer- assisted engineering; and and data processing. The government expects computer-assisted translation. to put up two-thirds of the $500 million it believes must be invested in microelectronics The list is noteworthy for emphasizing com- over the 1982-86 period.56 As in the past, much puter systems from the perspective of user of the money will come from the Ministry of needsnot only the last three projects, but also Defense. And as also in the past, the new mi- that on local networks. All are Software-inten- crioelectronics plan is but one piece of a much ""Mitterrand: Why Nationalization Will Work.- Wall Street "See R. T. Galkgher. "$20 Billion for French Electronics." Journal. Oct. 7. 1981. p. 27. Electronics. Sept. 8. 1982. p. 104: "Fourteen Projects." West 54"Microelectronics Plan: Win Market. Technology Independ- Europe Report. Science and Technology. No. 116. Joint Publica-. ence." West Europe Report, Science and Technology, No. 113, tions Research Service JPRS 81575. Aug. 18. 1982. p. 14. Among Joint Publications Research Service If' "S 81392. July 29. 1982. the five that were droppednot necessarily permanently-7-was p. 10. a supercomputer effort.

403 400 international Corripetitivene,3S in Electronics

sive, a field in which France ;s I n a relatively Calcul must be termed a failure, although CII- good position. The plan does not neglect con- HB's troubles had multiple sources. In semi- sumer products. France hopes to increase its conductors, Le Plan des Composants seems to presence in consumer electro.hus markets have functioned much better. Even so, the throughout Europe, with Thomson moving ag- largest French producerThomsoncontrols gressively into new generations of products like only a quarter of the domestic market, with a VCRs, electronic toys and games, and home inmarket share in all of Europe that is perhaps formation systems." 7 percent. Most of Thomson's sales are ;n The Socialist Government faces severe obsta- discrete semiconductors; the company has no cles in implementing such a vast program. In more than about 2 percent of the European IC addi ion to the $20 billion in planned invest- market. Although Thomson appears to have ments, a considerable increment compared to benefited considerably from technology- recent expenditures within theindustry, assistance agreements with Motorola, as have France has an inadequate supply of men and Matra arid Eurotechnique through their joint women with training and skills in .electron;cs; venture with American partners, French elec- the shortfall is reckoned at more than a thou- tronics firmsalong with most European man- Sand engineers yearly and at least three times ufacturersremain heavily dependent on as many technicians. The development plan. foreign sources of MOS and microprocessor contemplates an extensive training effort, in- technology. cluding the establishment of several new The history of French electronics policy schools. Moreover, foreign firms with invest- shows that strong government direction can- ments in France may resist some elements of not by itself produce a competitive industry. the program. Joint venture participants, for in- At the same time, the French seem to be learnL. stance, could prove less willing to transfer tech- ing how to make their electronics policy func- nology when the partner is a nationalized con- tion more effectively. cern. But in the end, money will probably be the limiting factor; boosting France's R&D rx penditures from 2.percent of GDP hi 1982 to United Kingdom 2.5 percent by 1985 is extraordinarily ambi- In contrast to the French, with their reliance tious. And, with nearly three-quarters of R&D on centralization and government action, Brit- carried out in government-controlled inFtitu-, ain's industrial policy has been closer to that tions, France runs a real risk of stifling innova- of the United Stateslargely ad hoc, not well tion and new ideas. coordinated. There is at least one major difference: the United Kingdom during the 1970's Future Prospects began to experiment with a variety of novel While the hallmarks of French industrial pol- measures intended to directly affect the actions icy remain the.samean elite corps of officials, of industry. Ranging from programs to encour- centralized policymaking, and a preference for age applications of microprocessors to govern- sectoral policy along with a tradition of state !tient investment in the semiconductor venture intervention in the affairs of industryMitter- Inmos, these initiatives are far different from rand's philosophy does represent 1 turn away the arms-length approach to industrial policy from the market orientation of Gi:;card d'Es- of the United States (U.S. policies related to na- taing. It is too soon to assess the effectiveness tional defense are, as usual, the exception). At of the new avenue, but past results give some the same time, these policies some of which insigh-t§.-Gwvernment-elfo-rts under Le-P dlt a tt racted- considerable-attention-i n-other-parts of the worldwere pursued withlittle sense of direction. Only in its sup, Jrt of Interna- ""First Details Published on Electroni::s Plan." West Europe Report. Science and Technology. No. 12a Joint Publications Re- tional Computers Ltd. (ICL) through procure- search Service JPRS 81904, Sept. 20, 1982, p. 7. ment practices, R&D funding, and other con- Ch. 10National Industrial Policies 401

vadtional policy tools has Britain shown much The Labor Government which came to pow- consistency over the longer term in policies er later in the decade continued this general toward elearonics. orientation, and picked up the pace by establishing a number of Economic Development Measures to aid ICI. in several respects shn- Committees to deal with specific industries. ilar to the somewhat earlier French effort to The Electronic Development Committee, set up build and strengthen CII-HB, date from the late in 1964, produced a series of reports that iden- 1960's. In the latter part of the 1970's, the tified problefns and proposed strategies for United Kingdom's electronics policies, as in overcoming them: But by the end of the dec- many countries, turned .owards semiconduc- ade, the planning experiment had run afoul of tors; a group of programs-Ire developed to persistent conflicts with the macroeconomic promote IC technolugy and pplications. Even policies that Britain's leaders were determined so, neither today nor at any point over the past to pursue; economic, planning came to be decade does the British example show much viewed as a failure, and the visibility and in- evidence of a coherent view of industrial pol- fluence of the Economic Development Com- icy.. mittees waned.°0 Early Experiments The More Reconfi Context Certainly Britain has had ample incentive to Since the beginning of the 1970's, U.K, in try new approaches; since the early 1960's, pol- dustrial policy has been a Bodge-podge. As in icymakers in the United Kingdom have sought ill( 'United States, consistency has been found ways of grappling with the nation's lackluster moutly in the area of national defense. A host economic performance by most measures the of government offices, themselves subject to poorest among industrialized nations. During periodic reorganization and changes in direc- the 1950's, macroeconomic policies had been tion, have been involved in policies affecting assumed sufficient for, revival. Rut continuing Britain's electronics industry. The National inflation, along with persistent wage disputes, Research Development Council, set up as early convinced the ruling conservatives to move to- as 1.946 to provide financial support for joint ward a more active government role. The Na- research ventures under the Ministry of Tech- tional Economic Development Council (NEDC) nology, is one example. The Ministry of Tech- was established in the early 1960's as a forum nology also had jurisdiction over the Industrial where business, labor, and government could Reorganization Corp. (IRC), establi..,ned in 1966 air their ideas about the future direction of the to aid industrial restructuring. Under the au economy.5 Inspired by the prestigious French thority of the Industrial Expansion Act, the Commissariat du Plan, NEDC was empowered Ministry of Technology engineered the merg- to produce 5-year plans intended to reduce ers creating the computer firm ICL, asdis- uncertainty about the directions of government cussed in more detail below. The IRC likewise economic policy. Planning responsibility fell provided financial backing and other encour- mainly.on a National Economic Development agements for a series of mergers that enlarged Office attached to the NEDC. GEC, the British General Electric Co. But the interventionist IRC was abolished in 1971, about the time the Ministry of Technology became part of the larger Department ofTrade For an outline of the origin and role of the NEDC. see T. Smith. "The United Kingdom." BigBusiness and the Stare.op. cit., pp. and Industrywhich has since again been di- np tho_narlier yprirq of Rritain'sindlts- trial policy is drawn from this source. Also see S. Blank. "Brit- ain: The Politics of Foreign Economic Policy, the Domestic Econ- omy. and the Problem of Pluralistic Stagnation," Between Power "They still exist. however. The committee for electronics re- and Plenty: Foreign Economic Policies of Advanced Industrial cently issued a report urging a comprehensive spctoral policy States. P. J. Katzenstein (ed.) (Madison. Wis.: University of Wis- for the industry. See "Prescription for Electronics." Financial consin Press. 1978), pp. 114ff. Times. Apr. 30. 1982, p. 16.

4 05 402 International Competitiveness in Electronics vided, leaving a Department of Industry and ICL a Department of Trade, The formation of ICL was preceded by much Among other agencies active in industrial less rhetoric concerning the need to create na- policymaking, the National Enterprise Board tional champions than in France, but the (NE13) has had considerable leverage because emergence of ICL in 1968 was similar to that of its ability to provide direct financing to of CII-HB. ICL benefited not only from gov- British firms. Esablished in 1975, NEB has ernment fin,:Tscino, but from aid for R&D and concentrated on startups such as the semicon- the promise of public sector purchases of its ductor manufacturer Inmos, to which it gave products. Britain's Government encouraged about $90 miilion in equity capital (ch. 7). In the series of mergers by which the company quite different .realms, the Science Research was formed, and supplied about $12 million a Council and the Advisory Council for Applied year until 1976 to stimulate its growth," Research and Development, set up in 1976, are Despite this,CL never emerged as a viable intended to supply policy_ guidancr nr, such coml)etitnr in le world computer industry. Al- topics as applications of new lechnowgies and though still holding more than a' third of the the education and triii)ing of engineers. U.K, marketlargely the result of government The number of government bodies involved procurements coupled with "Buy British" persuasion aimed at private firms--ICL has had in Britain's industrial policy provides one ex --- planation for the random approach to pro - little success outside the United Kingdom. grains in electronics. In France, relatively clear Within Britain, the Central Computer Agency, lines of authority link the various parts of the responsible for government purchases, gave bureaucracy dealing with electronics; certain perhaps 90 percent of its orders to ICL during agencies have the lead role in certain areas. By the early 1970's." This is a major reason why comparison, the British approach is uncoordi- the United Kingdom joins Japan as one of only nated. In further contrast to the situation in two countries where American computer man- France, Britain has never been very comfort- ufacturers and their subsidiaries do not have able with government intervention in the af- at least half the installed base, fairs of businessrather surprising consider- ICL is known for its software, butlike most ing the size of the public sector. Not only do computer manufacturers outside the United government plus the nationalized firms employ Statesmissed the shift. toward small systems. about a quarter of the British labor force, but The company has also been handicapped by publicly owned enterprises account for more the lack of a strong local semiconductor in- than 10 percent of the country's industrial out- dustry. Since the latest government initiative put and in recent years about a quarter of total a package of loan guarantees totaling nearly capital investment.'" Nationalized firms in in- half a billion dollars, and the installation of a dustries like steel and automobiles have re- new management team headed by a longtime ceived more attention from British policymak- executive of Texas Instruments' U.K. subsid:s ers than electronics.. Still, the United King- iarythere have been signs of revival.'" - dom's approach to the electronics industry does reflect a belief that government can Not long before jumping back in to try to save strengthen existing firms as well as create new ICLin part because mergers or takeovers in- ones. volving American companies were rumored 62"Technology and Trade Policy: Issues and, an Agenda for Action," op. cit.. p. 58. 63C (IP rarmoy, "SubsidyPniiriec in Flritain, Franrpiinri Wect Germany: An Overview," International Trade and Industrial Policies. S. J. Warnecke (ed.) (New York: Holmes & Meier, 1978), p. 38. "'P. Maunder. "Government Intervention in the Economy of "E. Bailey, "Britain's Role at Ailing ICL," NewYorkTimes. the United Kingdom." Government Intervention in the Devel- May 18, 1981, p. Dt; S. Love, "New Talent Spurs Britain's ICL," oped Economy, op. cit., pp. 131-137. Wall Street Journal, Mar. 1, 1982, p. 27.

o Ch. 10National industrial Policies 403

the Thatcher government had sold off the pub- firms. Tellingly, electronics and communica- licly held 25 percent of the' cornpany's stock: tions manufacturers have spent less of their ICL's checkered past thus illustrates the "stop- own money on R&D than the government has go" quality of industrial policy in the United contributed; in 1975, private firms spent 113 Kingdom. The uneasy relationship between million pounds on R&D, publicly held corpora Inmos and the government has followed a sim- tions 36 million pounds, and the government ilar pattern, one in which Inmos faced con- 130 million." In the United States, the impacts siderable uncertainty over whether the Thatch- of military spending on electronics have been er administration would provide the second in- far overshadowed by the vigor of the commer- stallment of capitalanother 25million cial industry; the British case has been vastly pounds :hat the company was counting on. different. Research and Development Other Policies Toward Electronics Beyond support of ICL, the U.K. computer Among the more intriguing programs of the industry benefited during the late 1960's and U.K. Government have been those aimed at uti- early 1970's from R&D funding provided lization of microelectronics. In the midst of a through the Advanced Computer Technology lengthy debate on the q destion of whether the Project, which provided up to half the costs of country needed an indigenous capability to de- projects dealing with hardware or software. sign and manufacture advanced ICs, the Mi- The British have also attempted to aid their croprocessor Applications Project (MAP) was electronics industry through efforts like a established to encourage companies in any in- preproduction order program, in which the dustry that could to incorporate these devices government purchases newly developed prod- in their products. MAP, which began in 1978 ucts ,and leases them to usersor "clients." and has been somewhat reluctantly continued After a trial period, the client, prepares a report by the Thatcher government, funds up to 25 on the new product and 'hen must either buy percent of the costs of product development. or return it. Other programs have supported Increased support is provided for microelec- software development and marketing, as well tronics-related programs in schools and col- as microelectronics. In addition to contracted leges, principally teacher training. A third ele- basic research, paid for by both civilian and ment consisted of a consciousness-raising cam- defense agencies, commercial product develop- paign aimed at 50,000 managers in private in- ment has been financed through government dustry, with MAP funds supporting seminars cts, particularly to ICL. to educate corporate decisionmakers on the virtues of the new terthnology. Government Nonetheless, the United Kingdom has been spending through MAP totaled nearly $100 a poor performer in R&Dmore precisely, in million over a 3-year period." development. Although British scientists continue to do excellent basic research, industry A related program known as MISPthe Mi- has been reluctant to invest heavily in R&D di- croelectronic Industry Support Programme, rected at commercial products and processes; also started in 1978aids firms in developing between 1967 and 1975, real R&D expenditures and manufacturing ICs. MISP was stimulated by industry declined. Furthermore, sectors like by a report prepared for the NEDC which electronics have suffered from a lack of capable stressed the importance of design and process- engineers. While government R&D expendi- ing expertise; a central goal was mass produc- tures have been heavily concentrated on electronics and aerospacein 197573D percent went to electronics and communications alone "K. Schott, Industrial Innovation in the United Kingdom, Canada and the United States (London: Contemprint, July 1981). this spending, largely motivated by military p. 12. needs (table 77, ch. 10), has had little percept- "'Microelectronics, The New Technology." Department of ible effect on the competitiveness of British Industry, London. 1981. p. 23. 40? 404 International Competitiveness in Electronics

tion'capability in standard devices."' Among the firm might be sold, and following a deci- the steps t 'iken wend the establishment of a joint sion by Inmon executives to locate a produc- venture between 0,13 British firm GEC and Fair- tion facility in the depressed area of South child, then still American-owned, to manufac- Wales, funding was continued. In late 1981, ture a broad-based line of ICs. More ambitious NEB reported $22 million in pretax losses, was the decision to establish a greenfield firm, more than half accruing from its holdings in Inmos. Inmos; losses are to be expected during the early years of such an enterprise, and it is still NEB's announcement in June 1978 that In- too early to judge the success of this recent en- mos would receive equity funding from the try into the world semiconductor industry, but government to design and manufacture n-chan- the qualms of the conservative government are nel MOS circuits, starting with memory chips, not surprising. generated a good deal of controversy within Great Britain (NEB has considerable independ- Has Britain's Approach Worked? ence in making such decisions). The attempt to replicate a merchant semiconductor firm on The answer, implicit in much of the discus- the American modelcomplete wth execu- sion above, is that it has not. While some of the tives experienced in U.S. companies and a de initiatives in electronics may eventually have sign center in Colorado Springswas ,C. move positive results, U.K. industrial policy as a directly into the central arena of worldwide whole has suffered from lack of consistency competition. The risks were high. Ininos was even during periods when the same party has to begin production of 64K RAMs in 1981a been in powerand from a rather odd, if not target which slipped, but as it turned out, no chaotic, mixture of policy instruments. Even more than those of a number of well-known the direct beneficiaries, British electronics American firms. companies, have not been very enthusiastic about the government's support efforts, view- Inmos is a unique experiment; the govern- ing them as favors likely to be withdrawn on ment committed 50 million pounds, split into short notice." Some executives in British in- two installments, with the hope not only of dustry could be described as not only skeptical creating a first -rank semiconductor company, but cynical about their government's policies. but also of luring talented British engineers Nothing like the symbiotic relationship be- back from employment with foreign firms (one tween business and government in Japan, or reason for the Colorado Springs location). Stim- even France, has emerged in the United King- ulating end-users in Britain was another major dom. objective. While there are still many doubters, Inmos appears to have had reasonable success The grab-bag character of U.K. policies to- in developing its first products. The company ward electronics has stemmed in part from the has plans for a new family of microprocessors, inconclusive nature of debate over the need for as well as a broad line of memory chips. Be- a continuing British presence in semiconduc- coming profitable may be more difficult. tor manufacturing. Many took the position that, so long as British industry applied ICs Since the election of the conservativeThatch- in its products, there was no need to have er government in 1979, efforts such as MAP bme-grown design and production capabili- and MISP have been scaled back. The conserv-. ty. Others held that, lacking an IC design and atives' review of the Inmos venture revived productionobase, applications would always lag public debate over the company's prospects. those in other countries. Rather than coming After-a Lonsidelableperiod-ofttmei La hity,dui- ing which it appeared that NEB's holdings in "See D. Imberg and J. Northcott. Industrial PolicyandInvest- ment Decisions (London: Policy Studies Insitute. 1981). pp. 72-73. "'See Microelectronics Into the 1980's (Luton. England: Also J. Northcott and P. Rogers. Microelectronics In Industry: Mackintosh Publications Limited. 1979), p. 27, for a summary What's Happening in Britain. No. 603 (London: Policy Studies of the Sector Working Party report to NEDC. Institute. March 1982). especially ch. 8. Ch. 10National Industrial Policies 405 to a decision; the British have tried to have it yet that this will produce positive results. To both wayssupporting Inmos, though never be fair, industrial policiesof whatever stripe whole-heartedly, while also pursuing applica- are a limited tool when the overall economic tions and technology diffusion through MAP situation has been as grim as Britain's. While and MISP. Similar patterns, over a longer time U.K. industrial policies may seem neither effi- period, have characterized the government's cient nor effective, they have perhaps been dealings with ICL; after many years of public asked to do the impossible. suppor',r, the Thatcher government withdrew, only to find itself forced to the rescue of the West Germany faltering computer manufacturer. Industries like steel and automobiles show similar oscilla- Industrial policy in the Federal Republic of tions in government attitude. Germany (FRG) has been distinguished by reliance on the market. Objectives have been The fact is that foreign firms have already allowed to remain vague beyond the level of captured major shares of most British elec- macroeconomic policy, where stability has tronics markets, except where the government, been paramount. But if private sector actions itself is the customere.g., the defense sector. have been central, this does not mean the role Outside the government market, ICL presents of the pubic sector has been negligible. Follow- little challenge to its American and Japanese ing a "social market philosophy," the West Ger- competitors, just as British firms now hold only man Government has helped reconcile nation- a small share of U.K. semiconductor sales. al, regional, and interest group concerns: The Thus, ICL's agreement with Fujitsu, entailing Act for the Promotion of Stability and Eco- marketing of Japanese-built mainframes in nomic Growth (Gesetz zur Fi5rderung der Sta- England, also involves purchases of Fujitsu bilifdt and des Wachstums der Wirtschaft) pro- semiconductors. While ICL has also negotiated vides a set of tools to coordinate economic pol- for rights to U.S. and Canadian technology, icymaking among government, management, Britain does not seem as well-placed as France and labor aimed at "macroeconomic equilibri- to make use of foreign know-how, and may find um." While avoiding extensive planning, pol- that itis already too late for technological icymakers have paid consistent attention to independence. structural adjustment; since the mid-1960's and the tenure of economics minister Karl Schiller, In sum, many of Britain's industrial policy it has been widely accepted that policy instru- efforts in electronics seem to have been too lit- ments could be deployed to "rationalize" mar- tle and too late. The formation of Inmos and kets and ease structural change. Especially the creation of MAP and MISP came at the since the mid-1970's, the FRG has also pro- close of the 1970's, by which time American vided considerable support for R&D. Although and Japanese suppliers were firmly established proponents of an avowedly sectoral approach in the U.K. market. to industrial policy have become more vocal, Industrial policy has been doubly difficult be- it is still true that industrial policies are market- cause of the stagnant British economy. As eco- oriented, with limited reliance on public own- nomic troubles continued, the government cut ership compared to a number of other Western back its R&D support, making progress in in- European nations, and a strong commitment dustries like electronics still less likely. Recent- to open international trade. Nevertheless, the ly, the Thatcher administration has tried to German Government- has sometimes taken streamline industrial policymaking by merging strong and direct action on the sectoral level the National Research Development Corp. and when--eeononns- problems_have_arisen___ the NEB into a "British Technology Group." The Institutional Setting One goal has beet z-) temper the activist policies of the NEB, which enjoyed considerable Economic and industrial policymaking in the autonomy in the past. There is no indication FRG combines elements of decentralized deci- 403 406 International Competitiveness in Electronics sionmaking with representation by major in- The Federal Government also holds majori- terest groups, including labor. The ministries ty shares in five banks, while cities and states of Finaince, Economic:3, Labor and Social Af- have their own financial imititutions. One of fairs, and Research and Technology are among the nationalized banks, the Kreditanstalt fin. the mere influential in terms of policies affect- Wiederaufbau, is a development bank that pro- ing industry. Macroeconomic policies are de- vides funds to commercial lenders.'') While the veloped and implemented by a number of agen- financial communities are major seats of in- cies. The Ministry of Finance submits 5-year fluence over industrial policy in both West Ger- plans. Since the early 1960's, a five-person many and France, they function quite different- Council of Economic Adviserscomprised of ly in the two countries: rather than selective academics not otherwise'attached to the gov- credit for favored firms and industries as in ernmenthas been responsible for macroeco- France, German banks have supported fiscal nomic forecasting. The Council, also prepares and monetary policies oriented toward aggre- annual reports on the health of the West Ger- gate growth. man economy. Money supply is the responsi- If economic and industrial policymaking in bility of the Deutsche Bundesbanklegally in- West Germany is less centralized than in dependent of the government, though closely France, the lines of responsibility are more tied toit.Policies and analysis related to clearly drawn than in Great Britain. While the economic and industrial development are cen- Research and Technology Ministry (BMFT), tered in the Ministry of Economics. The tends to approach industrial policy with Lander (state) governments help formulate eco- a perspective quite different from that of the. nomic as well as regional development policies. Economics Ministry, the division of authority A joint Federal-Lander panning committee, for is more or less predictable and consistent. Ger- instance, draws up regional action programs many's parliamentary system has seen few identifying growth points (schwerpu.!...torte) to changes in government since 1949; when a dif- be promoted via investment grants. West Ger- ferent party has come to power, overall objec- many has emphasized regional development tives such as maintaining the country's export perhaps more heavily than any other Western strength while controlling inflation have been industrial nation, with the Lander Govern- retained. ments central to these efforts.°9 As chapter7pointed out, financial institu- Policymaking Processes tions have a special place in the West German A distinctive feature of the German system pclicymaking structureas they do in France is the broad representation of interests, the ef- and Japan. Executives of the central Bundes- fort made to integrate diverse points of view. bank keep in close contact with public officials, The Stability and Growth Act empowers the and normally act in support of the govern- Federal Government to provide "orientation ment's economic policy. The Bundesbank's data" for policy measures to be "simultaneous- control of the money supply gives it direct in- ly and mutually agreed upon" by Lander and fluence over the value of the deutsche mark. local governments, labor unions, and employ- During the years of rapid economic expansion, ers' associations. In the late 1960's, "concerted particularly the early 1960's, the bank helped action" incomes policies were developed, maintain an undervalued currencya strategy aimed at consistency in approach among gov- that strengthened the export competitiveness ernment bodies and socioeconomic groups.on of German goods but earned a good deal of crit- budgetary matters as well as wages and prices. icism from thecnnntry'c trading_partnerq Cannarteci-action-was-not-an a empt-tcr-sup-----

't0G. de Carmoy, "Subsidy Policies in Britain, France and West Germany: An Overview," International Trade and Industrial Pol- 7°E. Owen-Smith, "Government Intervention in the Economy icies, S. J. Warnecke (ed.) (New York: Holmes & Meier, 1978), of the Federal Republic of Germany." Government Intervention p. 52. in the Developed Economy, op. cit., p. 176. Ch. 10National Industrial Policies 407 plant the monetary and fiscal policies of the unions traditionally exert pressure on the Federal Government, but was intended as an government through political activism, often adjunct and complement to these, the basic ob- confrontational. Even with Mitterrand and the_ jective again being the creation of an environ- Socialists in power, this is not likely to change ment conducive to economic growth. Like co- much. determinationwhich ensures labor a voice in plant operationsconcerted action sought to Policies Toward Electronics integrate labor and other interest groups into Despite its stress on macroeconomic tools, the mainstream of policy formulation. Today, West Germany has, over the years, instituted concerted action has fallen into disuse, but a considerable number of policies directed at speculation on its revival regularly surfaces; if specific industrial sectors. Some have been nothing else, this indicates the persistence of in portions of the economy where government the view in Germany that sound economic and ownership has been widespreade.g., energy indwArial policies depend on broadly based and banking. In contrast, sectoral involvement consensus-building. in electronics has been mostly restricted, to Indeed, institutionalized participation by R&D; compared to both France and the United major social groups appe I's to offset much of Kingdom, military involvement has not been the fragmentation that otherwise might seem prominent. Moreover, in further contrast with to characterize industrial policymaking in the these two countries, when FRG officials at- FRG. As in a number of other Western Euro- tempted to encourage a "rapprochement" pean countries, notably the Scandinavian na- among Siemens and several other computer tion:I, Germany's industrial policy is marked manufactufers, the large and powerful Siemens by concern with labor issues. In 1974, the concern resisted quite successfully." When BMFT and the Ministry of Labor and Social AEG-Telefunkenafter Siemens the country's Affairs set up a joint research program on "hu- largest electrical and electronics producerfell manization of the workplace." Directed not G1-1 hard times, the private .:sector at first dealt only at health and safety issues, the program with tilt: crisis on its own. A consortium of 24 aims as well to identify and encourage organi- ommercial banks engineered a massive res- zational changes that would increase job satis- cue effort, with financing totaling more than' faction (ch. 8). A number of studies gponsored half a billion dollars." Only when the bankers' by the program, which is oriented s'±rongly to- efforts proved insufficient did the government ward field experiments and employee partici- step in with a package involving further loan pation, have explored impacts of autoLoation guarantees and export credits.74 As this implies, and computer technologies." and as chapter 7 described in more detail, cooperation among induStry and financial institu- The systematized participation of labor in tions in the Federal Republic has been com- West Germ -.ny is especially noteworthy,in con- monand an increasing subject of parliamen- trast to Japan or France. In Japan, organized tary scrutiny and public criticism, on grounds labor is in factif not always in appearance that the power of the banks is too great. relatively powerless; the consensus so clearly visible in Japan comes, not from full participa- Despite efforts such as .the aborted Tele- tion, but from a rather passive acceptance by funken rescue, government influence has not other groups of policies that business and gov- been exercised as directly in electronics as in ernment have agreed on. In France, organized "Jequier, op. cit., p. 217. labor is vocalwith a marked radical cast 73K. Done, "The Last Chance Rescue,"Financial Times,June but labor participation in setting policy has not 14,1982. Under West German law, banks can own equity in pri- been internalized as in Germany. French vate firms and act as brokerage houses (ch. 7). German banks held about 40' oercent of Telefunken's stock. ""Research on the Humanization of Work," Ministry for Re- 74See "Germany's Telefunken Insolvent,"New York Times, search and Technology and Ministry of Labor and Social Af- Aug. 10,1982, p. Dl. Nonetheless, the firm entered bankruptcy fairs, document No. 2181/74v. in mid1982.

411 408 International Competitiveness in Electronics sectors such as energy-(where subsidies have funken never achieved much success in com- contributed to high domestic coal prices), steel puter systems. Siemens remains the largest (where firms such as Salzgitter are publicly German-owned computer manufacturer, some- owned). or shipbuilding (where a range of pol- what ahead of Nixdorf in sales (ch. 4, table 42). icy initiatives have been marshaled to shelter But Siemens' production is far less than that the industry from decline)." Still, since the late of IBM's German subsidiary; Siemens has 1060's the West German Government has never appeared to view computers as a major sought ways of strengthening the nation's com- piece in its corporate strategy. The company puter industry. While on the whole the German has only about 20 percent of the German com- electronics sector has been the strongest of any puter market, and less than 10 percent for in Europe, it has shared the common weakness Europe as a whole. Nonetheless, Siemens con- in computer. Part of the reason appears to tinues to receive by far the largest share of have been that the bigger electronics firms government funds for R&D in computer tech- Siemens, AEG-Telefunken, SELwere already nology." The contrast with Nixdorfa man- heavily committed to other lines of business ufacturer of business-oriented rriaicomputers consumer products, telecommunications, elec- is striking. Nixdorf is an aggressive world- trical machinery. Much like such American wide competitor in its chosen markets, much companies as RCA or General Electric, these in the American mold; the company has ac- large and diversified enterprises never devel- complished this with little government assist- oped much strength as computer manufactur- ance. ers. From the standpoint of the German Gov- Again in common with other European elec- ernment, there really was no computer indus- tronics firms, a number of German manufac- try as such to support. As a result, it proved turers have pursued ties with American and difficult to devise effective policies for encour- Japanese enterprises, one aim being technol- aging either the technology of computing or ogy acquisition. In 1978 Siemens purchased 20 commercial production. As in many other percent of Advanced Micro Devices. More re- countries, government procurements have cently, the company negotiated an agreement been channeled to local firms. Nevertheless, in with Japan's leading producer of computers, contrast to Japan and France, the FRG has Fujitsu; Siemens now markets several of Fu- largely avoided attempts to shield the industry jitsu's IBM-compatible mainframes in Europe. from foreign competition, relying instead on Such arrangements have brought criticism of domestic supports and subsidies.* A major government support for Siemens as failing to thrust of German efforts has been to stimulate promote an indigenous computer industry. utilization of computers through training programs and applications support. Research and Development Support Although benefiting from government fund- Financial subsidies for Siemens' computer ing amounting to more than 100 million efforts have been part of a considerably larger deutsche marks (something over $50 million program of technology development in the in fact a relatively small fraction of West Ger- FRG. Total R&D expenditures grew more than many's total subsidies during the 1970's for 60 percent in real terms between 1969 and computers and information processing), Tele- 1980, increasing from 2.1 percent of GNP to 73See Owen-Smith, op. cit.: p. 174 on coal prices; p. 184 on 2.3 percent; the West German Government has Salzgitter;p.173 on shipbuilding. `Even so, a recent trade dispute shows thatin Germany as 766,. 1978, the West German Government had supplied Siemens elsewhereforeign firms are often discriminated against, In a withacumulative total of 351 million deutsche marks(nearly case similar to AT&T's choice of Western Electric over Fujitsu $200 million) for the development of large- and medium-sized in fiber optics, Bremen University was forced to reverse a deci- computers. See "Sixth Report of the Federal Government on Re- sion to purchase a computer E y sie m from Burroughs. The con- search," Federal Minister for Research and Technology, Bonn, tract went to Siemens at a price comiderably above the American 1980, p 82. Thd computer support programs of the BMFTnow bid. See "'Technology and Trade Policy: Issues and An Agenda seem widely viewed as failures; they have been drastically scaled for Ac',1on," op.' cit., p. 49. back. Ch. 10National Industrial Policies 409 - strongly supported R&D, which has accounted ly, with the responsibilities of government of- for 3 to 4 percent of the Federal budget dur- ficials limited mostly to coordination and ing the pa_ I decade." Since the initiation of the avoiding duplication. Siemens has received 25 second data processino program in 1969, the to 30 percent of the money, with Telefunken electronics industry, and particularly the com- and Valvo each getting 10 percent or more. In- ponents sector, has.received a substantial frac- dustry seems to have regarded the program as tion of this spending; 30 percent of German useful but not of great impact; the major Ger- R&D funds have gone to electronics (and elec- man electronics firms have traditionally had trical equipment) -- % greeer fraction than in strong commitments to R&Dincluding basic any of the other cc. listed in table 77. But researchand government money appears to while governmeL _uppurt for electronics as a have gone mostly to efforts that the private sec- whole increased steadily during the latter part tor has judged marginal. Indeed, a principal ra- of the 1970's, it appears to have peaked at the tionale has been to finance projects with time end of the decade, with the more :recent con- horizons too long for industry to justify. traction stemming from disappointing results in computer technology." While most of the money in this first major 0 VLSI program went to big companies, the FRG Disillusionment with support for computer has also paid a good deal of attention to smaller systems has led industrial policymakers in the firmsof which there are more than a thou- Federal Republic to reorient their programs sand in electronics.8° In contrast to the market toward microelectronics. Here funds have orientation of other German industrial policy gone toward device physics and processing initiatives, the BMFTa relatively new agen- technologies, as well as IC design and develop- cyhas designed an array of sector-specific ment. Between 1974 and 1978 relatively modest programs aimed at small enterprises and sums were spent by the government on Micro- growth industries like electronics and biotech- electronics R&Dabout $30 million annually. nology. Small technology-based firms in the A somewhat more ambitious effort began in Federal Republic often face difficulty in rais- 1979. Like the computer R&D which it sup- ing capital. As in most countries other than the planted, VLSI support at first centered on the United States, venture capital markets are min- large electronics manufacturers. In its first pro- iscule. Viewing this as an obstacle to innova- gram, the government contributed about $300 tion; and with motives much like those lead- million over)the period 1976-82, ;with industry ing to the creation of the National Enterprise participants putting up matching funds and Board in Great Britain, the FRG Government competing on a proposal basis." The German set up a venture financing company (Deutsche VLSI program has been much less centralized Wagnisfinanzierungsgesellschaft) in 1975. This than Japan's; participants work independent organization purchases minority interests in German firms --with the intent of backing innovative developmentswhile giving the pro- ""Sixth Report of the Federal Government on Research," op. prietors preferential rights to buy back the equi- cit., p. 75; "FRG's Position in World R&D Community Assessed,!' West Europe Report, Science and Technology, No., 72, Joint Pub- ty if their business succeeds. lications Research Service JPRS 78876,,Sept. 1, 1981, p. 40. "See "Sixth Report of the Federal Government on Research," The turn toward support for smaller compa- op. cit.. p. 53. While the budget, for government expenditures nies has also been reflected in the BMFT's latest on electronics R&D was schedided to increase from about 350 million deutsche marks in 1975 and 1976 to more than 600,million in 1980 and 1981, expenditures in these later years were w"General Scheme of the Federal Government's Research and cut back considerably from the amounts originally planned. Technology Policy for Small and Medium-Sized Undertakings." " M . Gold, "West Germany Reported About To Launch $300 Ministry for Research and Technology and Ministry of Econom- Million VLSI R and D Plan," Electronic News, Apr. 30, 1979, ics. 1979 updating. Also seeInnovation in Small and Medium p. 1; "Eurcipean Semiconductor Industry: Markets, Government Firms(Paris: Organization for Economic Cooperation and De- Progratris," West Europe Report, Science and Technology, No. velopment, 1982), pp. 133-139; and G. Kayser, "Small Business 134. Joint Publications Research Service JPRS 82686, Jan. 20, Policy in the Federal Republic of Germany," European Small 1983, p. 32. Business Journal, vol. 1, winter 1983, p. 39.

413 99-111 0 - 83 - 27 410 International Competitiveness in Electronics semiconductor program. The 3-year effort, be- .technologies to the private sector, as well as ginning in 1982 and funded at about $45 mil- developing them. lion per year islike a number of Britain's Joint government-industry financing on a more recent initiativesdirectedprimarily at project basis is the rule. FhG instituteswhich applications.," This new program comes on top together employ more than 2,500 peoplepro- of a 40-percent increase in microelectronics vide technical advice to smaller firms, coop- R&D support that had already been scheduled. erate with universities, and function as tech- Most of the applications money will be chan- nology conduits.Institutesare organized neled to small firms, with one of the objectives around technical disciplines; one concentrates being job creation; of 1,000 grant applications on semiconductor devices and processingtech- received during the first 6 months, two-thirds nology (the Institute for Solid State Technology were from companies with fewerthan 200 em- in Munich), another on computer systems (the ployees. Administration is the responsibility of Institute for Information and Data Processing, the VDI Technology Center, established by Karlsruhe). Several others work in areas less the BMFT in 1976 specifically to help small- directly related to electronics. and medium-sized firms develop and apply microprocessor technology. The Institute for Solid State Technology, one of the more successful of the Fraunhofer lab- The Fraunhofer Gesellschaft" oratories, can serve to illustrate the FhG model. Loosely associated with the Technical Univer- West Germany's attentiveness to smaller en- sity of Munichthe Institute's director holds terprises does not stop with microelectronics. a chair there, and perhaps 20 students work The Ministry of Economics supports more than at the laboratorythe Institute employs near- 80 industrial research associations, while the ly 100 people, about half of them engineers or Fraunhofer Gesellschaft (Association of In- scientists. This makes it the largest organiza- stitutes of Applied Research, FhG)comprised tion of its type in West Germany, and perhaps of some 25 institutes which function as R&D in Europe. Founded in 1974, housed in its own laboratorieshas as one of its major respon- building away from the university, and grow- sibilities the diffusion of technology to in- ing largely through the initiatives of its direc- dustry, especially small companies. tor, internationally known for his research in Strengthening the FhG, which was founded semiconductor technology, the laboratory gets in 1949 to perform applied research and engi- 70 percent of its annual fundingabout $5 mil- neering development on a contract basis, has lionvia separately budgeted R&D projects. been one of the more intriguing BMFT initia- The BMFT typically provides a major share of tives. The FhG remained small unti!, a govern- project budgets, the remainder coming from ment decision in 1969 madethe chief vehi- one or more industrial sponsors. In essence, cle for support of applied research. At this the government shares risks with industry. point funding began to increase rapidly. A re- Two of the Institute's staff members are paid examination of FhG goals in 1973-74 led to a directly by the BMFT to advise and consult strengthening of its mandate for transferring with small- and medium-sized' companies. Much of the laboratory's work is concerned "'See "Increased Government Funding for Microelectronics," with processing technology; prototype circuits West Europe Report, Science and Technology, No. 92, Joint Pub- can be fabricated, along with smalllots of spe- lications Research Service JPRS 80133, Feb. 18. 1982, p. 5; "Special Microelectronics Program," West Europe Report, cialized devices such as sensors and ICs for Science and Technology, No. 113, Joint Publications Research medical applications, The Institute also oper- Service JPRS 81392, July 29, 1982, p. 13: ates an X-ray lithography facility at West Ger- "=Much of the information in this section is based on interviews. See, also H. Keller, "30 Jahre Fraunhofer-Gesellschaft: many's synchrotron storage ring in Hamburg. Rilck-und Ausblick," FhC Berichte 3-79 (Munich: Fraun- hoferGesellschaft, 1979), p. 3, and Vertragsforschung filr Wirt- Perhaps the most noteworthy aspect of the scha ft und Staat (Munich: Fraunhofer- Gesellschaft, 1981). FhG and its mandate from the BMFT is the ori-

4iil Ch.10National Industrial Policies 411 entation toward commercial technologies. The the best placed of any in Europe. But in the institutes are not basic research organizations more likely event, progress in electronics in the that function remains with the Max Planck Federal Republic will dependas it has in the Gesellschaft. Nor do they function as govern- paston domestic actions, public and private. ment laboratories, although the relationships between individual FhG institutes and govern- Past government policies, when directed at ment agencies vary considerably; the ties are electronics, have not been notably successful. closest among the six that carry out R&D fi- Nonetheless, German industry has a sound nanced by the Federal Ministry of De ense. The base to work from. Siemens, if not a leader in institutes are a conscious attempt to speed computers, probably has the best semiconduc- commercialization of new technologies and dif- tor technology of any company in Europe. fuse R&D results through industry. One way in (P is strong in linear circuits because of which the FhG does this is simply to provide a its emphasis on consumer electronics, but Sie- venue for bringing representatives of Federal, mens was virtually the only European man -. Liinder, and local governments together with ufacturer that recognized the importance of MOS ICs at an early date.) Germany's domestic industry and the universities. The Fraunhofer production of ICs has grown as a percentage experiment is an attempt to compensate or the of consumption in recent years, a sign of weak links that exist in Germanyas in most Siemens' continued technical strength and per- countriesamong these groups, especially haps of positive results from government R&D where commercial technologies rather than programs." But the entire consumer elec- basic research are involved. Likewise, the deci- tronics sector in West Germany, not just AEG- sion to accept defense-related projects in 1955 Telefunken, has faltered under the pressure of was based on the belief that it was better not Japanese competition. ZVEI, the Central Asso- to isolate defense R&D, but to combine it with ciation for the Electrotechnical Industry, has civilian work in hopes that each would benefit. claimed that increased sales of imported home (Defense-related projects now account for 20 entertainment products have been a direct to 25 percent of the FhG's effort.) cause of shrinkage by such firms as Grundig Within Germany, the Fraunhofer Gesell- AG, and consequent losses of jobs; at the end schaft has won high marks for facilitating tech- of 1982, Grundig and Philips filed an antidump- nology transfer while avoiding direct govern- ing complaint against Japanese producers of ment involvement in decisions on directions VCRs.85 In computers, West German firms and priorities, but its comparatively small have less than 5 percent of world sales. Nix- budgetabout 230 million deutsche marks in dorf has chalked up respectable profits and ex- 1981, something over $100 millionlimits the ports by concentrating on smaller business- assistance that flows to any one industry." oriented systems; the company has done this on its own, without significant government aid. The Future As the example of Nixdorf showsa lesson repeated in other countriesindustrial policy If a joint strategy for the European Commu- nity in electronics comes to passa prospect is no substitute for well-managed private firms. that seems slight, as discussed in the next sec- West Germany has thus maintained its position, but not so improbable as a few years ago tion in the second tier of the world electronics West. Germany's industry would probably be industry. Can it compete in the years; ahead when faced with both American and Japanese ",Vertragsforschung kir Wirtschaft und Steal, op. cit. In 1978, 35.7 million deutsche marks 1818.7 million) went to the four Fraunhofer institutes ,involved in work related to microelec- "G. Dosi, Technical Change and Survival: The Europea» Semi- tronics and information processing; see "Cooperative R and D conductor Industry (Brighton, U.K.: Sussex European Research Programs To Stimulate Industrial Innovation in Selected Coun- Centre, Sussex European Papers, May 1981.) triesWest Germany," Department of Commerce, National Bu- J. Gosch, "German Consumer Firms Face Bad Times,''Elec- reau of Standards, Office of Cooperative Technology, November tronics, Sept. 11, 1980, p. 97; "Japanese VTRs Are Target of EC 1979, p. 69. Antidumping Case," Wall Street Journal, -Disc. 24,_1982, p. 9,

41.5 412 International Competitiveness in Electronics firms? At present, many of the industry's prob- government, then, is far from laissez-faire. In lems stem from the brciader dilemmas of the contrast to the British case, sectoral initiatives FRG economy; high interest rates and low prof- have been pursued with a good deal of consist- it margins have made it difficult for German ency over time, although such policies have not mnpanieS, which have traditionally borne the necessarily entailed extensive involvement by oulk of such expenses themselves, to maintain government officials. The West German case high levels of spending for R&D and new cap- does underscore the critical importance of ag- ital investment. In recent years, Siemens has gregate policies as necessary (if perhaps not accounted for as much as 12 percent of all West sufficient) to sectoral development. German industrial R&Dfor many observers Industrial policy in Germany has benefited this alone signifies imbalance." from a better sense of timing than in the United West German firms also face a critical deficit Kingdom. Government support for R&D in in technical manpower, despite mounting un- electronics began to pick up in the late 1960's, employment in the nation as a whole. Accord- and has continued to growthis despite an on- ing to the Association of German Engineers, going debate between the BMFT, which favors 16,000 jobs have been vacant for lack of people expanded sectoral thrusts, and the Economics particularly in electrical, mechanical, and Ministry, which continues to stress aggregate civil engineering; in 1980 only 3,600 students measures. While R&D programsincluding were enrolled in technical universities able to funding for VLSI research and the efforts of accommodate 4,700.87 Such problems are in no the FhGhave not advanced the competitive position of the German electronics industry in sense unique to West Germanythe question' any very obvious or dramatic sense, they ap- is whether government policies will help to pear to have nurtured it in a variety of less resolve them. direct and visible ways. Unlike electronics policies in nations which have tried to leapfrog How Effective Are West German the competition, the German approach has Industrial Policies? been one of broad support for more basic kinds Industrial policy', has a less distinct identity of research, in the hope of returns over the in the FRG than in\many other countriesat longer run. least it is harder to summarize. On the one As the Federal Republic struggles with ris- hand, the approach has been more market ori- ing unemployment and continuing economic ented than in France;',certainly planning and stagnation, such policies will be severely tested. coordination on the French model are absent. Formulated in a time of overall growth, there On the other hand, the West German Govern- is no guarantee that the FRG view of industrial ment has consistently supported industrial policy will prove adequate to deal with the development through macroeconomic meas- adverse conditions promised by the rest of the ures and by integrating a broad range of 1980's. Germany's problem is much the same perspectives and interests into the policymak- as that faced by the United States. ing process (critics in some countries might regard this as !a weakness). The role of the The European Community "According to Siernens' annual report for 1980, the company I In West Germany and elsewhere in Europe, spent over 3 billion; deutsche marks (about $1.6 billion), more I than 9 percent of worldwide sales, on R&D. Over 90 percent of concern over technology gaps vis a vis Amer- the money came from Siemens' own funds, the rest from govern- ican and Japanese competitors has led to peri- ment contracts and grants. For comparisbn, U.S. firms during f odic proposals that the European Community he same year spent the following amounts as a percentage of/ sales: Amdahl, 15.8 percent; IBM, 5.8 percent; Data General, 10.0' /(EC) d4velopa j,k)int polidy toward electronics. percent. See "Spending for Research Still Outposes Inflation,'} Rapid increases \in consumer electronics ship- ithin.,:ssWeek, ',sly 6, 1981, p. 60. Siemens is clearly committed B 1mentslfrom Japan have Stimulated talk of im- toeeping up in technology. Tagliabue, "Germany's Economy Stumbles," New .York port restraints, but a cominon effort in R&D has Times. Apr. 13, 1981, p. Dl. been the most frequent suggestion. A 1980

1 Ch. 10National Industrial Policies 413

study by Siemens, for instance, held out little carefully designed to avoid areas where coun- hope for indigenous semiconductor industries; tries and companies compete directly. Fund- the conclusion was that continued growth in ing, planned ,`o be about $45 million over 3 European sales would probably benefit Japa- years, will be contingent on substantial con- nese firms the most, with the U.S. market share tributions from the industrial participants, dropping from three-quarters to less than two- which number a dozen of Europe's largest elec- thirds by 1985.88 tronics firms (1CL, Siemens, Nixdorf, CII-HB, Although much of Europe has suffered sim- Philips. Olivettithe planning effort began ilar problemsthe twin maladies of recession with company managements rather than gov- and inflation, a perceived slowdown in tech- ernment officials). EC planners hope the effort nological advance, rising labor costs, unem- will expand within a few years to encompass ployment, low rates of capital investment, slip- more ambitious targetse.g., projects analo- ping competitivenessjoint responses have gous to Japan's government-sponsored R&D been slow to appear. In 1980, the EC's industry ventures in supercomputers And fifth-genera- commission proposed a European strategy in tion systems. It remains to be seen, however, electronics that would have included govern- whether the Europeans will manage to coop- ment-funded programs to develop semiconduc- erate effectivelyand, if they do, whether tor processing equipment, as well as an ad- cooperation in basic research will make-much vanced communications network linking the difference, given that many of the large Euro- members of the Community.89 The proposal, pean electronics companies have always per- which would have required modifications to formed high-quality research but have had dif- national procurement policies, stalled when the ficulty translating the results into commercial French dismissed it as insufficient while the products. British dithered over the implications of exposing ICL to open procurements. So, while the Japan EC countries nave recognized the need for a japan is the exception to many rules in the more unified approach, national concerns have international electronics industry. Government thus far remained paramount. policies evolved along with the industry; they The latest attemptwhich bears the name have consistently supported private firms, Esprit (European Strategic Programme of Re- directly and indirectly. Subsidies have been search in Information Technology)got under- substantial, though not inordinately large com- way in mid-1982. At first directed chiefly at pared with other countries. Both financial sup- semiconductor processing, in part because port and indirect measures have been careful- Europe has been heavily dependent on im- ly targetedbenefiting some parts of Japan's ported processing equipment, Esprit will also electronics industry much more than othersa support work on chip architectures for VLSI, feature that has attracted much attention in the device modeling, and computer-aided circuit United States. Consumer electronics, for exam- design and testing.99 The program has been ple, has not been a major focus of government policy compared to microelectronics and com- ""Growth of Electronics Market in Europe Seen Benefiting puters; nevertheless, diiring the period of con- Japan," New York ,Times, Nov. 28. 1980, p. D3. solidation and concentration that extended ""Europe's Electronic Strategy is Modest, But It Still Isn't Easy." The Economist, July 26, 1980, p. 63. Over the years. the through the 1960's, the government maintained EC Commission has produced a variety of elaborate proposals a series of barriers to imports and foreign in- and studies, to little evident effect. See, for example: "New Infor- vestment that effectively limited competition mation Technologies," Sept. 1, 1980; "Proposal for Council Reg- ulation Concerning Community Actions in the Field of Micro- in consumer electronics to local firms.'" Lib- electronic Technology," Sept. 1, 198C; "The Competitiveness of European Community Industry," Mar. 5, 1982; all Commission r.""Sources of Japan's International Cu-npetitiveness in the Con- of the European Communities. Brussels. sumer Electronics Industry: An Examination of Selected Issues," "D. Fishlook, "Why Europa Wants Esprit," Financial Times, prepared for OTA by Developing World Industry and Technol- Aug. 3, 1982, p. 13; J. Smith, "Can Europe Cooperate on Re- ovy, .1.nc. under contract No. 033-1010.0, pp. 31-46; see also The search?" Electronics, Aug. 25, 1982, p. 85." U.S. Consumer Electronics Industry (Washington, D.C.: Depart-

41Y 414 International Competitiveness in Electronics eralization began only in the late 1960's, as the tors at the core of Japan's emerging industrial government's attention turned elsewhere; by structure, the structure that will keep the coun- this time Japan's consumer electronics industry try's economy growing and competitive into had become well established. Foreign invest- the next century. A unique feature of elec- ment controls on monochrome TV production tronics policy in Japansince copied by other facilities were relaxed in 1967, on color pro- nationsis official sanction and promotion, duction 2 years later (in some contrast with not only of the industry as such, but of elec- European governments, Japan limited inflows tronics as the epitome of a broad array of of foreign capital as well as products). Like- emerging technologies (including CAD/CAM, wise, the tariff on color TV imoortsformerly robotics, composite and ceramic materials, and 30 percentdropped to 71/2 percent in 1971. biotechnology); the policies of Japan's Govern- Similar measures were adopted to protect the ment toward electronics are in fact aimed at fledgling computer and microelectronics goals transcending conventional sectoral industries. boundaries. These policies, for years, have also been consciously directed at leapfrogging other TV manufacturers clearly benefited from nation's technologiesanother aspect of the government support of broadcasting, from the Japanese strategy that governments elsewhere, array of direct and indirect trade barriers that particularly in Asia, have tried to emulate. In Japan erected during the postwar years, and several respects then, Japan's use of the tools from policies that encouraged exporting. Still, of industrial policy has been innovative; Jap- direct and positive supporte.g., for R&D and anese policymakers have been both more am- product developmentwas modest compared bitious and more experimental than, for in- to the attention lavished on information processing.Beginning in the 1960's, computers and semiconductors have been at the center of policies toward electronics and "the information industry." As these sectors grew, Japanese pol- icyakers shifted directionaway from the complex of measures for protecting domestic industries that had been the hallmark of the government's approach during the 1950's and 1960's-, toward more positive measures. Rather than simply sheltering local companies, the government sought to actively strengthen Ja- pan's capability in data processing, with the aim of moving into world markets. Financial subsidies, primarily for R&D, were a major vehicle, along with other, less direct supports for.. research, as well as measures to encourage and facilitate applications of new technologies. An example of the latter is the Japan Electronic Computer Co., which buys data processing equipment from computer manufacturers and leases to users (ch. 4). Today the information industries are viewed as the flagship of the knowledge-intensive sec-

Photo credit: Bell Laboratories ment of Commerce. September 1975), pp. 12-13. andUnited StatesJapan Trade: Issues and Problems(Washington. D.C.: Light emerging from glass filaments used General Accounting Office. ID-79-53. Sept. 21. 1979). ch. S. in fiber-optic communications

41 d Ch.10 National Industrial Policies 415 stance, their court lerpo rts in France or the with membership drawn from the private sec- United Kingdom. tor. The efforts of people like YonejI Masucla The only other public agency with significant have fed the broad consensus which evolved ongoing jurisdiction related to electronics is the among :coders in Japanese business and gov- Science and Technology Agency (STA), under ernment concerning the critical importance of the Prime Minister's Office. In size and re- electronics, and particularly computers. Active sources, STA cannot rival MITI. It does, how- since the mid-1960's on advisory councils to ever, coordinate the government's budgetary the government, Masuda was responsibleas outlays for R&D and related exper ditures, pre- Executive Director of the Japan Computer paring, for example, an annual, ",science and Usage Development Institutefor the 1972 re- Technology White Paper." STA also funds report, "The Plan for an Information Society: search projects, including contract research by Japan's National Goal :award the Year 2000." private firms, through its New Technology De- Respected academic and author of more than velopment Corp.93 STA influence over nuclear, 20 hooks, as a government advisor Masuda ad- ocean, and space technologies has been more vocated a comprehensive national plan for extensive than in electronics. "computerization" in Japan, including government investment in future-oriented projects This is not to say that other government agen- such as a "coinputopolis," or computerized cies do not develop policies that affect the Jap- city, and a computer peace corps. Masuda's anese electronics industry. They do, but on a ideaswhich are well within the mainstream less regular basis than MITI and STA; more- of this brand of futurism, based on the assump- over, the influence of other agencies tends to tion that the production of information will be less direct. The Ministry of Finance (MOF) gradually overshadow the production of ma- has jurisdiction over macroeconomic matters terial goods, eventually comprising the next e.g., fiscal and monetary policy. In recent stage in economic developmentheavily influ- years, growing budget deficits have forced the enced I'vlITI's (the Ministry of International MOF to weigh proposals for sectoral assistance Trade and Industry) vision of a future infor- more carefully; competition for funds among mation society.92 Most Japanese policymakers electronics and other industriesas well as take a more pragmatic view, but the visionary with government objectives other than indus- outlook of Masuda and others like him helped trialdevelopment--has become stiffer. Ti crystallize a broadly based consensus on the MOF also exercises a good deal of influenc importance of computer technology. over the Bank of Japan, while public corporations such as the Japan Development Bank can The Institutional Setting channel funds to favored companies through loans and grants (ch. 7). Long-term projections In contrast to the United States or the United by the Economic Planning Agency include Kingdom, a well-defined group of government forecasts of output by sector of the economy agencies in japan bears the responsibility for that are widely regarded as reliable guideposts official policies toward the electronics in- to future business prospects. While neither dustry. Both policy development and imple- public nor private banks need subscribe to the mentation are centralized in MITI, specifical- government's investment priorities; they often ly its Information Machine Industries Bureau. put money into sectors targeted by such plans. Satellites attached to MITI include the Agen- cy for Industrial Technology, with functions An independent body, the Fair Trade Com- in R&D, and the Information Processing Indus- mission (FTC)though peripheral in industrial tries Advisory Council; a prestigious group policy compared to MITI or the MOFhas 91Cagakt, Gijutsu-cho/Kankyocho (Science and Technology 92Y. Masuria, The Information Society as PostIndustrial Socie- Agency/Environment Agency) (Tokyo: Kyoikusha, 1979). pp. 48, ty (Tokyo: Institute for Information Society, 1981), p. 29. 77.

413 416 snternational Competitiveness in Electronics often resisted policies formulated by those ahead (in some energy research areas, such as agencies. Examples include legislation exempt- nuclear power, this has already occurred). ing sectors like electronics from provisions of Japanese antitrust law to facilitate "collabora- Policymaking in Japan tion" among firms for "rationalizing" the in- Japanese indus :ial policy is built on close dustry." The FTC has repeatedly, though sel- consultation among business leaders and gov- dom successfully, opposed WTI recommenda- ernment officials. Corporate executives rou- tions for antitrust exemptionsbut in contrast tinely participate in both formal and informal to the Japanese petroleum industry, where the discussions concerning policies toward elec- FTC has frequently investigated particular tronics. It is an overstatement to claim, as some companies, electronics firms have seldom been observers have, that in Japan industry tells scrutinized apart from matters of rebates and government what to do, while in France gov- resale price maintenance. Even the public out- ernment tells industrybut this does convey cry over price-fixing among color TV manufac- a sense of the difference. The Information turers fueled by media reports of dumping Processing Promotion Advisory Council, for charges against Japanese firmsin the United instance, brings together representatives of Stateswas assuaged informally ?ether than by Japan's leading electronics firms to discuss FTC decision; MITI persuaded the companies MITI proposals. While such advisory councils involved to lower domestic prices by 15 per- meet relatively infrequently, and rarely have cent. Legal challe,igas to the business activities a determining voice in policy development, of Japanese electronics firms have come pri- they serve to mobilize business interests and narily from abroad: in the United States alone, help form a consensus in support of the even- opanese electronics companies have been in- tual outcome. Advisory councils are only one volved in more than 30 lawsuits, the mejority such forum. Representatives of the many elec- over dumping." tronics industry associations in Japan interact In addition to these traditional actors, other with officials from MITI and other agencies agencies and organizations have recently through a wide network of public and semipub- found more prominent roles. The intermin- lic institutions. Several organizations bring isterial Council for SCience and Technology together government, industry, and universi- has been active in developing and coordinating ty leaders to stimulate work On computer soft- large-scale R&D programs. The Ministry of Ed- ware; the Information Technology Promotion ucation has lailriched its own 3-year VLSI proj- Association (IPA), for one, had a 2.78 billion ect. Diet (pug liamentary) committees dealing yen budget (about $13 million) in 1980, raised with sciervx and technology have become from both public and private sources. Es- more visib'e. Local governments have started tablished in 1970, IPA organizes programs to court new technology-based industries; through which private corporations and IPA Kawasaki has put together a plan calling for staff conduct joint research on problems such transformation into a "microcomputer city," as computer-aided design or software packages while Hiroshima has organized a council to for small businesses." study the impacts of high technology on its Similarl/theJapan Information Processing established industrial base. Given this prolifera- Development Center (JIPDEC)a semipublic tion, science and technology policy in Japan organiztion with a staff of 150, the bulk of may become more politicized in the years whoare engineerswas established in 1967 wi the support of MITI and the Ministry of "Kijoho no Kaisetsu (An. Explication of the Law for Special P sts and Telecommunications. JIPDEC's pri- Measures for Specified and Information Industries) (Tokyo: Min- istry of International Trade and Industry, 1979). "Denshi Kogyo Nenkan, 1979 (Electronics Industry Annual, ,"Konpula Hakjusho-1979 (Computer White Paper-1979). Nihon 1979. Ministry of International Trade and Industry) (Tokyo: )oho Shari Kaihatsu Kyocini (Japan Information Processing De- Denpa Shuppansha. 1979), p. 303. velopment Associati;:n) (Tokyo: Konputa Ejisha, 1979). p. 94. Ch. 10National Inoustrial Policies 417 mary mission is the marketing of software. tion processing. Likewise, the project is only Loans and grants for some of its programs hive one of a number of follow-ons to earlier MITI- been provided by IPA. Operating with a Si) sponsored activities such as the VLSI R&D pro- million budget. JIPDEC carries out surveys on gram (discussed in ch. 5, as well as below) and information processing, conducts R&D, sup- the Pattern Informatio-i Processing System ports technical training and education, and en- Project (PIPS)." Such R&D efforts complement courages information exchange through sem- one another; they involve shifting groups of inars and publications. Examples of JIPDEC public and private sector participants drawn projects include a microcomputer promotion from a wide range of institutions, In parallel center and an Institute of Information Tech- with the fifth-generation computer project, nology for retraining technical specialists. MITI is sponsoring the supercomputer effort JIPDEC activities also led to the fifth-gen- mentioned earlier, along with a 10-year pro- eration computer project. gram on advanced microelectronic devices and work on optical measurement and control. De- Government-Sponsored Research and spite the funding that MITI provides, the Min- Development Projects istry's officials seldom attempt to guide or The fifth-generation computer effort typifies direct research, but confine their participation Japan's approach to R&Dbringing together to helping shape objectives and to administra- private sector firms, along with selected public tive functions. institutions. With funding from MITI and the Compared with other countries, Japan's ap- bicycle racing association, the fifth-generation proach to aid for electronics is unique in at projectwhich has attracted worldwide pub- least three ways: 1) government-supported pro- licityis overseen by a 22-member panel in- grams are multiple but carefully coordinated cluding representatives from Tokyo Universi- with one another; 2) they are oriented toward ty, companies such as Fujitsu, and MITI.97 facilitating the activities of industry, rather About half the roughly $500 million budgeted than telling industry what to do; and 3) the time for the 10-year effort is to be provided by the horizons are unusually long. The last point is government. A research association (kenkyu critical: the 8- or 10-year planning horizons for kumiai) was set up in 1979 to mobilize nine Jap- many current Japanese R&D projectswith anese companies for R&D on microelectronics every indication that, while projects will be devices and peripheral and terminal equip- adapted to evolving circumstances, continui- ment, as well as softwareall aimed at major ty will be preservedpoint to the depth of the strides in computing technology. JIPDEC's role government's commitment Certainly there are has been largely facilitative; the research as- few analogs in the United States, even in de- sociation now carries the primary responsibili- fense researchwhere the 6-year VHSIC pro- ty. The association's administrative staff has gram is the exception, not the rule. been drawn from employees of the participating companies, who are dividing the R&D Cooperation in Research and Development effort. Observers in the West often misconstrue the As discussed in more detail in chapter 5, the nature of Japan's "cooperative" R&D efforts. fifth-generation computer project is far from While corporate leaders and government ofan independent or all-inclusive effort; its work ficials do in some cases work closely with one is proceeding in a context of government-sub- "PIPS has been much less visible in the United States than sflized R&Das well as company-funded re- several' of Japan's other R&D efforts, but it played a major role searchaimed at related aspects of informa- in laying groundwork for the fifth-generation computer project. See H. Nishino, "PIPS (Pattern Information Processing System) ""Fifth Generatimi Computers," JIPDEC Report, Japan Infor- ProjectBackground and Outline," Proceedings of the 4th In- mation Processing Development Center, summer 1980. The dis- ternational Joint Conference on Pattern Recognition, Kyoto, Nov., cussion following also draws on interviews with MITI officials 7-10, 1978, International Association for Pattern Recognition, in the Information Machine Industries Bureau. p. 1152. 421 :116 International Competitiveness in Electron,cs another during ongoing projects, the more search.99 Individual firms did not cooperate on usual pattern has been a carefully planned divi- either product designs or processing technol- sion of Libor. MITI bureaucrats help initiate ogy. Thus, while the MITI-sponsored VLSI new projectsafter lengthy preliminary discus- project has become known abroad as a "coop- sions with industry advisory committeesby erative" effort, the actual extent of interaction winning budgetary approval. They also mon- among participating firms was limited; spokes- itor ongoing programs, evaluating progress and men for the Japanese electronics industry say judging success. Government officials are often that dividing the research enhanced the overall detailed to organizations like JIPDEC. Program success of the project. It appears that the or- administration is normally delegated to repre- ganizational form involved a compromise be- sentatives of participating firms, with the tween attempts to encourage individual inter- r,, search itself divided among these firms. Peo- actionswith objectives str.-..h as stimulating pio from different companies seldom work side personnel developmentand the more con- by side. crete technical goals. Certainly as the work undertaken by joint R&D projects in Japan The two government-supported VLSI proj- moves toward development, interfirm coopera- ectsparalleling one another in timeillustrate tion declines; a MITI-orchestrated follow-on to these patterns. The first, oriented toward com- this VLSI project, which began in 1980 and em- munications, was carried out by the public cor- phasizes chip designs and applications, takes poration Nippon Telegraph & Telephone (NTT) the fcrm of totally independent efforts by each under the aegis of the Ministry of Posts and participant. Telecommunications. The second, directed at applications of ICs to computers and much bet- The work of the "Research Association for ter known outside Japan, was sponsored by R&D on New Function Elements," also begin- MITI; with 40 percent government and 60 per- ning in 1980, can be viewed as another follow- cent private funding, the $300 million, 4-year on to the VLSI project; it illustrates the way in effort took the form of a research association which MITI-sponsored research efforts com- linking five participating firms. Three lab- plement one another. This association's labora- oratories divided the work: a shared facility tory draws on a larger group of companies. managed by the VLSI Technology Research As- Matsushita, Sanyo, Sharp, Oki, and Sumitomo sociation; the Computer Development Labora- Electricnone as strong in their technology as tory jointly run by Hitachi, Fujitsu, and Mit- the five companies that had participated in the subishi; and the NEC-Toshiba Information Sys- VLSI projectwill all be involved in one tems Laboratory. Staffs of the latter two lab- more of three major microelectronics develop- oratories came, not from the larger group of ment efforts.199 These are: participants, but from the companies operating Three-dimensional circuit elements them; the joint facility drew engineers and sci- which can be visualized as more or less entists from all five, as well as MITI employees conventional ICs stacked atop one an- from the AIST. MITI was deeply involved in other, increasing the density. planning and organization during the prelim- High electron mobility transistors inary stages. Later, the teams from the par- (HEMTs), one variety of which consists of ticipating companies independently carried out their assigned research tasks. Only in the extremely thin layers of semiconducting association's joint laboratory was a real effort 9,0Interview with Mr. Nebashi. IBMJapan and formerly at the at cooperationwith technical people from dif- VLSI Cooperative Laboratory, Nihon Keizai Shim bun, Jan. 19, 1981, p. 1. ferent companies working togetherunder- 100"FY82 Government Projects in Electronics Listed." Japan taken; this was a minor portion of the overall Repoil, Joint Publications Research Service JPRS L110676, July program, restricted to more fundamental re- 22, 1982, p. 55. Ch. 10National Industrial Policies 419

materials such as gallium arsenide or would an attempt at such an accounting be gallium aluminum arsenide; these struc- very meaningful. Indirect benefitse.g., tem- tures carry the potential for higher switch- porary exemptions from antitrust provisions ing speeds. hence faster computers. escape Quantification. Even when government Radiation- hardened devices suitable for funds flow directly to industryas in the cost- use in extreme environments such as sharing typical of joint R&D projects in japan, nuclear powerplants or outer space (resist- or the subsidies for the West German computer ance to heat and vibration is a related industry during the 1970'sthe real questions objective). concern the effectiveness with which the The first two especially will support both the money is spent. supercomputer and fifth-generation projects. Nevertheless, subsidies deserve special attention in the case of Japan because the U.S. elec- The Role of Universities tronics industry has argued that they have been a key to the competitive success of Japanese University-industry interactions in R&D are firms. Research funding is only part of the total nu closer in Japan than in other countries picture of industry-specific support, but as //again perhaps in some contrast to the common table 77 indicatesalid in common with other / perception. Close collaboration is rare, even industrialized countriesmore than a quarter though the rules prohibiting professors in the of all Japanese R&D expenditures, both govern- national universities from working for private ment-funded and industry-sponsored, have companies can be circumvented. Contract re- gone to the electronics/electrical machinery search and consulting by university faculty are sector. At the same time, government expend- more limited than in the United States. itures on research are not high compared to Japanese policymakers universary express other countries; considering only R&D, and the wish that university-industry relations be counting only expenditures directly related to improved, and that sufficient numbers of well electronics, public funding is quite small trained professionals be available to meet the about 1 percent of the total for 1978, according economy's needs. To date, however, little prog- to the Japanese Government.702 This is hardly ress seems to have been madenor, in fact, the whole story; it does make the point that have new policy initiatives directed at such R&D in Japan is primarily the responsibility of concerns emerged. As discussed in chapter 8, private industry. Japanese R&D is heavily con- Japan's colleges and universities have for some centrated on commercial applications; neither years been turning out more engineering grad- military technologies nor basic research get the uates than in the United States. Nonetheless, attention they do in other countries. Looking as in other industrialized countries, there has at all R&D spending, the private sector in Japan been concern over future shortfalls in the sup- provides over 70 percent of total funding ply of engineers and scientists; a resent survey more than in the United States, where industry covering the hiring plans of more than 1,600 spending accounts for 50 to 60 percent (table Japanese.firms points to stiff competition dur- 77). ing the-1980's for university graduates trained MITI's annual compilation of government in technical fields. 101 supports and subsidies for the "information industry" is the most comprehensive listing of How Significant Are Supports and nuln 1978, government bodies in Japan, including state and Subsidies in Japan? local, reportedly contributed 6.8 billion yen (about $34 million) to the total of 580 billion yen (about $3 billion) spent for R&D As for any country, it is impossible to place or, "electrical machinery." This includes household el'ectric a monetary value on the policy measures that equipment. as well as communications and electronics. Most benefit Japanese electronics companies. Nor of the R&D work is for development. SeeKagaku Gijutsu Kenkyu Chose(Report on the Survey of Research and Development, 101"Daisotsu Danshi Nobi Niketa" (Number of Male Graduates Prime Minister's Office. Statistical Bureau) (Tokyo: Nibon Tokei Declines).Nihon Keizai Shimbun,Aug. 27, 1981. p. 1. Kyokai. 1979), pp. 39-40, 94.

423 420 International Competitiveness in Electmni:s programs related to electronics. (For the United ogies e.g., health care, regional energy saving, States, no comparable data existin part be- computerized traffic control systems. The fig- cause no one agency has responsibility for such ures in the table also include money for con- programs). For 1980, the Japanese Government ducting surveys on computer usage, adminis- budgeted about S1.3 billion toward the develop- tering qualifying examinations taken by com- ment of the information industryexpendi- puter technicians, and the costs to the govern- tures encompassing much more than just the ment of special tax deductions extended to R&D programs highlighted above; a large frac- companies that train information processing tion of the total consists of loans and loan specialists. guarantees rather than direct grants.103 In- Taken together, it is the comprehensive na- cluded in the total, for instance, is the more ture of such programsnet their spending lev- than 5200 million that the Japan Development elsthat distinguishes Japan's policies toward Bank loaned to the Japan Electronic Computer electronics and other targeted industries.'" The Corp. for lease financing; this aids Japanese very fact that the government publishes an in- computer manufacturers by reducing the funds formation industries budget indicates the care they would otherwise have to commit to rent- with which the bureaucracy monitors develop- al and lease arrangements with their cus- ments in electronics and disseminates infor- tomers. as well as absorbing risks associated mation among government, business, and fi- with repurchasing. nancial circles. It is this attentiveness on the The computer industry has received a sub. part of government, and the fact that most pro- stantial share of direct subsidies. Budgetee grams are coordinated by MITI, that sets Jap- MITI expenditures for major projects closely anese industrial policies apart. Over the years, related to data processingincluding several funding by the Japanese Government has of those outlined aboveare listed in table 80. grown, but the significance of MITI's initia- The table is not inclusive, and is intended only tives goes well beyond financial support; in- to give an idea of the magnitudes of typical gov- deed, to look only at the money spent is to un- ernment expenditures. These sums are not large compared to R&D spending by industry "The "Research and Development Project of Basic Technol- itself in either Japan or the United States, or ogies for New Industries." established in late1981.is another example. The original plan called for total spending of about in comparison with government funding in $460 million over 10 years: however, the first year's expenditums other countries, Portions of such subsidies have have been scaled down by the finance-conscious MM. Private funded large-scale. long-term programs aimed corporations are being funded to participate in one of 12 R&D "themes." such as biotechnology and advanced materials. The at social applications of electronics technol- "New Function Elements" microelectronics projects mentioned earlier are also part of this umbrella program. See "A IST 1982." Agency of Industrial Science and Technology. Ministry of Inter- thWenshi kogyo Nenkan 1979,op.cit.. p. 340. na.tional Trade and Industry. Tokyo. pp. 6 and 7.

Table 80.Japanese Government Expenditures on Selected Projects Related to Computer Technology Budgeted expenditure (millions of dollars)a Prgject 1981 1982 Basic technology for nextgeneration computers $28 $22 Basic software technology 24 20 Microelectronics ("new function alements") 3.1 4.5 Supercomputer, R&D 0.14 3.3 Peripherals 4.8 2.6 Fifthgeneration computer R&D 0.07 1.7 $60.1 $54.1 ariscalyear basis. convened from yen at 220 to the dollar for 1981, 249 for 1982. SOURCE: "PY82 Government Projects in Electronics Listed," Japan Report.JointPublications Research Service JPRS L10678,

424i Ch. 10National Industrial Policies 421 derestimate the impacts of such programs. tional moves by Japanese electronics manufac- They have considerable symbolic and psycho- turers have for many years had the active sup- logical value in galvanizing the efforts of many port of Japan's Government. participants behind a set of goals shared by In the United States, many signs indicate that government and industry. Programs in elec- Japanese manufactuiers are now often recog- tronics have typically been aimed at breaking nized as peers, Technical exchange agreements bottlenecks viewed as critical to continued between American and Japanese electronics progress. Both the VLSI projectwhich, as out- companiesrather than outright purchases by lined in chapter 5, was intended to help 7apan Japanare on the upswing. Mitsubishi and catch up to the United States in digital MOS Westinghouse have arranged a joint venture to ICs=and the fifth-generation computer project, design and manufacture ICs. Hewlett-Packard with its software push, have been designed to is getting RAM technology from Hitachi. The serve such purposes. The supercomputer proj- U.S. Department of Defense has persuaded Ja- ect is quite different; not particularly. impor- pan to transfer defense-related electronics tant in Any commercial sense, it is first and technologies to this country (although what foremost intended as a highly visible symbol these technologies will consist of is far from of Japan'', ability to compete technologically clear). American semiconductor firms are set- with the United Statesfrom the Japanese per- ting up design centers in Japan, as well as prospective, supercomputers are one of the critical duction faciliteswhile Japanese firms do like- propaganda battlefields of the "computer war." wise in the United States, each seeking to draw Comparing Japan's industrial policy with ef- on the other's technical talent. forts in Britain or FranCe points to a major dif- Movement toward cooperation amidst on- ference: government policies in Japan are di- going commercial rivalries has not been con- rected at further strengthening a private sec- fined to the initiatives of private companies. tor that is vital and still expanding rapidly, not In response to criticism from the United States at revivifying a stagnant industry. Government and elsewhere that MITI-sponsored electronics programs in Japan complement the dynamism R&D constitutes an unfair subsidy, Japan has and international orientation of the country's suggested steps in the direction of international electronics firms; they have contributed to, but cooPeration. For example, foreign firms have not created, their competitive ability. been invited to participate in discussions aimed at an enlarged fifth-generation computer proj- Recent Trends ect having the form of an international joint Majc. thrusts of Japan's industrial policy venture.'" SuOh proposalseven if carridd have been aid and encouragement for exports through,would not by themselves stem the ris- of electronics and, to a lesser extent, overseas ---inglide of criticism aimed at Japan's industrial investment. The international activities of Jap- policies, as well as the country's indirect and anese electronics firms are especially visible nontariff barriers to trade. Still, if nothing else, in Asian markets, where interdependence is they are a sign of the confidence the Japanese growing (ch. 4). A study by the Electronic In- now:have in their own abilitieswhile also be- dustries Association of Japan forecasts strong ing a well-calculated public 1.1acions ploy. expansion elsewhere in the Far East, and urgers There ard two fundamental perspectives in Japanese firms to develop strategies of "accom- United States on questions of J6, an's sub-: modation " promotingJapanese investment sidies and indirect trade barriers. the one and technology transfer, while importing low- and, those who believe free flows o technol- technology, labor-intensive electronics prod- ogy to be a prerequisite for economy growth ucts from other Asian nations.'" -These interna- I05Denshi Sangyo no Kokusaika rio Hoko to sono Eikyo ni Khn- 1°6"Dai GoSekai no Conputa: NichiBeiQo no Kyodo Kaihatsu sun, Chosa Hokoku(Survey Report on Trends in the Internation- Shido" (Fifth-Generation Computer: Beginning of Joint U.S.- alization of the Eloctronics Industry and Their Influence, Part Japan-West European Joint DeveloPment).Nihon Keizai Shim- II on East and.Southeast Asia). op. cit., pp. 271-291, bun.Aug. 30. 1981. 425 422 International Competitiveness in Electronics and technical innovation call for equal access examination much of the appearance of inter- by U.S. firms to programs sponsored by the firm cooperation vanishes. The program was Japanese Government. The other view is held effective because it was carefully craftcd to by those who would prefer restrictions on out- help Japanese firms overcome specific weak- ward flows of U.S. technology in an effort to nesses that MITI and industry leaders had yeserve "technological security." As debates identified: emphasis on linear devices, a legacy in this country continue, the procurement, of production for consumer products; lagging R&D, and customs and standards activities of capability in the processing of large-scale ICs, Japan's Government will be scrutinized by because Japanese firms were dependent on partisans of both viewpoints. semiconductor manufacturing equipment from the United States; lack of experience in digital How Effective is Japanese Industrial Policy? circuitry among engineers and technicians. In contrast to governmentsupported R&D proj- Any judgment of the contribution of Japan's ects in West Germany or the United Kingdom, industrial policies. or government policies the Japanese were able to define their needs anywhereto international competitiveness in and agree on a program that would help them electronics rests in part on intangibles. Precise catch up to the United States. It is the consist- evaluations are impossible. What isthe ent and coordinated attentiveness to the prob- "worth" of the networks for information trans- lems and potentials of electronics (and other mittal.and consensus-building woven by MITI? industries) that dish_ guishes the policies of What are the costs and benefits of the am- MITI and the rest of Japan's Government more biguities and uncertainties surrounding an- than the character of individual programs or titrust enforcement in the United States? policy instruments. In judging the effectiveness of Japanese in- At a more general level, the long-term orien- dustrial nolicy, the starting point is its basic tation of policies toward electronicstypified thrusti -) cultivate rather than confine the na- by the fifth-generation computer project--also "tion's electronics companies.. The institutional distinguishes Japan from other countries. Fur- apparatus that has evolved over the years has ther, development of the electronics industry contributed far more than absolute levels of while a goal in itselfhas been pursued for financial assistance might indicate. The end larger reasons: electronics is viewed as the key result' has been effective mobilization of institu- to Japan's overall industrial development, the tional and human resources, 'comprehensive- first ingredient in the knowledge-intensive, en- ness in government efforts, a substantial degree ergy-efficient economy that the country's tech- of policy integration without rigidity. The focus nocrats are striving toward. of Western observers on cooperation between government and business only hints at how the Japanese industrial policies have certainly system works. not been univeital tri u phsefforts to prop up declining.sectorsTsY1) or to counter inter- With few exceptions, Japan's Government national market trends (petroleum) have not has used tho same policy tools to promote elec- been particularly successful. But policies tronics as other nations: in the early years, toward electronics have complemented the dy- tariff barriers combined with controls on for- namism of private companies already well po- eign technology and capital flows; today, sup- sitiOned both domestically and international- ports and subsidies for R&D and commercially. It is the congruence of public policy and -ization While the highly publicized VLSI R&D evolving shifts in industrial structure that, in project has been held out as a unique instance of cooperation; one that would violate anti- the end, is the hallmark of present-day Japanese trust laws in the United Statesunder closer policies toward the information industry. Ch. 10National Industrial Policies 423

Summary and Conclusions

Among nations that have set out to promote Can, then, industrial policies create com- their electronics industries, the policy tools parative advantage? The answer is clearly no. come from a common list: R&D funding, in- Competitive success in electronics, here and vestment grants and subsidies, procurement, abroad, depends on many factors, of which merger and trade policies. No avenue emerges government actions are only one. Taken alone, that can guarantee success in strengthening the public policies are seldom as important as the competitive ability of a country's electronics capabilities of a nation's private companies: firms. Under closer scrutiny, many of the pol- human resources and their utilization, includ- icies adopted by nations like Japansorrietimes ing the quality of management; costs and avail- thought to be unfair or uniqueare not so dis- ability of capital; technological ability in elec- similar from those used in other advanced in- tronics and the complementary infrastructure; dustrial economies, even the United States. overall market conditionsthese are more cen- Matters of timing, comprehensiveness, con- tral to international competition. Public pol- sistencyrather than, the types of policies icies can add or substract from them, but the adopteddifferentiate the industrial policies of ability of governments to compensate for weak- various countries. nessesor to reverse declines in competitive- nessis circumscribed. Although they can As competition in the international elec- either help or hinder industrial development, tronics industry has intensified, governments public policies alone do not determine have stepped in to help their own entrants. In directly or indirectlythe competitive standing the early 1960's, European and Japanese fears of electronics industries in any nation. over the "American challenge" sparked syste- Today, policyrnakers in the U.S. Government matic attempts to protect and strengthen do- must decide whether to continue the ad hoc mestic computer manufacturers. At that time, approach of years past or move toward meas- the preferred policy approach began with trade ures aimed more consciously at preservingand protectiontariffs, controls on flows of foreign strengthening the advantages that the Amer- investment and technology, discriminatory ican electronics industry draws from its setting procurements. Several countries encouraged and structure. If the choice is to develop a more mergers among computer firms. In the 1970's, comprehensive industrial policy, much can be as trade liberalization underthe General Agree- learned from studying foreign experience ment on Tariffs and Trade continued, indus- West Germany's Fraunhofer Gesellschaft, trial policies shifted away from overtly protec- Japan's VLSI project, Britain's schemes to pro- tionist and defensive approaches. Today, sup- mote commercial applications. But no recipe ports Thr R&D, indirect subsidies such as tax for success emerges from the countries that incentives, and other less direct measures com- have experimented with industrial policy. It is prise the fou.idations of public policies toward one thingto-,say that policies toward elec- electronics in virtually all countries. Except in tronics should'be in tune with overall changes threatened sectors like consumer products, in industrial structure and international mar- trade liberalization has been accompanied by kets; it is quite another to actually design and a parallel movement towardpolicies with sec- implement an effective industrial policy amidst', ondary rather than primary effects on interne- the ongoing uncertainties and ambiguities that - tionalflows of electronics goods. If there is an the political and economic con- exception, it is the United Stateswhere, leaving aside defense-related policies, the most text. prominent measures have continued to be reg- Governmeip poliCies, thenas illustrated by ulatory. the countries examined in this chapterare

427 424 International Competitiveness in Electronics generally tailored to the level of technological labor supplied by a long-established educa- and commercial development of firms in the tional system is only one example. Structural local industry. It is no coincidence that the na- features of the political and economic system tion which led initially in semiconductors, in in Japannatural resource endowments, ex- color TV, in computersthe United States isting capital markets, political stability, estab- made no attempt to devise a systematic policy lished mechanisms for policymaking, charac- orientation toward electronics. Where was the teristic systems of labor-management rela- need? Nor is it surprising that the countries tionshave .tended to shape and limit indus- with more comprehensive policies have gen- trial policy decisions, rather than the other way erally been those that have perceived them- around. selves at a disadvantage. But why have some A9 japan and other countries seem likely to countries been more successful in mobilizing discover, it may be easier to develop policies institutional resources to create sustained and aimed at catching up than to devise strategies coordinated industrial policies than others? for keeping up or jumping ahead. For one A variety of forces work to enhance the abili- thing, as internationalization of industrial and ty of government officials to design and imple- market structures proceeds, the influence of ment coordinated, timely, and comprehensive national governments will diminish. But the policies toward industries like electronics. A fundamental point is that in any 'industry or relatively centralized policymaking apparatus, technology, creating a new model is harder where a single agency or a select few have well- than following a recognized leader. Goli- defined' responsibilities, is one. The grab-bag ernment aid has helped electronics firms in nature of British policies mirrors the agencies other countries improve relative to American charged, at one time or another, with policy competitors; the situation for the United States development. In countries where government has beenand remainsdifferent. The leaders, officials belong to a respected civil service they be they American or Japanese, have to break are more likely to have the resources to analyze new grounda commitment that industrial and initiate actions with positive effects on in- policymakers in Japan have long since -ffiade. dustry. The dominance of political appointees Japan's publicly voiced determination to im- in the United States, and their rapid turnover, prove the technological base for the country's works against the kind of consistency seen in electronics industry stems from a recognition nations like Japan. So too does the laCk of un- that past successes have been built on the adap- derstanding of technology characteristic of tation and commercialization of technologies both bureaucrats and politicians in this coun- originating elsewhere, mostly within American try. An elite civil service does not ensure suc- firms. Now, Japan is a leader along with the cess, as the mixed record of French industrial United States. The public as well as the private policy shows. But especially in Japanwhere sectors in each face the need to devel- consultation and cooperation between industry op appropriate strategies for the years ahead. and government have been closer than in many Industrial policies for the 1980's and beyond other countries, if not so close as sometimes will be most successful where policymakers pictured in the Westconsensus is easier to grasp the dynamics of ongoing shifts In domes- achieve than in nations where adversarial rela- tic and international markets and industries. tions are the norm. Easier too is carrying To the degree that public policies ignore or at- through the actions that have been agreed on. tempt tocounteract,such forces, they will be Such factors have enhanced the effectiveness less likely to reach their objectives. Policies of industrial policy in Japan, the one nation that designed to complement and reinforce ongo- liar; so far managed to catch upin at least ing trends will be more likely to have positive some respectswith the United States. Of effects. This is not to say that-public polic s course, Japanese electronics firms have been cannot help shape these trends. If it is trueat favored by other circumstances as wellskilled the industries which fueled postwar ecomic

2 d Ch. 10National Industrial Policies 425 growthsteel, petrochemicals, automobiles well as electronics and information procesSing have attained a stage of relative maturity, then will contribute to growth in existing as well emerging technologies are indeed an appropri- as new industries. It follows that the appropriate focus of government policy. Technologies ate emphasis of public policy may not be elec- based on genetic engineering, advanced ma- tronics alone, but economic adjustment and terials, computer-integrated manufacturingas technological development more broadly.

423/

99-111 0 - 83 28 CHAPTER 11 U.S. Trade Policies and. Their Effects Contents

Page Overview 429 Tariffs; the Multilateral Trade Negotiations 430 Tariff Effects 431 Other Tokyo Round Agreements 436 Dumping 438 The Law and Its Administration 438 The Color Television Case 439 Prospects for Dumping Actions Elsewhere in Electronics 441 Subsidies and Countervailing Duties 442 Countervailing Duty Law and Its Administration 443 The Question of Indirect Taxes 444 Other Unfair Practices 445 Quantitative Restrictions and the Escape Clause 446 Orderly Marketing Agreements for Color Television Imports 446 Escape anise Proceedings in Color Television 447 Effects of Quotas and Other Nontariff Restrictions 449 The Escape Clause 449 Prospective Effects of U.S. Trade Policy on the Electronics Industry 4501 Consumer Electronics 45'1 Semiconductors 45;4 Computers 457' Summary and Conclusions 459

Table

Tdbit X. Page i 1. Itypothetical Example Illustrating Tariff Effects on a Product With Nominal Nlanufui:Inring Costs of 51,000 ; 431 CHAPTER 11 U.S. Trade Policies and Their Effects

Overview All trading nations develop policies dealing beneficiariesmany of whom do not realize with imports and exports. On, the import side,' they are paying less for some of the goods and such policies are usually intended to control services they purchase. Because of this imbal- flows of incoming goods judged harmful to the ance, governments often raise barriers for pp- domestic economy. Formalized export policies, litical reasons, sometimes creating serious as a general rule, are less numerous and typ- disruptions. The familiar example is the Smoot- ically intended to encourage overseas sales. To Hawley Trade Bill, adopted by the United the extent that international commerce is re- States in 1930, which raised the average U.S. stricted to trade and its financing, countries tariff to more than 50 percent and was one must export to be able to import, and vice ver- cause of a steep decline in world trade. More sa. Over time, exports will therefore approxi- recently, 'Japan has utilized a wide variety of mately equal imports. For such reasons, trade tariff and nontariff barriers to protect develop- policies seldom have first-order effects in deter- ing industries, including electronics. mining overall levels of imports and exports, but tend to guide and regulate tradeinfluenc- Near the end of World War H and after- ing, for example, the composition of a nation's wards, the United States took the lead in efforts imports. Policies can also be adopted to en- to establish a liberal world trade order. This courage exports so that needed importso.g., commitment has continued uninterrupted to oilcan be paid for. Most common remain im- the present day. American leadership has been port controls serving to limit threats faced by a major force in negotiations among trading domestic industries. nations aimed at moderating tariff and, more recently, nontariff barriers to trade. These ef- In recent years, the governments of indus- forts have taken place largely within the struc- trialized nations have, as matters of official ture of the General Agreement on Tariffs and policy, generally taken the position that unre- Trade (GATT), an organization now compris- stricted tradeor at least trade with minimum ing some 80 nations. GATT provides a forum impediments in the form of tariffs or similar for negotiations together with mechanisms for restrictionsbenefits all countries. Although resolving conflicts. a principle often honored in the breach, nations usually assume that relatively open trade is in While trade negotiators have made :consid- their self-interest. Countries import goods erable 70f; 1'C% in redthcing tariffs, nontariff which they themselves cannot produce as ef- measuiu,, -,Are proving less tractablewithin ficiently, and export products in which they GATT or on a bilateral basis. As more nations have a comparative advantage ich. 5).In develop industrial policies nominally for do- theory, everyone is better off. mestic reasons, the trade arena has taken on a new complexion: indirect and nontariff bar But while the benefits of open international riers have risen. as tariff walls have declined. trade are spread widely across society, the The result has sometimes been termed "the costs against which they are arrayed _tend to new prOtectionism." In essence, negotiators be concentrated. .Individual companies, their are struggling to fit the policy framework from employees, the cities and regions in which they an earlier eraGATT mechanism's have roots are located, bear the brunt of shifting patterns in the 1940'sto a radically different setting. of trade and competition. When imports rise, International corporations now compete in the injured parties are more vocal than the some parts of the world, cooperate in others,

432 429 430 International Competitiveness in Electronics ship goods between subsidiaries located in ternational markets. Much of their success has dozens of countries, and take advantage of na- been due to a global perspective and worldwide tional industrial policies where they can. Gov- operationsneither of which would have been ernments design policies to attract foreign in- possible without the open environment for vestment and technology under some circum- trade and investment created since World War stances, to keep it out under others. Trade II. To be sure, foreign countries have often related complaints by U.S. firms embrace not adopted policies intended to restrict inflows of only the old-style unfair practicesdumping or American-made electronics products. But in export subsidization to boost trade balances, most though not all cases, such restrictions predatory practices aimed at building monop- have had effects that were marginal or indirect olies or cartelsbut asymmetries in the "rules or both. While trade barriers have sometimes of tt.a game." The claim is that the industrial encouraged U.S. firms' to establish overseas and trade policies of other nations tilt the rules manufacturing facilities, for many years Amer- in their favor. Trade.negotiators will be faced, ican electronics companies had such advan- for years to come, With adapting rulemaking tages in technology and cost that they would and adjudicating procedures to these new have been potent competitors virtually regard- realities. less of the trade policies adopted by other nations (the principal exception has been Japan). This chapter briefly reviews the environment Still, these advantages have gradually dimin- for international trade in electrOnics under ished over time. GATT, then disCusses the trade policies of the United States, particularly as these relate to the Where technological change is less rapid and electronics industry. Only limited attention labor costs more significant, trade policies goes to other countries. The'chapter illustrates carry more, weight. In such products as televi- impacts of trade policies and discusses policy sion receivers, CB radios, and passive com- directions that may be important in the future. ponents, U.S. firms have-not been able to main- On the whole, the U.S. electronics industry tain advantages in technology or manufactur- has been helped by the Federal Government's ing cost. Here, liberal U.S. trade policies have trade initiatives during the postwar period. made it more difficult for American firms to Semiconductor and computer firms, in par- compete effectively most notably inthe ticular, have benefited from the opening of in- domestic consumer electronics market,

Tariffs; the Multilateral Trade Negotiations

The General Agreement on Tariffs and Trade the rest of the decade (another round of multi- provides the basic context for negotiations lateral trade negotiations before the end of the among nations concerning trade,and, where 1980's is unlikely). This Multilateral Trade needed, for adjudicating disputes. Other Agreement was implemented in the United bodies, including the Organization for Eco- States by the Trade Agreements Act of 1979.' nomic Cooperation and Development (OECD) Earlier negotiations under GATT had fo- and the United Nations, play more limited roles cused on tariffs; although the Tokyo Round e.g., collecting statistics. GATT isthe pri- MTN resulted in further cuts, negotiators con- mary vehicle for multilateral trade negotiations centrated on such matters as quotas, customs (MTNs), the latest of whichthe so-called procedures, product standards, and public sec- Tokyo Round, concluded in 1979resulted in tor procurement practices. Examples of Tokyo an agreement which will bethe principal framework for international trade over at least 'Public Law 96-39, July 26, 1979. Ch. 11U.S. Trade Policies and Their Effects' 431

Round topic!: qf special relevance for trade in Tariff Effects electronics include: As taxes on imported goods, tariffs directly A revised subsidies code, intended to, pro- affect price competitiveness. From the view- vide a framework for dealing with national point of the country imposing them, tariffs c'in industrial policies having the indirect ef- Nerve multiple purposes. One effect is normally fect of subsidizing exports or otherwise af- to raise domestic prices; tariffs permit local fecting trade flows (as by giving domestic firms to manufacture at higher costs while re- products advantages over imports). maining competitive in the marketplace, pro- (---\. Staging of tariff reductions for semicon- tecting domestic industries from foreign rivals. ductors. accelerated by Japan in 1981 after Alternatively, governments impose tariffs to extensive bilateral negotiations with the counter unfair trade practices such as dump- United States, a similar acceleration of ing or export subsidies, or to retaliate against Japanese tariff reductions on computers restrictions by other nations. following a year later. An agreement on government procure- The impacts of tariffson trade patterns are Ment, where again negotiations between not always so straightforward as the nominal the United States and Japan concluded, at percentage rate would indicate; "real" rates of the end of 1980, in a bilateral accord more protection may exceed nominal rates by signifi- liberal than that arrived at under the MTN cant amounts. Table 81 gives a hypothetical but framework. not unrealistic examplea product (which might be something like a computer Aerminal) In the United States, passage of the Trade with a nominal produaior` cost of $1,000, pur- Agreements Act of 1979 was accompanied by chased components constwating 80 percent of a reorganization of trade-related activities car- this, final assembly the remainder. The table ried out by executive order. As discussed be- compares two cases: 1' final assembly overseas, low, responsibility for dumping and counter- with the complete ,tc in imported and sub- vailing duty investigations moved from the ject to itari,of dercent; and, 2) final Department of Treasury to Corn, fee, while assembly in the United States, with compo- a new Foreign Commercial Service was estab- nents imported at a tariff rate of 5 percent. In lished in the Department of Commerce in place both cases, the components are assumed to be of the commercial officers attached to the purchased abroad at the same cost. (Transpor- Department of State. At the same time, the Of- tation costs are ignored.) As shown, assembly fice of the U.S. .Trade Representative was given in the United States gives a cost advantage of the job of coordinating international trade ne- $60. The real protective effect with respect to gotiations on a continuing basis. This reorga- the operations carried out domesticallythe nization followed mounting criticism of the "effective rate"would then be $60/$200, or fragmentation and diffusion of responsibility 30 percent. This percentage can be interpreted for trade matters within the executive branch. as the amount by which domestic costs of as-

Table 81.Hypothetical Example Illustrating Tariff Effects on a Product With Nominal Manufacturing Costs of $1,000

Foreign assembly U.S. assembly Cost of components $ 800 $ 800 Tariff on imported components (5%) 40 Cost of assembly 200 200 $1,000 $1,040 Tariff on imported system (10%) 100 Total cost in the United States $1,100 $1,040 SOURCE: Office of Technology Assessment. 432 International Comootitivenes in Electronics semhly could exceed foreign costs before from 6.6 to 5 percent., Stagingthe sequence American firms would begin to lose competi- of stepwise reductionsvaries by product; the tiveness. As a result, even where nominal tariff most common pattern is yearly cuts over the rates are identical, protective effects can dif- per:od 1979-87 of about one-eighth the total fer; each case must be considered individual- negotiated concession. Likewise, duties on of- ly. The example in table 81 is not atypical in fice and computing equipment will fall from that tariffs or, parts and components gen- araverage of 5.4 to 3.6 percent. In certain, erally 'owe' Wm a tariffs on final pfuducts; cases, the United States did not grantreduc- wbrve this is kilo case, effective tariffs are tions.`Jot surprisingly, these were generally aliv,!ys higher than nominal tariffs. products where imports have caused problems for domestic manufacturers. Tariffs on color Tariff Changes in the Tokyo Round MTN TVs, for example, will remain at the current and After level of 5 percent. Indeed, for items subject to section 201 escape clause findings discussed Nominal tariff levels on electronics productsbelow), of which this was one, U.S. negotiators vary a good deal, with the Tokyo Round re- had no authority to offer concessions. sulting in significant changes for microelectronic devices and computers. A:, :.%entioned Tariff reductions agreed to by countries above, tariffs on both semiconductors,,,1.com- :11 have been important export markets for puters weir 111i! subject of bilateral negotiations American electronics firms were generally between the L itited States and Japan subse- somewhat largerthough with important ex- quent to the multilatern= agn eruortt ftl ceptions. Many nations have maintained con- 1981, at =reed to Ince tiffs siderably higher tariffs than the United States; tegrated .rcA.4i (ICs) to 4.2 percent as of the shipments of ICs into the European Communi- beginning of 1983. Originally, they were to ty (EC), for instance, have been taxed at 17 Lave dropped in stages, reaching the 4.2 per- percenta duty that the Europeans declined cent level only in 1987. U.S. tariffs on ICs went to reduce.4 The tariff wall has been steep from 6 to 4.2 percent in 1982. Somewhat later, enough that both American and Japanese firms as part of a larger package of tradeconcessions, the Japanese Government announced a parallel 'MTN Studies, Vol. 6, Part 5. Agreements Being Negotiated reduction in tariffs on computers. The cuts, at the Multilateral Trade Negotiations in GenevaU. S. Inter- film 7 to 4.9 percentthe U.S. levelwent into national Trade Commission Investigation No. 332-101, Subcom- effect at the beginning of 1983, rather than in mittee on International Trade, Committee on Finance, U.S. Senate, August 1979, p. 251. Computer parts, as well as periph- 1987 as again originally schediiled.2 eral equipment, can be imported duty-free from some countries as a result of the Generalized System of Preferences (GSP), under As part of the Tokyc Round, the United which the United States, the European Economic Community. States granted a variety of tariff concessions and Japan have agreed to give preferential tariff treatment to on imports of electronic products,but these products manufactured in developing nations. However, imports of such products into the United States under the GSP are ex- cuts will not have much impact because most pected to remain small. Of those that do enter this country, many U.S. tariffs were already low. The reductions originate in Americaff-owned facilities such as Texas Instru- seldom amounting to more than a few percent- ments' plant in El Salvador. 'A group of nations that did not join the European Communi- age pointswill make littledifference in tyincluding Austria. Switzerland. Portugal, and several of the landed costs of imports. For example, the Scandinavian countrieshave formed the European Free Trade average level of tariffs on components (in- Association, EFTA. In contrast to the ECwhich has common cluding passiv6 devices such as resistors and tariffs on importseach EFTA member sets its own duty levels. Once inside an EFTA country, however, goods can move free- capacitors, as well as semiconductors) and ly within either EFTA or the EC without further tariffs. To keep telecommunications equipment will decline exporters from channeling all goods through the EFTA member with the lowest duties, the Association has adopted a complex set of rules of origin. U.S. firms have sometimes charged that 1J. Robertson, "Japan Offers To Speed Up Tariff Cuts." Elec- these rules are significant trade barriers. See Consumer Electron- tronic News. May 31, 1982. p. 1. ics Market in Europe (London: Frost & Sullivan. Inc.. 1978), p. 95.

3,F Ch. 11U.S. Trade Policies and Their Effects 433 have built plants within the EC to avoid it. Secondary Effects of Tariffs European countries did cut tariffs on a variety of other electronic components and on In addition to raising the costs of imports communications equipmentbut for commu- compared with domestic goods, tariffs can nications especially, nontariff barriers remain have a variety of less direct impacts on trade a strong impediment to trade. Average EC tar- and production; for instance, they may stim- iffs on office and computing equipment will ulate local investment by foreign manufactur- drop from 6.9 to 4.9 percent. Overall, the Com- ers seeking to avoid the extra costs borne by munity's reductions will have little effect on imports. The complex patterns of U.S. direct competitiveness because American electronics investment in electronics have been shaped by products generally had significant price (or tariffs among many other factors. Foreign elec- technology) advantages in the European mar- tronics firms have also invested in the United ket even at the-old tariff levels. The European States, particularly in the consumer sector; case Is a general one: reductions in tariffs by European and Japanese firms hold majority or other countries will seldom have large net ef- partial ownership positions in U.S. electronics fects on U.S. exports of electronics, if only companies ranging from producers of color because nontariff barriers have usually been TVs (Magnavox, Quasar) to those designing more significant (nontariff measures and their and manufacturing sophisticated ICs (Ad- impacts are discussed in more detail in a later vanced Micro Devices, Fairchild) and com- section). puter systems (Amdahl). Tariff barriers are seldom the sole cause of Reductions in Japan's tariffs must also be foreign investmentand may be minor factors kept in perspective. The protective barriers that compared with the desire to locate R&D and/or shielded the Japanese computer industry dur- production facilities closer to markets, or to ac- ing its earlier years have been coming down quire state-of-the-art technical knowledge. Still, for some time. In 1978, duties on mainframe tariffs can sometimes be a major consideration. computers were cut from 13.5 to 10.5 percent, In 1978, Nippon Electric Co. (NEC) opened a tariffs on peripherals from 22.5 to 17.5 percent. semiconductor plant in Ireland specifically to The further reductions to which Japan agreed be within the European Community.5 Produc- are no surprise given that Japanese computer tion from this factory is not subject to the 17 manufacturers are now highly competitive in percent EC duty; semiconductors can also their home market. Likewise, accelerated stag- be sold in European Free Trade Association ing for ICs is evidence of the domestic in- (EFTA) nations free of tariffs. NEC, like the dustry's strength; Japan's Government was many American firms th;it had made earlier therefore willing to grant concessions in order European investments, took advantage of what to reduce trade frictions with the United States. is in essence a single market in Western Eu- EC countries did: not feel they had this option. rope. The opportunity to-reduce costs in such a market, combined with the investment incen- Although both the EC and Japan have low- tives provided by the Irish Governmentwhich ered some of their tariffs on consumer electron- was seeking jobssufficed to attract NEC. icsbut not on color TVsthis will have little Ironically, while both U.S. and Japanese firms effect on U.S. exports, which have not been have been able to treat Europe as. one large large. In Japan, prospective importers of col- market, local manufacturers have seldom been or TVs face, in addition to tariffs, a commodi- able to manage this. The rather parochial at- ty tax levied on 17 categories of consumer titudes of both corporations and governments goodsincluding automobiles, home appliances, and camerasthat adds 15 to 20 percent SR. H. Silin, The Japanese Semiconductor Industry: An Over- to the cost of imported as well as domestically view (Hong Kong: Bank of America Asia. Ltd.. January 1979). produced TV receivers. p. 161.

436 434 International Competitiveness in Electronics within the EC have hindered indigenous devel- commodity-like products such as memory cir- opment. The Japanese case is quite different. cuits. To the extent that such patterns continue, There, relatively high tariffs on imports of elec- they will imply that thetariff walls whichpro, tronics were combined with restrictions on for- tected Japanese semiconductor thatiltlaCttiftiB eign direct investment--imposed by the For- for so many years contributed to their eventual eign Investment Law of 1950 as well as strin- competitive success by making it expensive for gent exchange controlsto protect the local these companies to import for their own needs. industry." On the other hand, price competition fueled Secondary effects also arise when imports by imported components has probably bene- subject to tariffs are incorporated into final fited U.S. electronics firms that manufacture products. While intended to shield domestic final products. Sectors like consumer elec- manufacturers, say of components, these tariffs tronics and computers have gained from lower may have the unintended consequence of rais- cost and better quality componentsthe con- ing costs for firms making the final product sequences of heightened competition. Wide- perhaps harming' their competitiveness and spread foreign sourcing of components by eventually leading to demands for further pro- American manufacturers points to the poten- tection. Protection extended to the American tial conflicts of interest with respect to import steel industry, for instance, has increased costs restrictions that often arise between purchasers for U.S. automobile companies. and suppliers. In the electronics industries of some countries tariffs and other trade barriers have Tariff Treatment of Offshore Manufacturing created incentives for internal production and American-made components incorporated in vertical integration. When selecting vendors, imported goods have been exempted from tar- companies weigh prices along with such fac- iffs for almost 200 years, The current version tors as quality and delivery schedules. High of the law is embodied in items 806.30 and product manufacturers to integrate backward, 807.00 of the Tariff Schedules of the United particularly where domestic suppliers have States. Under specified conditions, shipments been protected because they were too weak to from overseas, plants benefit from duty-free compete effectively. Such factors have been at treatment of the value of materials or parts sent work in both the EC and Japan; where many abroad for processing or assembly and then re- firms whose primary end products have been turned to the United States. Without this pro- computerS or communications systems have vision, re-imports after offshore assembly established internal semiconductor operations. would be subject to tariffs on their full value. The tendency has been especially pronounced Because the tariff ex?mptions in items 806 and in Japan, where American semiconductor 807 lower the cost of overseas production rel- products were not as freely available as in ative to the no-exemption case, they Kritplicit- Europe. ly encourage American corporations to split production between domestic and foreign In the longer term, vertical integration plants. U.S. electronics firms began investing where semiconductor facilities produce for inin production facilities in developing countries ternal as well as external salescould lead to as early as the 1950's. While central to cost scale economies that smaller U.S. merchant competition among TV and semiconductor firms may not be able to match. While Ameri- manufacturers, offshore production has been can firms have had the advantage in flexibili- a secondary element in the strategies of U.S. ty compared with their integrated JapaneSe firms making computers and business ma- competitors, and in products where innovative chines. Although labor unions have tended to design has been critical for market success, oppose 806/807 On grounds that they encourage they have not fared so well in mass-produced "exports" of jobs, the evidence concerning the ',United StatesJapan Trade: Issues and Problems (Washing- actual extent to which this occurs remains am- ton. D.C.: General Accounting Office. September 1979). p. 27. biguous (see app. B).

43 / Ch. 11U.S. Trade Policies and Their Effects 435

'I he 806 and 807 provisions differ in scope. tiler processing. Offshore plants owned by Item 806.30 is restricted to metallic articles, American semiconductor firms have been ex- sent abroad for processing, which undergo still tending their operations downstream, shipping more processing after their return to the United completed rather than semifinished ICs back States. Silicon wafers qualify under the typical to the United States. manufacturing sequence outlined in chapter 6. Item 807.00, on the other hand, requires neither Semiconductor devices and TVs are nct the that the articles be metallic, nor that they be only electronic products to enter under 806 and further processed upon their return. However, 807. Modest volumes (in dollar terms) of ca- there are three other conditions, not required pacit-.;rs and vacuum tubes come in under under 806.30: 1) the items. must have been ex- 806.30. Under 807.00 the list is much longer; ported in a state ready for assembly; with no it includes office machinery, communications additional fabrication needed; 2)they.niust not apparatus, watches, stereo and high-fidelity lose their physical identity; and 3) they must equipment, and many types of components. not have been advanced in value or improved. As the size of 806/807 flows indicates, the in condition except through the assembly.proc- tariff exemptions have had significant impact ess.7 By value, the largest category of 807 im- on the global structuring of the American elec- ports consists of automobiles incorporating tronics industry. Companies have rationalized parts originating here. Other major items in- production by shifting manufacturing to parts clude clothing made from fabrics cut in the of the world where costs are lower. In only a United States. Under both 806 and 807, tariffs few cases have the tariff exemptions been are levied at rates.equal to those for equivalent deciding factors, but they have certainly made articles made wholly overseas but are based it easier for U.S. firms to move abroad. As dis- only on the value added abroad.' cussed in chapter 9 and appendix B, the effects Of the two statutes, 807 accounts for the on employment of such transfers are difficult greater value of imports by far, in electronics to evaluate. Depending on the assumptions, as in other product categories. Total value of they can be negative or positive. Even so, in all 807.00 imports in 1980 was $13.8 billion, at least some cases the choice may not be pro- compared with $237 million for 806.30.8 Total duction here versus production there, but pro- value of 806.30 electronics imports in 1980 was duction there or no production at all. only $55 million, continuing a steep decline Fri any event, much of the electronics indus- from more than $250 million 3 years earlier.9 try today is globally integrateda trend to The major electronics imports under both which items 806.30 and 807.00 have contrib- statute items are semiconductors and parts, uted. The consequences span a considerable some of which qualify under either provision. range. U.S. firms have retained competitive- Item 807.00 imports of semiconductor devices ness in product lines where they would other- increased nearly threefold-- during the period wise face marked cost disadvantages. Less-de- 1978-80, reaching $2.45 billionsomething veloped countries have been helped to indus- ever three times the value of color TV ship- trialize, while outward flows of American tech- ments entering under 807.10 Imports under nology have been accelerated. Some domestic 806.30 are being replaced by those under employment opportunities may have been sac- 807.00 because of the 806 requirement for fur- rificed. From a policy perspective, many of the impacts by now appear irrelevant. The laws have been on the books in one form or another 'Imports Under Items 806.30 and 80Z 00 of the Tariff Schedules for decades, and are not likely to be rescinded. of the United States.1977-80 (Washington. D.C.: U.S. Interna- tional Trade Commission Publication 1170. July 1981). p. 4. As tariff levels continue to come down, such °Ibid.. B-2. ex imptions become more marginal to deci- lhid B46. B48. The duty-free values run about two-thirds of the total value. sions on production locations; indeed, wage "Ibid., pp. B-15. B-17. Color TV imports under item 807.00 levels rater than tariff exemptions have nearly can be found in ch. 4, table 14. always been the determining factor.

433 436 International Competitiveness in Electronics

Other Tokyo Round Agreements stating that invitations to bid should allow adequate time for foreign companies to respond. Ten distinct understandingscomprising the Obviously, considerable latitude remains for Multilateral Trade Agreement (MTA)came hindering foreign respondents, but grievance from the Tokyo Round negotiations. Some of machinery is to be established for handling the these, each covering a particular subset of trade complaints of parties alleging discrimination. issues, are irrelevant to electronicse.g., that on dairy products. In other cases, little of sub- The MTA procurement code could have far- stance is changed under the new language; this reaching effects if it functions as written: The is the case for the antidumping and subsi- governments of industrialized nations are dies/countervailing duty provisions discussed major customers for many types of goods; if in later sections. the provisions are fully implemented, these markets would be opened to foreign suppliers. Other MTA provisions pertinent to interna- In actuality, this is not likely to happen very tional tracde in electronics deal with: rapidly. Imagine the repercussions in the government procurement, United States if the General Services Admin- technical barriers to trade, and istration bought 5.000 Toyotas for the Federal import licensing procedures. motor pool. These agreements could yield dividends in the The second,of the listed agreementsthat re- form of increased exports by U.S. electronicS lating to technical barriers to tradetackles, manufacturers, but are not likely to have much or presumes to tackle, the collages of PoliciPs effect on imports of electronics products. used by governments in many countries to re- "Could" because the rather general nature of duce import volumes via diScriminatory tech- the MTA makes infractions difficult to pin- nical standards or regulations. This code is not point. A series of test cases is likely, focusing tightly written, and leaves a number of loop- at first on more blatant departures from the in- holes that could easily be employed to evade n:',--mingful compliance. For instance, govern- tentions of the codes. ments can promulgate regulations or product The first of the three provisions listed above, standards different FT0171 international stand- that covering government procurement, calls ards for national security reasons, to prevent in essence for nondiscriminatory treatment of deceptive, practices, to protect health and safe- foreign firms seeking access to government ty, to preserve the environment, and finally to purchases. That is, foreign and domestic bid- help with "fundamental technological prob- ders are to be treated the same. Exceptions lems." Such rationales liAve been marshaled in related to military sales and national security the past to defend regulations that discriminate will doubtless be interpreted broadly. The stip- against foreign firms and, without much clues- ulations -which cover purchases above about , tiOn, will continue to be so used in the future. $200,000 -are rather far-reaching; they include, This agreement, it is fair to say, is long on for example, state and local as well as national rhetoric but short on substance. governments. On, the other hand, developing With the exceptions noted above, technical countries are not bound by this part of the regulations and standards are to be written so MTA, and virtually none have signed it. as not to discriminate among potential sup- The government procurement agreement pliers or be undue impediments to interna- also addresses matters such as technical spec- tional trade. Where a country's regulations can- ifications and notification of bidders, which not be harmonized with international stand- have considerable impact in practice. Techni- ards, GATT and other interested parties are to cal specifications are, where possible, to be receive full notification of differences. Like- htlegrarl nn infnpnotinnol nearFnprnnninCs cfnnrinrrle 1AI; CO1011 nrotnrn ntorli foci-inn nrnnnt'-4nrCIC Ch. 11U.S. Trade Policies and Their Effects 437 parties to the MTA were at least in principle immense complexity and diversity of nontariff willing to accept the notion of relatively free harriers in various parts of the worldand -access for foreign suppliers. should be regarded as no more than a first step. Whether or not the MTA code on technical It represents an attempt to broaden the com- barriers will have significant effects on common ground among participating nations, mov- mercial practices remains to be seen. In terms ing beyond questions of tariffs and other direct of U.S. exports, the extent tc which standards impediments to trade while holding to the and regulations elsewhere impede shipments premise that has guided negotiations since the has not always been clearleaving aside such original Reciprocal Trade Agreements ACt, well-known examples as the procurementprac- before World War IIthat free and open trade tices of NTT (Nippon Telegraph and Tele- is good for all concerned, with the distribution phone) in Japan. U.S. exporters, in electronics of benefits improved by concessions to less-de- as in other iIclustries, have generally attempted veloped nations. to sell goods abroad that are as close to their The ultimate impact of the Tokyo Round on domestic production as pogsible. In some nontariff barriers, and on future trading pat- cases, exports have been stifled not by foreign terns, remains to be seen. As a statement of in- standards but by'the unwillingness of Ameri- tentions, the agreementsincluding the new can firms to cater to foreign-market conditions. subsidies codeare commendable. From an The third agreement relevant to trade in elec- operational perspective, the verdict is less tronics i!.; that on import, licensing. The text sets clear. Governments seeking politically accept- forth rather general stipulations intended to able reasons for eliminating some of their reg- simplify procedures associated with permits ulatory clutter can begin; countries intent on and licenses, making it more difficult to, use maintaining trade pintection will not find licensing procedures as nontariff barriersand themselves severely constrained. The course especially to single out and discriminate of the world economy will also play a role; against particular countries. Because import governments are loathe to reduce nontariff bar- quotas or Orderly Marketing Agreements fre- riers during periods of stagnation. quently involve licensing requirements, com- In the context of electronics, the Tokyo panies attempting to gain, or hold market share Round agreements have already had some ef- when such quotas are in effect have a special fect. For example, the U.S. Government has interest in equitable treatment. Perhaps the been able to convince Japan to soften its stand most important provision in the import licens- on exempting NTT from the provisions of the ing code states that any enterprise fulfilling the new procurement code. NTT, a major pur- importing country's legal requirements "shall chaser of high-technology communications and be equally eligible to apply and be considered switching equipment, is notstrictly speak- for a license." The only exception relates to ap- ingan agent of the Japanese Government. But plicants in developing countries, who are given its exclusion from the government procure- preference. Governments signing the MTA ment agreement created -a whirlwind of pro- also agree, in awarding licenses, to take into test from spokesmen for tt.S.,industry; who account: 1) economic order quantities or lot believed the.exemption to be symbolic of con- sizes, 2) past import performance of the appli- tinuing efforts by Japan to evade the intent of cant, and 3) "reasonable" distribution, of li- the MTA while subscribing to its language. censes to new importers. After prolonged discussions, the japanese This brief review of MTA provisions points Government persuaded NTT to open its pro-- out the central difficulty now faced by inter- curements to foreign bidders." Thus far, there national_ . trade negotiatorsnontariff barriers. ,,See, for example, T. J: Curran, "PoliticsandHigh Technology: 438 International Competitiveness in Electronics have been few foreign sale.; to the communica- opened by the agreements. Without this aware- tions giant. ness, and without the pressure on foreign governments that such awareness can generate, In electronics and other industries, the even- the agreements will have less effect. Equally tual consequences of the MTA for nontariff important will be attitudes of officials in im- barriers will depend on factors such as aware- porting countries who have responsibilities for ness among exporting farms of thepossibilities monitoring and enforcement.

Dumping The practice of dumpingselling goods in Act contains strong sanctions against pred- export markets at less than their home-market atory dumpingthat intended to eliminate price, or under some circumstances at less than competition and increase market powerits costis one of the unfair trade Practices re- application is narrowly circumscribed. An ac- stricted by GATT. In essence, dumping is a tion filed in consumer electronics under this form of price discrimination; it is proscribed statute remained before the courts for some in export markets for_the same reasons as in years, but more generally the stipulationthat domestic marketsbecause price discrimina- the plaintiff demonstrate predatory intent tion can be used to drive out competitors and makes it unlikely that the Revenue Act of 1916 construct monopolies. In recent years, as will form the basis of future dumping find- American industries have faced stiffer comings.13 This leaves the Trade Agreements Act petition from imports, the number of dumping of 1979 as the primary mechanism for anti- complaints has climbedfrom 11 in 1974 to 44 dumpiroz enforcement. The 1979 Act modified in 1982.12 U.S. law to conform to the revised GATT antidumping code negotiated during the Tokyo In electronIL:s, most of the dumping cases Round." Although the Antidumping Act of have involved consumer products; there have 1921 was repealed and the Tariff Act of 1930 been lengthy proceedings concerning TV re- amended, with a few exceptions the substance ceivers, as well as products like CB radios. An- of the changes was minor. tidumping complaints were among the first attempts by American TV manufacturers to stem According to U.S. law, dumping is the sale the rising tide of imports in the late 1960's and of foreign goods in the United States at less early 1970's. As other portions of the industry tha.n "fair value." The 1979 Act transferred face increasing import competitionnot only responsibility for less than fair value deter- from Japan, but in lower technology products minations to the Department of Commerce; from developing countriesthe number of fil- earlier, the Department of the Treasury had in- ings may continue to grow. In recent years, vestigated dumping complaints and made fair American semiconductor firms have frequent- value determinations. The new act also shortly accused Japanese manufacturers of dump- ened the timetable for investigations, and ing, but have not filed formal charges. changed the definition and determination of fair value somewhat; fair value had formerly The Law and Its Administration been defined as foreign market valuebasically U.S. Administration-of the Antidumping Act of 102i (Was/v.=_--- U.S. antidumping law is now contained in ington, D.C.: General Accounting Office, Mar. 15, 1979). two statutes: the Revenue Act of 1916 and the ""The Agreement on the Implementation of Article IV of the General Agreement on Tariffs and Trade," 'Agreements Trade Agreements Act of 1979. While the 1916 _. _ . Ch.11U.S. Trade Policies and Their Effects 439

the selling price in the country of origin, or, an analysis based on such considerations, after where such information was not available (the which the six commissioners (appointed by the goods might not be sold at home), the selling r)resident and confirrhed by the Senate for price in third countries. Prices formed the basis 9 -yt.ar terms) voteeach making their own of comparison; the law allowed sales at less judgments as to injury or threat of injury. Com- than cost provided the manufacturer also sold missioners, singly or jointly, prepare written below cost elsewhere. if goods were sold only opinions that explain their reasoning. If a main the U.S. market, the old law specified that jority of Commissioners find injury, the remedy a "constructed value" based on estimated pro- is assessment of a special dumping duty in duction costs be determined. In essence, cur- tended to equalize prices between home coun- rent law extends the use of cost-based con- try sales and those in the United States. These structed values to cover fair value determina- antidumping duties are assessed and collected tions where goods are being sold below cost by the Department of Commerce (formerly either at home or in third-country markets." Treasury). Foreign firms th:. t, for whatever reasons, sell below cost at home cannot do so in the United The Color Television Case States without risking dumping convictions, even under circumstances where this would The long and complex history of antidump- not otherwise be judged an unfair competitive ing complaints in consumer electronicsstill tactice.g., when cash flows remain positive not fully resolvedwas no doubt cane of the even though full- costs might not be covered. reasons for provisions in the Trade Agree- Earlier, sales at less than cost constituted ments Act of 1979 transferring responsibility dumping only in narrower circumstances. This for enforcement from the Department of Treas- provision of U.S. trade law, which is not con- ury to Commerce; advocates of stricter admin- sistent With definitions of dumping in most istration of the law felt that Treasury officials other countries, has meant that the Department had been less than diligent, in part because of of Commercenow responsible for antidump- the Department's traditional commitment to ing enforcementloften finds itself estimating open trade. overseas production costs, an exercise f7aught Complaints that Japanese firms were dump- with uncertainties and possible distortions." ing TVs in the United States began in 1968 with Statutory relief isi available only when sales a filing by the Electronic Industries Associa- in the United States at less than fair value are tion (EIA). This initiated what has perhaps found to cause or threaten to cause "material" been 1)- lengthiest case in the history of U.S. injury to a U.S. industry, or to materirlly retard antidumping law." The EIA complaint alleged the establishment of a domestic ind,istry." Re- that the Japanese were able to maintain low sponsibility for establishing injury or threat of prices in the United States for predatory pur- injury rests with the U.S. International Trade poses because prices in Japan were kept ar- Commission (ITC)an independent agency of tificially high by import barriers. The Japanese the Governmentwhich weighs factors such manufacturers acknowledged that retail prices as actual or potential declines in output, sales, were higher in Japan, but held that the dif- market share, profits, and employment. In the ferences were caused by higher taxes and by usual course of events, the ITC staff prepares a complex and costly system of marketing and distribution.

I 'Section -7 -7a, Trade- Agreements -Act of -1979, -Also -sec J. It took 3_years for the Department of Trees- Sklaroff, "United States Antidumping Procedures Under the ury to complete its investigation, findingin Trade Agreements Act of 1979: A crack in the Dam of Nontariff Barriers,"Boston College International and Compamtive Law March, 1971that the Japanese had indeed /440 international Co, /

dumped TVs in th then went to ITC Later that year. IT concluding that t the U.S. industry assessment and dutiesat that tit Treasury. Import required to post a duties. Fixing the size intended to elevt level of TV price lengthy process. tion provided by judged inadequa 'mpetitiveness in Electronics

\ Photo credit: RCA Color TVs undergoing long-term tests te American market. The case ing the duty rate. Eventually, the Department for determination of injury. resorted to constructed value estimates based 'C returned a positive finding, on commodity taxes collected by the Japanese he dumped TVs had injured Government. and clearing the way for the An extraordinary number of claims and collection of antidumping counterclaims accompanied the efforts of me still the responsibility of Treasury and the Customs Service to determine ters of TVs from Japan were and collect these duties. Not only were Amer- 9 percent bond toward these ican manufacturers of TVs and components involved, but also the unions representing their of the antidumping duties employees. Arrayed un the other side were the ate prices of imports ;o the Japanese manufacturers, their U.S. represent- in Japanproved another atives, and the American firas which had been The wholesale price informa- importing TVs s_from Japanmostly large retail- failanese manure-auras-was-- er s-s-u c h-at--See te and in some cases false, tracted course of the disputes also mirrored nrrucarnrrinnt Ch. 11U.S. Trade Policies and Their Effects 441 between the Customs Service and other parts predatory pricing? Would Japanese firms for of the Treasury Department.19 By 1980, only a dozen years or more willfully cut the prices about $13. million in dumping duties had been they charge in the United States below those collected. Moreover, assessment of duties for the market would otherwise set because in 1975 and later :.,,fears has never been completed, some still longer term they seek to monopolize pending final resolution of disputes covering the market? Does selling imported goods at less earlier periods. Not only have duties been at than costnow effectively prohibited by the issue, but also civil penalties for alleged illegal 1979 Trade ActaksZays..,constitute an unfair rebates to importers as a means of circumvent- business practice? Siill, reg-drdless of how these ing the added duties. questions are viewed, the fact is that Japanese With the transfer of antidumping enforce- firms were found under U.S. law to have ment to the Department of Commerce in 1980, dumped TVs. Injury to the domestic industry a new agreement was negotiated with import- was established. American manufacturers of ers. Commerce. agreed to accept a total of about TVs have been entitled to trade protection but $75 million, rather than pursuing in the courts have not received it. The uncertainty and con- duties which the Department estimated at near- fusion created by the these long and convoluted ly $130 million for the period 1971-79. The proceedings has probably done more damage EIAoriginal plaintiff in the dumping pro- to the industry than the dumping itself. ceedingsand its allies then claimed that the The modifications to U.S. antidumping law actual dumping liability was $700 million or incorporated in the Trade Agreements Act of more, and challenged the Commerce Depart- 1979 address some of the procedural problems ment's proposed settlement in the courts; a illustrated by the TV case. Not only has respon- 1981 decision allowing the settlement to stand sibility for dumping determinations and the was appealed to the Supreme Court. Late in assessment of duties been transferred from 1982, the Supreme Court denied the appeal; Treasury to Commerce (by Executive Reorga- evidently Commerce's negotiated settlement nization Plan No. 3, effective Jan. 1, 1980), but can now proceed.20 the ITC injury investigation now begins immediately, rather than awaiting a positive find- This 15-year historywhich has still not ing of dumping. The concurrent investigations come to an end, and during which the com- for which the act sets relatively short time plexion of the American consumer electronics schedulesare intended to speed the process. industry changed irreversiblydramatizes the If future dumping investigations are shorter iequacies of U.S. antiduinping procedures because of the 1979 Act, this will limit uncer- as a means of relief from "unfair" import com- tainties and disruptions, reducing costs for petition. The lessons hold for other industries both defendents and plaintiffs. This would also as wellwitness the example of steel. Not only make it more difficult for domestic firms to use is enforcement slow, complex, and suscepti- dumping proceedings in "strategic" fashion to ble to delay by various parties, but the legal deter foreign competitors from entering U.S. definitions of dumpingwhich, in the United markets; dumping complaints can discourage States as in many other countries, predate market entry through the threat of future GAVI.seem remote from the realities of busi- penalties as well as by imposing legal costs on ness competition. No one argues that preda- defendants. tory practices should not be outlawed, but what relevance, for example, does the relationship of home market price to export price have to Prospects for Dumping Actions Elsewhere in Electronics Wighunan, "Charges U.S. Blocks $400M Duty on Japan 442 International Competitiveness in Electronics but rare in senile nductors or computers. The single case involving semiconductors, in 1972, led to a finding of dumping but not injury. Im- ports of c,mputers into the United States have been at luw levels, leaving little reason to expect complaints. If computer imports were to rise and dumping to be alleged, less than fair value pricing would be difficult to establish; at least for large systems. The complexity of such systems, the difficulty of establishing com- parability in performance, and the high R&D expenses that must be borne, complicate pricing comparisons. Moreover, selling prices for data processing systems often include service or software charges that arehard to isolate. Pricing structures in the computer market particularly the establishMent of "quality- adjusted" pricinghave already created formidable difficulties in purely domestic antitrust P17010credit: Perkin-Elmer actions where predatory pricing has been at Plasma etching system used in semiconductor fabrication issue.21 Less than fair value' determinations based on foreign market prices or constructed values would be still more troublesome, at least for mainframes. The problems are not so intractable for small systems and peripherals, with resulting cost advantages from the learn- where significant import penetration is in any ing curvemight well occur before dumping event more likely, while personal computers pro,:eedings could be resolved, even under the sold at retail could be treated much like other accelerated timetable of the 1979 Act; consumer products. moreover, the same factors lead toadvance pricingwhich is not in general illegal. There- The characteristics of the semiconductor in- fore, while antidumping actions may continue dustry also work against antidumping proceed- to be filed in more mature sectors of theelec- ings.Large-scale ICs--including computer tronics industrye.g., consumer ,Iroducts, memory chips, where import saleshave in-, where the technology is relatively stable and creased rapidlyexperience relatively short price competition based on low production product lifetimes. Coupled with the large econ- costs intensedumping allegations /inhigh- omies of scale in IC production, and the im- technology sectors seem less likely to escalate pOrtance of yields, deep market penetration- from verbal attacks on imports to formal complaints. In high- technoicgy industries, products 21See, for example, R. Michads..'lledonic I'ricris and the Strut can be obsolete by the timedumping actions lure Of the Digital Computer Industry." Journal of Inthistriai Eco- nomics.vol. 28,March 1979, p.263. have been resolved./

Subsidies and Countervailing Duties

-innothr, criortrumFrnm privortising in the form of credits or guarantees extended Ch.11U.S. Trade Policies and Their Effects 443

agreements limit interest ratesto levels that R&D support) and what are clearly export sub- can be below the market rate but not too far sidies (e.g., low-interest loans to foreignpur- below.22 In such forms, export subsidies have chasers) will always be ambiguous. As nations become one of the most common nontariff pursue increasingly sophisticated industrial measures affecting international trade; sub- policies, it becomes still more difficult to draw sidies with domestic objectives also havecon- that line. sequences for exports. Although important in capital goods in- Countervailing Duty Law and dustries, neither export financing nor export Its Administration promotion have played major roles in electronics outside of telecommunications.23 In GATT and U.S. law provide remedies paral- contrast, subsidies with ostensibly domestic ob- leling those for dumping where American jectives have become a major tool by which firms and industries are injured by export sub- governments promote their electrolVr:,s indus- sidies. In dumping cases, private firms set the tries; these have less direct and visible effects prices at issue, while prices are distorted by than export credits, making them difficult to direct or indirect government action in thecase countervail or to negotiate over. While revi- of subsidies.- Importing nations then impose sions to the GAIT subsidies code were a cen- countervailing duties for essentially thesame tral item on the agenda for the Tokyo Round purpose: to eliminate price differentials created negotiations, little progress was made; the by the unfair trade practice. In principle, the changes were basically matters of procedure. countervailing duty is set at a level that bal- Distinguishing export from domestic or inter-- ances the effect of the subsidy. In practice, the nal subsidiesthe latter of many forms but administration of countervailing duties in the universally employedis central to a workable United States is even more problematic than code but fraught with practical difficulties. for antidumping dutil.!s. Measures adopted by governments that have U.S. countervailing duty legislation is found the effect of subsidizing domestic electronics in two statutesthe Tariff Act of 1930 and the industries range from grants for basic research Trade Agreements Act of 1979. As in thecase to regional development incentives. Because of antidumping law, responsibilities in counter- any such policy, even relocation assistance for vailing duty cases are splitthe Department of displaced workers, could in principle help Commerce investiaates foreign export subsi- firms to exportby cutting costs, raising prof- dies (this responsibility was again lodged with its, or improving technical capabilitythe. Treasury until 1980); ITC determines injury. dividing line between measures that most peo- If ITC votes any of three findingsinjury toa ple would agree are domestic subsidies (e.g., U.S. industry, threat of injury, or impediments to the establishment of a new U.S. industry "When Canada proposed a financing package for New York City subway cars at an interest rate of 9.7 percentwell below then a countervailing duty equal to the net the agreed international minimum of 11.4 percentthe Cana- value of the subsidy is to be imposed on the dian Government defended this as necessary to meet France's imports. offer. See "Reagan Decides U.S. Should Not Match Financing fin NOW York Subway Car Sale," U.S. Import Weekly, July 14, Under the 1979 Act, the test turns on "mate- 1982, p. 448... nror a discussion of export financing, see R. E. Shields and rial" injuryincluding actual or potential R. C. Sonksen, Government Financial Institutions in Support. declines in output, sales, market share, cash of 1 -9. Exports (Washington, D.C.: Center for Strategic and In- flow, profits, productivity, capacity utilization, ternational Studies, Georgetown University, September 1982). A nifire general review of U.S. export policies is Report of the employment, wage levels, or the ability to raise President on Export Promotion Functions and Potential Export capital." It: earlier years, U.S. law did not re- Disincentives (Washington, D.C.: Department of Commerce, quire that injury be found before countervail- 444 International Competitiveness in Electronics subsidies was enough. Although differing from levy such taxes can exempt or rebate them as GATT language, this provision had preceded they wish. The United States has less latitude the establishment of GA,TT and been retained than nations with extensive arrays of indirect under a grandfather clause. The original law taxes. had been passed before the turn of the century, After the Trade Act of 1974 had been passed, but no countervailing duty was imposed by the Zenith challenged rebates of Japan's commodi- United States until 1967. ty tax on exported TVs under the act's provisions. The American manufacturer sued the The Question of Indirect Taxes Department of Treasury, claiming that rebated indirect taxes in Japan constituted subsidies What then is a "subsidy" under GATT and/or and that Treasury had failed to properly inter- U.S. law? As might be expected, the definitions pret the new law.26 Treasurycountersued, have been controversial. There has been a ma- claiming that decades of acceptance by all par- jor legal action in consumer electronics, the ties of its past practices had effectively ratified case hinging on whether exemptions orrebates these practices. Four years later, in 1978, the of indirect tax( :s on exported goods constitute Supreme Court upheld Treasury'sposition)rui- a subsidy. bath the old and newGATT sub- ing that rebated commodity taxes do not con- sidies codes permit indirect taxese.g., val- stitute subsidies under U.S. law. ue-added taxesto be rebated, but not direct taxes. (Direct taxessuch as corporate income Countries with commodity or value-added taxesare levies based on factors of produc- taxes generally levy them on imports as well tion like capital or labor, indir,..1ct taxes are as domestic production.Thu, within a coun- those levied on the product itsail.) The assump- try having indirect taxes the impacts are, at tion underlying this rule is that indirect taxes least in principle, neutral: both imports and can be easily included in pricesand passed domestic goods are subject to a tax based on along to consumers, while direct taxes cannot their value. However, matters are not really this (they depend, for instance, on annualized profit simple. Exports from a nation like the United levels). If the full indirect tax is passed through States that relies on direct taxation may be to the purchaser, profits to the seller are unaf- burdened with higher selling prices reflecting fected. Under these circumstances, a rebate or higher corporate taxes, thus at a disadvantage exemption of such taxes on export sales would in maf.kets where indirect taxes are the rule not constitute a subsidy under theusual defi- (countries with substantial revenues from in- nitions. direct levies normally tax personal and corporate income at correspondingly lower rates). Compared with its trading partners, the Furthermore, foreign manufacturers shipping United States relies less heavily on indirect to the United States may reap benefits: after leviessales, excise, and value-added taxes receiving rebates on indirect taxes at home, and more heavily on direct taxation, primari- such firms face no compensating border tax ly of cort,oraie and personal income. Many adjustments when their goods eater the United European nations impose a value-added tax Statesthough they generally must pay tariffs. (VAT) at each stage of the production process. They are free to sell in 'a market where the In Japan, consumption taxes of 5 to 30 percent prices charged by domestic firms may well apply to items such as automobiles, electrical have to cover higher corporate taxes. As a appliances (including TVs), and a variety of result, nations that rely heavily on indirect luxury goods, while excise taxes apply to other taxes can be presumed to have advantages in dasses,25 Under GATT rules, countries that international tradealthough the size and

2 5 Flint) ri.Stimulation Prrigrams in the Maior Industrial Coun- Ch. 11U.S.Trade Policies and Their Effects 445 significance of the advantages can be difficult electronics, fv::vvever, the proportion is much to judge. lowerperhaps in the range of 10 percent." VAT systems have sometimes been suggested Other countries have registered complaints for the United States, in part --cause of their with GATT against the DISC mechanism,argu- potential for stimulating exports (assuming cor- ing that it functions as an export subsidy but porate taxes were reduced at the same time). does not qualify as an exemption from indirect The effects of such a shift in U.S. tax policy taxes." Despite a finding by GATT that DISCs on specific firms and industries would depend do constitute subsidies, no country has yet im- on factors such as:27 posed countervailing duties on U.S. exports, compensating reductions in income taxes, nor has the United States offered to repeal the as well asthe overalltax liabilities (and legislation that permits DISCs. (Recently, the profitabilities) of the firms in question; Reagan administration has proposed an alter- the extent of vertical integration charac- nat've to DISCs, as pointed out in the next teristic of the industry, along with the chapter.) place of particular firms in the chain of production; Other Unfair Practices fractions of revenues stemming from ex- Section 337 of the Tariff Act of 1930which ports;, was amended in 1979 but has seldom been used price elasticity of demand for each product deals with unfair competition in interna- affected; and tional trade not already covered by antidump- design and implementation of the system ing or countervailing duty laws. Most of the for collecting the VAT or other indirect tax complaints filed under section 337 havecon- and (optionally) rebating it for exp'irted cerned patent infringements, as in Apple's goods. complaint to ITC over counterfeit computers, While the merits of VATs have thus far not but in yet another case concerning imported been seriously debated in this country, since TV receivers, two American manufacturersac- 1971, U.S. law has provided a mechanism cused Japalese firms of illegal predatory pric- the Domestic International Sales Corp. (DISC) ing practicesspecifically, of cutting prices in intended to put American exporters on a the United States below costs inan effort to more even footing with manufacturers in coun- drive American firms from the market.3' When tries having indirect taxes. DISCssubsidiary imports are involved, price-cutting complaints corporations whose activities are confined to are usually filed under antidumping or counter- selling goods in export marketspermit U.S. vailing duty statutes, but section 337 actions firms to defer a portion of tax liabilities from can also be brought if conspiracy or intent to profits on overseas sales. Several thousand monopolize is alleged. In this instance, ITC DISCs have been established, primarily by proceeded with a section 337 investigation larger American corporations with substantial even though the Department of Treasury volumes of export business. In recent years, 2,This estimate is based on a survoy of 325 member firms by more than half of all U.S. exports have been the American Electronics Associatior,. Because most of themem- bers of the Association are smaller companies, it probablyun- channeled through DISCs.28 For exports of derstates the actual fraction for electronics. See Capital Forma- tion, Part 1, hearing, Senate Select. Committee on Small Business. Feb. 8 and 10, 1978, p. 53. "See A Value-Added Tax for the U.S.? Selected Viewpoints 3°See J. M. McGuire, "The GATT Panel Report on Domestic (New York: The Tax Foundation, Inc.. 1979). International Sales Corporations: Illegal Subsidy Under GATT " "Export Stimulation Programs in tlw Major Industrial COUll- Interrational Trade Law Journal, v11. 3, 1978, p. 387. tries, op. cit., p. 319. "ITC Investigation 337TA23. filed jam 15. 1976. 446 international Competitiveness in Electronics claimed exclusive jurisdiction under counter- case was terminated whenITC issued consent vailing duty law. The theory behind the com- orders prohibiting predatory pricing and spe- plaint was that assistance given Japanese TV cial purchase inducements .'color TVs. Fu- manufacturers by their governmentthough ture section 337 complaints by American elec- not necessarily bounties or grantswithin the tronics firms are perhaps most likely as at- definitions of countervailing duty statutes- tempts to expedite relief, given the slow pace might still constitute a conspiracy to restrict of past antidumping and countervailing duty trade, an unfair practice under section 337. The investigations:

Quantitative Restrictions and theEscape Clause

Over the 9ast two decades, tariff levels have the result of unilateral action by Japan rather been reducud by international agreement to the than negotiations between two governments point that, for many goods and in many ad- exporting nations have entered into OMAs of vanced economies, they are no longer a major their own volition. factor in market outcomes. Nowhere is this The United States negotiated its first OMA movement plainer than in electronics.With covering imports of color TV receivers in 1977 tariffs largely closed off as a legitimate vehi- with Japan. Under the conditions, Japan agreed cle for protection, industries exposed to the to limit shipments of color TVs tothis coun- rigors of international competitiontogether try for a 3-year period; no more than1,560,000 with their employees and political supporters complete sets and 190,000 incomplete sets have sought other forms of relief. Along with were allowed each year. Exceptfor being the many other nations, the UnitedStates has in- outcome of bilateral negotiations, thecolor TV creasingly fallen back on import quotas. By OMA was equivalent to a quota of the type out- whatever nameOrderly Marketing Agree- lawed under GATT. ment, Voluntary Restraint Agreementquotas limit shipments originating in particular'coun- The stop-gap nature of this first OMAcov- tries. Under GATT, unilaterally imposed quo- ering a single troublesome exporterwasil- tas are explicitly disallowed except to correct luminated when Taiwan and South Korea took persistent balance of payments deficits, and up the slack (ch. 4). Itquickly became necessary then are to be Jmpo ra r y . Nonetheless, quotas to extend quotas to these twocountries if the have proliferatedtypically on a negotiated U.S. industry was to be effectivelyshielded. bilateral basisz--with the path often :;leared by OMAs were negotiated with Taiwan and South "escape clause" actions permitted under Korea late in 1978, to expire at the same time GATT. An outline of the escape clause mech- as the Japanese quotaJune 30,1980. Imports anism in U.S. lawsection 201 of the Trade from Taiwan were limited to roughly half a mil- Act of 1974follows the discussionbelow of lion units, plus twice as many incomplete sets quotas on color TVs. (without picture tubes), over the year-long period beginning Jul- 1, 1979. Koreanship- Orderly Marketing Agreements for ments were restricted to about 300,000TVs.32 This extension to other countries illustrates a Color Television Imports common failing: wheninitially directed against The only direct quotas on U.S. electronics a single exporter, quotasmust often be wid- imeec-tg have heen termed Orderly Marketing ened as now competitors step inthe seriesof Ch. 11U.S. Trade Policies and Their Effects 447 tifiber Agreement being the classic case. Note predictable, were decision; by Korean and that table 14 in chapter 4 shows imports from Taiwanese firms to follow the Japanese had in Mexico more than doubling over the period establishing assembly operations in the United 1976-80during which time the OMAs with Ja- States. pan, Korea, and Taiwan took effectwhile shipments from Singapore increased more Escape Clause Proceedings in than seven times. Virtually all the imports from Color Television Mexico enter under item 807.00 of the Tariff Schedulesmeaning they are shipments by Why have the United States and other na- ArrLA `can -owned firmswhile both 807 and tions resorted to quotas? Partly because quan- non -807 imports from Singapore have gone up titative restrictions are administratively clean sharply; Singapore now ships more TVs to the simple to monitor. More important, for a har- United States than Korea did at the time the ried government, quotas may seem the best ON1A with that nation was negotiated (table choice among a set of generally unattractive 14). Might there be pressure for quotas with alternatives. The color TV case illustrates the Singapore at some future time? Or other Asian. political dilemmas that often foster such deci- countries? If so, could unrestricted Mexican sions. shipments be justified simply because they are intracorporate transfers of U.S.-based multi- The OMA with Japan followed a series of nationals? legal actions initiated by the U.S. industry in attempts to stem rapid increases in imports. As As expected by the American negotiators, discussed earlier, dumping charges against the Japanese manufacturers responded strategical- Japanese came first, but for a variety of reasons ly to the OMA. To avoid the new restrictions, duty collection was repeatedly postponed. they not only invested in Taiwanese and South American firms together with labor unions rep- Korean manufacturing facilities but opened resenting their employees continued to press assembly plants in the United Statesa desir- for import relief via other avenuesone being able consequence from the viewpoint of the Zenith's countervailing duty suit, metdioned Federal Government because these plants earlier and destined ultimately to fail. The would help maintain domestic employment, avenue that finally proved successful began d4fusing some of the pressure from labor with an appeal filed in October 1976 by a group unions. As these U.S. plants came onstream, of 'companies and unions for relief under the Japanese shipments of color TVs (but not of escape clause, section 201(b) of the Trade Act subassemblies) diminished. By 1980, Japan's of 1974. This provision, following article XIX exports of completed and nearly completed sets of GATT, permits trade restrictionsindepend- were no longer considered a threat, and the ent of questions concerning fairnessif im- OMA with Japan was allowed to expire on ports are found to be causing serious injury or schedule.. Of course, the possibility of a new threat of injury to domestic producers. The pur; quota continues to shape business decisions by pose is to allow a temporary respite or escape Japanese exporters. from import competition while industries ad- OMAs with Taiwan and Korea, on the other just to new conditions. The protective meas- hand, were renegotiated to cover the period ures adopted in such cases, termed safeguards, through June 30, 1982 at new levels (Taiwan: need not be quotashigher tariffs are one 31- 400,000 sets in the first year, 425,000 in the sec- ternative. ond; Korea: 385,000 sets in the first year, In terms of the color TV OMAs, two features 575,000 in the second), .after which they too of the escape clause_ mechanism are note- _ iinwnri In nnri offine rnncram csrra ana.n Arr,nitisl,r. ;..e+ ne,ird.v eicsr.s_ 448 International Competitiveness in Electronics liberalization. Without this change, protection try did not in fact stand together on this), re- for the American industry via the escape clause ceded. But the OMAs also accelerated a proc- would almost certainly`have been precluded. ess begun earlierthe establishment of U.S. operations by Japanese TV manufacturers, and Another feature new to the 1974 Actcon- later Taiwanese and South Korean firms. Sony cerning the role of ITC in the investigation of had initiated the trend in 1972; since then, injurybore on the ultimate outcome of the many others have followedas described in color TV case. Under earlier law, when injury chapter 4sometimes by taking over the plants was found ITC recommended remedies to the of ailing American rivals. Wholly owned Jap President, who could either accept or reject anese subsidiaries now supply perhaps one- them. The 1974 Act added a time limit, stipu- third of the U.S. market (table 10). If American lating that the President respond within 60 manufacturers expected to recapture the do- days to an ITC injury finding. Furtherand mestic market, or if they anticipated a slacken- most significantthe act provided that what- ing in price competition, they were disap- ever action the President took could be over- pointed. ridden by a simple majority of Congress.* Thus, the options available to the executive The hill range of consequences provides branch had been narrowed, the hand of those other causes for reflection. OMAs did not stop advocating import relief strengthened. The the transfer of U.S.-owned production facilities threat of reversal by Congress greatly increases to foreign countries, a movement,that had be- pressures on the Executive, for whom the col- gun earlier. Zenith, for instance,continued to or TV case posed a dilemma. The ITC Commis- shift TV manufacture to offshore plants in sioners determined that the U.S. industry had Mexico and Taiwan. Still, if the industry does suffered injury, andwith only one 'dissenting not appear to have gained materially from the voterecommended a large tariff increase. If quotas, it is likely that further losses were the President took this course, an international avoided. trade dispute of major proportions would That competition did not abate is shown by almost certainly havl been precipitated. On the price data collected by ITC over the period of other hand, rejecting the ITC recommendation the initial agreement with Japan; retail prices would bring with equal certainty the prospect for color TVs (19 inch and smaller) remained of reversal by Congresseven more embarras- essentially constant during a period of severe sing. Under these circumstances, the White inflation in the U.S. economy." Even for large- House finessed the entire problem by negotiat- screen sets, where U.S.-owned firms continued ing with Japan for voluntary restrictions. Dis- to dominate the market, prices increased only cussions carried out between the (then) Office abobt 6 percent. While price stability also mir- of the Special Trade Representative and`the rors cost-cutting improvements in both product Japanese Ministry of International Trade and and process technologies, it seems :clear that Industry (MITI) led to the OMA. competitive responses by Far Eastern Manufac- What have been the consequences for the turers were the chief cause. During the same U.S. TV industry? That the political victory had period, many household appliances rose in any very substantial impact on its competitive price by 50 percent and more. vitality can be questioned. Imports were cut Nor did profits recover. While OMAs re- back, and the frontal assault by Asian firms ar- duced import market sharesin the 18- and rested. The specter of U.S. manufacturers be- 19-inch categories, penetration declined from ing totally overridden, which underlay the ap- about 30 to 10 percent during the first year peals by industry and labor (thoughihe indus, Ch. 11U.S. Trade Policies and Their Effects 449

in terms of the competitive position of the U.S. to be that restrictions may work better in pro- industry, this apparent benefit was partly off- tecting what are essentially infant industries, set by the output of Japanese firms assembling at least if combined with other policies support- TVs here. Capacity utilization rates of domestic ing industrial development. in the United firms improved but profitability did not follow. States, on the other hand, quotas intended to The ratio of net operating profits before taxes protect mature industriesnot only color TV, to net sales, which had declined from E.7 per- but automobiles or steelhave had ambiguous cent in 1:i;72 to a loss in 1974 of 1.2 percent, cuZcomes. has been rn:1M:i.g at less than 2 percent inre- cent years, as pointed in chapter 4. While Could quantitative restrictions effectively a'tfAs helped preserve domestic employment shield other portions of the U.S. electronics in- opportunities, they provided no more than dustry should imports surge as they have,say, modest relief from competitive pressures. in semiconductor RAMs (random accessmem- ory circuits]? Probably not. Early in 1982, amidst consternation created by heavy import Effects or Quotas and Other penetration figures for 64K RAMs, Hitachi, Fu- Nontariff Restrictions jitsu, and NEC all announced accelerated time- Many nations have utilized restrictions other tables for assembly in the United States. These than tariffs to regulate trade in electronics. moves were clearly aimed at heading off for- Japana major beneficiary over the past three mal complaints. If dumping or escape clause decades of vigorous advocacy by the United proceedings had been instituted, the parallels States of open international tradehas em- with color TV would probably have beenrep- Tiloyed nontariff restrictions frequently and ef- licated still further. As for color TVs, Japanese fectively as part of its economic development firms already have enough volume in the U.S. strategy. Among the more blatant nontariff bar- IC market to attain the scale economies needed riers created by the Japanese has been MITI's for standardized products. In general, quotas definition of domestically produced comput- are not a promising route to improved competi- ers. These are confined to systems manufac- tiveness for high-technology American in- tured by firms in which majority ownership is dustries like electronics. Japanese.35 Machines built within Japan by American-owned firms are "foreign"despite The Escape Clause the fact that IBM-Japan, for instance, employs As mentioned above, GATT permits govern- some 13,000 Japanese and only a handful of ments to come to the aid of domestic industries Americans. MITI has preferred that purchas- threatened by imports. But before protection ers of computers chose "domestic" equipment, can be extended under the escape clause pro- using controls over foreign exchange to help vision in section 201 of the 1974 Trade Act, ITC enforce its wishes; although exchange controls were dismantled in 1975, MITI continues to must return a finding that "an article is being monitor the market, and reportedly advises imported into the United States in such increased quantities as to be a substantiaicause customers to buy Japanese computers.36 of serious injury, or the threat thereof, to the That nontariff restrictions appear to have domestic industry producing an article likeor been more effective in achieving their osten- directly competitive with the imported arti- sible goals in Japan than in most nations il- cle."37 Fairness or unfairness is not part of the lustrates once again that evaluating industrial_text. The rationale is to pi ovide a time-inter- policy-meaStiriiS-kSeldoinstraightfOrward. vat during which the threatened, industryand One lesson of the Japanese experience appears its workers can adjust to the (new) competitive 450 International Competitiveness in Electronics

Revisions to U.S. law in the 1974 Act made' fad tapered off. Nonetheless, imports captured itconsiderably easier for an industry to the vast majority of 1979 sales.4° demonstrate injury and thus qualify for protec- .Despite questionable effectiveness in past tion. As noted above, relaxation of the provi- cases in electronics, the escapeclause remains sionAhat relief be contingent on a rise in im- a tempting vehicle for portionsof the U.S. inports stemming from tariff concessions or dustry that find themselves hardssed by ship- other forms of trade liberalization by the ments from overseas. First and foremost,it United States was instrumental in the color TV does not require that imports be linked to un- action, Furthermore, previous incarnations of fair behaviora condition that has often the escape clause. required that increased improved difficult to satisfy.' in dumping or ports be a major cause of injury. The 1974 ver- countervailing duty actions. Furthermore, in- sion changed the adjective to substantial,de- jury can be defined in terms of narrow product fined as "important and not less than any other categories. The law requires only that injury cause." This standard is considerably weaker, be demonstrated in "that portion or subdivi- and all vise equal makes it easier for belea- sion of the producer which produces the like guered industries to secure protection.38 or directly competitivearticle;" the market in Other than the color TV case, only one suc- which such injury cccurs can be limited to "a cessful escape clause action involvingelec- major geographic area of the United States."'" tronics products has been advanced sincethe The implications can be appreciated by recall- passage of the 1974 Trade Act.This was filed ing the typical competitive strategies of Japa- An late 1977 after a fourfold increase in imports nese exporters. In both consumerelectronics of CB radius. ITC worded its findings strong- and 5e`miconductors, exporters selected specialized market niches where American man- ly: "...serious injury is clearly immineni and threatens the domestic industry with extinction ufacturers seemed vulnerable, the intent being unless remedial action is taken toenable U.S. to gain a substantial market sharewithin this producers to compete on more equal price niche and then diversify. Thus Japanese semi- terms."39 The President responded by raising conductor manufacturers concentrated on 16K impOrt duties .frOm 6 to 21 percent. Afterthe RAM chips, taking advantage of a shortfall in first year, the duties decreased inincrements, U.S. production capacity to quickly gain some reverting to their original level at the end of 40 percent of the Ameriban market. Underthe kThe impact of this period of provisions of the Trade Act of 1974, an export the trc trade tariprotection on the CB radio industry is dif- strategy of this type could be subject to fiult to judge, largely because sales dropped restraints. recipitouslyfrom around 5 million units in 1978 to only 2 million the next yearasthe CB 40Electronic Market Data Book1980(Washington, D.C.: Elec- .1.v. R. Cline. N. Kawanahe. T. 0. M.Kronsio.and T. Williams, tronic Industries Association, 1980). p. 49. "19 U.S.C. sec. 2251(b). Dumping and countervailingduty Trade Negotiations in the 7'okyo Round: A Quantitative Assess- hasis'as well: ment (Washington. D.C.:. Brookings Institution,1978), p. 203. statutes invite cdrnplaints on a narrow product line "'U.S. Import WeeklyReference File (Washington.D.C.: during 198 ?, more than 120 separate investigations in carbon Bureau of National Affairs. 1979), p. 58:0106. steel products alone were undertaken by ITC.

Prospective Effects,olf LLF-Trade Policy on the Elecf(onics Industry

- To what extent, then, might the panopoly of the course of international competition, but Ch. 11U.S. Trade Policies and Their Effects 451 result of transferring responsibility foran- out sharply, whereas the benefits are visible tidumping and countervailing duty provisions mostly in the aggregate. Is it possible that the from Treasury to the Department of Commerce U.S. economy is too open, given changes in the has be& a more sympathetic hearing for Amer- international marketplace? Does the absence ican busineisand as a consequence, the fil- of effective import controls in electronics make ing of more complaints. Much also depends on the domestic industry overly vulnerable to in- the complexion of ITC, which shifts as Presi- roads by foreign industries? dentially appointed commissioners come and go. Changes in the definition of injury have mad'rotection at least in principle easier to obtain; these too, in the ordinarycourse of Conbumer Electronics events, serve to encourage demands for trade restrictions. Thus far in electronics, trade ac- International competition has generated con- tions have centered on consumer products; as tinuing pressures for trade protection in con- Japanese manufacturers ,step up their price sumer electronics, yet the current situation of competition in semicondtictors and computer the U.S. industry is partly a consequence of equipment, there may be filings in these prod- domestic competition. Today, two American- uct categories. Furthermore, complaints by, the owned firms, Zenith and RCA, account for U.S. industry are increasingly centered on sub- roughly 40'percent of U.S. color TV salesas sidies and other tools of national industrial they have, rather consistently, for many years. policye.g., Japan's R&D programs. By and Although import penetration increased dra- large, trade negotiations and the GATT have matically during the 1970-77 period, the brunt proved unable to deal with such issues. of the sales losses were borne by other, manufacturers. In a single year: Magnavox and Certainly protectionist sentiment has been Motorola each saw their domestic TV sales rising over the past half-dozen years. Is a turn- drop by more than 15 percent. These Market ing point in the American attitude toward trade declines led rather directly to the sale of their a real possibility? For some 50 years, the United TV operationsto North American Philips and States has taken the lead in international ne- Matsushita, respectively. While only four do- gotiations to lower barriers to trade. From the mestically owned producers remained in 1.983 first Reciprocal Trade Agreements Act in the (compared with 17 in 1970), they have been 1930's to the Tokyo Round MTN, U.S. policies joined by. more than 10 foreign companies have supported liberalization as being in-the manufacturing or assembling sets here. Con- Nation's long-term interests. Both major polit- centration has not increased significantly and ical parties have for the most part accepted the no one firmAmerican or Japanesehas come underlying preMise of these policies: open close to dominating the market. Policy deci- trade leads. to an efficient allocation of global sions by the U.S. Government stimulated the resources, as a result of which the American influx of foreign capital, although OMAs probe people will, more often than not, find them- ably influenced timing more than decisions to selves better off. Other nations have benefited invest. Foreign-owned plants in the United as well. The export-based economic growth of Statestogether with continuing imports from West Germany and Japan owes much to the Japan and other Far Eastern nationscreated openness of the large and affluent U.S. market, relentless pressures on American TV manufac- while American consumers have gainedaccess turers, even while quotas were in force. U.S. to a greater variety of products, as well as lower firms shifted production abroad to reduce prices resulting from foreign competition. On costs, at the expense of jobs herebut con- the other side of the ledger are the costs of dis- sumers have benefited via low prices and high- location and adjustment that follow upsurges quality products. Still, only the largest and ofimnnrte Moe,. mete tont] to rinrIct 1-Itanly;11, 452 International Competitiveness in Electronics

For other consumer goods, Federal policies isting U.S. trade policies could shield domestic have rarely come into playarid seem unlike-, firms producing the consumer electronic prod- ly to, at least in a manner similar to events in ucts of the foreseeable future. Indeed, cost the TV industry. Far Eastern firms are today pressures -will probably- continue_to drive a the unchallenged leaders in most other con- good deal of production by American entrants sumer electronic products; thereis virtually no offshore. U.S. manufacturing left to protect when it comes to portable radios, monochrorric TVs, stereo/high fidelity equipment, the simpler Semiconductors pocket calculators, or electronic watches. More Unlike consumer electronics, where the ef- important for the future, the United States has forts of U.S. manufacturers have been largely lagged in the development of new products confined to the domestic market, production such as video cassette recorders (VCRs). At and sale of semiconductors is carried out on every step, Japanese manufacturerslead in a global basis by the major U.S.merchant product development or are working in parallel firms, as well as by Japanese andto a far or in cooperation with Americanfirms, the lesser extentEuropean producers. U.S. lead- primary exceptions being electronic toys and ership has meant that domestic manufacturers games and home computers. If products like have not, as yet, sought direct Government as- -video disk players achieve mass-market suc- sistance in combating imports. For many years, cess, the Japanese will be early andformidable sales expanded rapidly; American suppliers competitors. were often hard pressed just to keep pace.The Under such circumstances, protection for number of domestic companies serving the American manufacturers would have to come market tripled during the 1960's, while imports through legal provisions that have not in the were until recently almost entirelythe- ,inter- past been exercised. When applying the injury divisional, shipments of U.S. multinationals standard in antidumping proceedings, ITC ex- (ch. 4). amines whether imports are harming a domes- The picture began to change at the end of the tic industry, are likely to damage it, or are 1970's. Aggressive competitive tactics and mar- preventing such an industry from being estab- ket successes by Japanese firms have had im- lished. The last of the three possibilities has pacts in many parts of the world, but from the seldom been relevant because existing indus- perspective of U.S. policymakers the domestic tries have normally sought dumping investiga- market has been the focus. Japanese companies tions. But in principle, the clause could be a are providing the first real competitionfrom basis for reliefif imports were priced at less abroad in the experience of most of the Amer- than fair valuefor products that are not even ican industrycompetition that has driven being made in the United States, such as VCRs. them to seek the attention of the Federal Gov- On the other hand, the escape clause injury ernment. The most publicized examples.of Jap- standard would have to be considerably anese inroads have been the 16K and 64K stretched. Here, there are only two possibilities: RAMs sold in large numbers to manufacturers an existing industry must beseriously injured of computers and microprocessor-based sys- or threatened. Given productleadership over tems. In. 1980, Japanese manufacturers cap- seas, with imports achieving a sizablemarket tured about 40 percent of the U.S. andworld share from the outset, serious injury to an ex- market for 16K chips, partly because of inade- isting industry probably could not be demon- quate capacity in American plants. By 1982, strated (assuming new types of products) un- the Japanese share of next-generation 64K less the standard was applied in a novel and RAM sales was running at about 70 percent. unintended way. Given the general ineffective- More than any other event, the rapid inroads Ch. 11U.S. Trade Policies and Their Effects 453

The tempest over 641 RAMs in the spring vision by industries suffering from foreign of 1982 may prefigure future trade disputes in competition.42 Nor were antidumping proceed- the high-technology products of this and other ings initiated for 64K RAMs. Later in the year, industries. As publicity over inroads growing-demand caused-prices to firm,-defus- by Japanese imports, the Departments of Com- ing allegations of dumping and redirecting lob- merce and Defense began examining the im- bying by domestic semiconductor firms toward plications for national security. At the time, reciprocity legislation. Nonetheless, following only Texas Instruments and Motorolaamong this turn of events, Japan got another unpleas- U.S. merchant firms were producing 64K chips ant surprise: the Justice Department began an in quantity. Prices had been dropping rapid- investigation of six Japanese semiconductor ly, driven not only by declining manufactur- manufacturers, premised not on price-cutting ing costs, but by recessionary pressures leading but on restricting shipments to the United to price cutting in the Japanese market as well States in order to raise prices.43 To con- as here: Worldwide production capacity for siderable extent. such episodes illustrate the in- 64K RAMs may have exceeded demand for a ability of the traditional tools of trade policy time. to deal with events in a fast-moving, technologically based industry like microelectronics; When the Semiconductor Industry Associa- they also illustrate the multiplicity of actors tionand Motorola specificallyaccused the populating U.S. trade policy and enforce- Japanese of dumping, though without filing menta multiplicity that some would charac- complaints, the Japanese responded by an- terize as leading to confusion and disarray." nouncing plans to move some production to the United States. Meanwhile, the Commerce/ In any case, the U.S. merchant semiconduc- Defense study had begun, evidently at the in- tor industry has not thus far sought direct pro- stigation of the latter agency. Among the possi- tection. Rather, American firms and their trade ble outcomes of the Commerce/Defense study, association(s) have continued pointing to fea- three appeared at the time to be among the tures of Japanese industrial policies and busi- most likely: ness practices they feel are unfair, urging the U.S. Government to exert pressures aimed at ending them.45 In addition, industry executives 1. A dumping complaint against the Jap- have sought Federal actions that would im- anese, self-ir; Hated by Commerce. prove their own ability to competelobbying 2. A section 232 investigation, based on the in favor of R&D tax incentives and reductions national security implications of U.S. de- in capital gains taxes, as well as calling atten- pendence on Japanese ICs. tion to engineering manpower shortages. Many 3. A complaint through GATT, probably concerning issues of reciprocal market access.

The section 232 alternative is noteworthy for 42See "Specialty Steel Industry Attacks Draft Report by Com- illustrating the variety of instruments that merce on Ferroalloy Study," U.S. Import Weekly, July 21, 1982, governments can bring to bear in trade-related p. 478. Ferroalloys are used in making steels. The investigation maiters. Part of the*Trade Expansion Act of was requested by a trade association of U.S. suppliers. Past applications of sec. 232 have been restricted to petroleum imports. 1962, this rarely used statute permits the Presi- 43A. Pollack, "Inquiry Puzzles Chip Makers," New York Times. dent to limit importse.g., by tariffs or quotas July 7, 1982, p. D9. ""Two dozen or more Federal agencies exercise some degree where such shipments "threaten to impair of responsibility over foreign trade and investment policies the national security." No section 232 pro- "Opening Statement of Senator Roth." Government Organiza- ceeding was started in this case, but a recent tion for Trade, hearing, Committee on Governmental Affairs, investigation of ferroalloy imports by the Com- U.S. Senate, June 4, 1981, p. 2. 454 International Competitiveness in Electronics

of the tax changes advocated by semiconduc- faced in gaining access to patents and other tor firms were in fact. implemented by the Eco- technical resultsis viewed as further evidence nomic Recovery Tax Act of 1981 (ch. 7). Indus- of asymmetries favoring the Japanese. In sum, -try-leaders have also-asked-the-Federal Govern- USsemiconductor_manufacturers believe that ment to negotiate for easier access to the Jap- a closed market shelters Japanese competitors, anese market, claiming that sales in Japan are leading to economies of large-scale production virtually impossible except for products that that translate into low prices. Investments in local companies do not make. The U.S. indus- production facilities within Japan are one way try's position is that asymmetries vis a vis Japan the U.S. industry sees to counter the threat. result in a competition in which the two sides are playing by different rules, with the advan- As the discussion above implies, trade out- tage to the Japanese. comes depend on complex sets of competitive As in earlier export thrusts, Japan's semicon- relationships, Consider, as an example, the ductor shipments have been concentrated in question of scale economiesa matter more a few product types. The choice has been mem- complicated than sometimes implied. A por- ory circuitsparticularly dynamic MOS RAMs tion of the cost savings, associated with produc- in part because of the close coupling between tion scale come via learning curve effects; he Japanese efforts in microelectronics and com- who gains an early edge in market share en- puters. Worldwide, Japan's RAM sales have in- joys lower costsperhaps permanently. The creased more rapidly than those of American Japanese, in the simplest view, "learn" by pro- firms, with European manufacturers the big ducing for the domestic market, then penetrate losers. At the end of 1979, just four compa- foreign markets based on low costs and low niestwo Japanese, two Americanproduced prices. But learning economies are not quite nearly two-thirds of the total world merchant so straightforward. Some costreductions are output of 16K RAMs. Two years later, a pair functions of cumulative production volume; of U.S. firms confronted six Japan producers others depend on time as wel1.46 In the latter in the baffle for worldwide market share in 64K case, obtaining a large early market share RAMs; while other American, manufacturers would not confer the same cost advantages. were ramping up production or preparing to The extent to which market penetration results enter with their own designs, Japanese com- in lower costs, therefore, may be product- panies were investing heavily in additional pro- specific; for some types of ICs, cumulative duction capacity. volume might matter much more than for others. The limited evidence available suggests Still, except for Mostekwhich is rapidly that costs for logic circuits depend more heavi- - diversifying its product line--RAMs have not ly on time, memory costs more on scale.47 As dominated the sales mix of any American com- a consequence, the advantages of access tothe pany. Thus, while Japanese incursions have U.S. market could be considerably greater for had drastic impacts on the RAM marketaf- products such as RAMs than for at least some fecting prices and profits, as well as market other types of ICsperhaps one reason Japan's sharessimilar shocks have not yet been felt exports have been so heavily weighted toward in other products. A major concern of U.S. pro- memory devices (the comparatively straightfor- ducers is that this experience will be repeated elsewhere, denying them the learning and scale economies so important for competitiveness, 4ea W. Webbink, The Semiconductor Industry: A Survey of Structure. Conduct and Performance (Washington, D.C.: Federal and cutting into the profits they need to gen- Trade Commission. Bureau of Economics, January 1977), pp. erate cash for expansion. Moreover, MITI- 49ff. sponsored R&D efforts like the VLSI project are " "Management Committee Report to the Management Review Committee C.n Assessment Rennrt. IBM. October 25. 1971 f PX Ch. 11U.S. Trade Policies and Their Effects 455 ward design of RAM and other memory cir- The vulnerability of the U.S. semiconductor cuits is also a major factor). industry to foreign competition is therefore a function first and foremost of technology. Even so, patterns linking design and develop- American firms with the ability to be consist- ment, manufacturing, and marketing tend to ently early to market with1,,,wproducts have repeat fairly consistently from product to prod- generally had less to worry about. Thus far, uct (see app. C on the 4K RAM). Technology most imports have been standard circuitsa dominates production and marketing strategies situation that could certainly change if semi- in .the early phases, because the firSt to offer conductor manufacturers in Japan or else- a new chip design may-possess a near-monop- where begin to design more innovative devices. oly--whether the chip is an innovative product At the same time, there are real limits to a tech- or simply an incremental advance. After intro- nology-based strategy. Incremental payoffs duction, prices decline slowly as firms recoup from R&D may diminish over time. Although R&D costs by selling to customers willing to new types of microelectronics products could pay premium prices for leading edge devices. still open new mass markets, signs of tech- Eventually, other manufacturers enter the mar- nological slowdown have begun to appear. In- ket, forcing prices downsometimes to a point evitably, the industry will mature, with greater below production costs. The final phase in the competition from imports a predictable conse- product cycle finds prices low, with many of quence: slower rates of technological change the earlier participants unable or unwilling to make it easier for foreign firms to catch up and compete; as a consequence, prices may even keep up. begin to rise once more. Industry structure is also important. The high-volume merchant manufacturers in the United States have coexisted with a fairlt1 large The competitive dynamics of the industry re- number of small firms for many years, the lat- volve around such factors. Some companies at- ter typically specialists filling market niches of tempt to be leaders, bringing innovative prod- less interest to bigger companies. Structural ucts to market early and capitalizing on the changes are underway in the domestic indus- liigher prices they command; in the United try, partly in consequence of heightened inter- States, Intel has become known for this strat- national competition, partly because the capital egy, which depends on heavy expenditures for requirements of advanced circuits make pur- design and developmentas well as abandon- suit of VLSI difficult for small companies. The ing products when they, begin to mature and changes are of two types, as discussed in pre- margins fall. Other firms manufacture a diver- vious chapters. First, diversified American cor- sified line of more mature devices,' concen- porations are purchasing or merging with trating on process technologies as a ,route to formerly independent, semiconductor firms. low costs. For such companiesNational Sem- Second, foreign enterprises are continuing to iconductor has been an example---Lmarket pen- take ownership positions in U.S. manufactur- etration is vital; as a result, they are particularly. ers. Such marriages have occurred, on the one vulnerable to import strategies that also empha- hand, because roreign electronics manufac- size market position. Still other entrants par= turers want quick access to evolving tech- ticipate largely as a byproduct of internal op- nologies, and on the other hand, because the erations. Semiconductors may account for only a small fraction of their business, but if they U.S. partners have needed infusions of capital. use substantial quantities in their other .end How will these structural shifts affect com- productsas do Japanese manufacturers such petition? As American semiconductor manu- as Fujitsu or Hitachithey may be able to sell facturers become larger and more diversified invirnr, .7,11 ,,,,Tre,r4111 / 456 International Competitiveness in Electronics

1 lines of business.Such)companies have the tors have no doubt contributed to the reluc- flexibility to shift resources internally, to meet tance of American semiconductor manufactur- the competition on a prirle basis, to invest more ers to press formal trade complaints against heavily in R&D and in 0.oduction equipment. Japanese exporters.. To, the extent that American semiconductor The old wayserecting trade barriers to firMs have had difficulty in financing expan- shield domestic industriescan damage U.S.- sion, with possibly harmful effects for U.S. based companies quite aside from any possi- j6irnpetitiveness, consolidation should help. bility of retaliation by foreign governments. /Multinationals also tend to have 'free trade" Texas Instruments produces 64K RAMs in Ja- /Perspectives because they depend on doing pan for export to the United States. In con- r ! business overseas, and are less likely to press sumer-electronics, the OMA on color TVs from / trade complaints. On the other hand,some Taiwan restricted shipments by RCA and Ze- observers believe large firms to be less en- nith, both of which had substantial investments trepreneurial and more cautious, and that con- there. Japanese semiconductor manufacturers solidation will reduce the prObability of innova- would quickly shift production to export plat- tion, making technical leadership more diffi- formsin many cases, the t,,..1mo offshore sites cult to maintain. While small firms will con- favored by American firmsin the event of re- tinue to existmany new ones have been strictions on shipments of ICs from Japan. started recently structural shifts of the type They would also move more production here. visible in semiconductors have charaCterized When those affected include U.S.-based mul- the maturation of many industries; rather than tinationals along with foreign firms having focusing on the-'llipposed virtues and liabilities significant interests in the United States, tradi- of small and lar e firms, it is perhaps more per- tional protective measures become let-, prac- tinent simply to observe that they have dif- tical. Such dilemmas have arisen, or are likely ferent strengths .and weaknesses in a given to, in other industries as wellautomobiles, competitive context. computers, possiblyaircraft,chemicals, energy, pharma. .uticals. Increasingly, firms More to the point in terms of trade policy, in such industries are tied by a multitude of what are the implications of greater numbers co-production and joint venture agreements, of foreign manufacturers with active semicon- irrespective of, the locations of their head- ductor design and production facilities in the quarters. As a consequence, the "inside-out- United States? This tendencyTarticularly evi- side" or "good guy-bad guy" distinction be- dent in recent investments by Japanese firms comes a difficult one for policymakers to draw; to mitigate trade frictionscombined with uni- in such a world, trade policies directed to an versal fOreign involvement by the larger Amer- older order may simply be overrun by events. ican p oducer§, has made semiconductor man- ufact re' one of the more international of the In any event, the American semiconductor worl/rs industries. Attitudes toward trade and/ industry has not attempted direct action to investment shaped by the traditional concerns staunch the flow of imports, much less with- of domestic firms and their workers may not drawn to a protected position in the United fit the realities of such an industry. In micro- States. Instead, while continuing to lobby the eleictronics and computers, the notion of "our" Federal Government, U.S. merchant firms have firms versus "theirs" is an oversimplification moved boldly to maintain their competitive- when so many companies operate on a global ness worldwide. This is one reason the in- scale. Corporations engaged in bitter trade dis- dustry's leaders have been more vociferous putes in one part of the world may establish . overwhat they view as unfair domestic sub- joint ventures elsewhere; cross-licensing is the sdies by the Japanese Governmentand over r;vrvIe r-Irsrsaofn virifb rmo nnnthar utrhorrs imnorlimorOc to their num AtiPMT1tq to qP11 nr Ch. 11U.S. Trade Policies and Their Effects 457

From the viewpc5int of American companies U.S. exports and imports' of semiconductoir that purchase semiconductors, the outcomes products will continue to be transfers between of intensified competition have been beneficial: / divisions of multinational corporations. Links a wide range of product offerings, low prices,',, among U.S. and foreign firms will certainly high quality. Would. these benefits have fol- persist and may well strengthen. Trade poliCies lowed even if U.S. producers had chosen strat- i. dealing with dumping or countervailing duties .egies or trade protection? This`must remain. an are unlikely tu be very relevant, if only beCause open questionbut to the extent that other inof the pace of technological changewhich dustries offer parallels, the benefits would/have can render :he products in question obSolete (heen smaller and slower to arrive before the proceedings have run their course. As the technological leads of U.S. mieroelec- /ironies firms narrow, and competition con- ( tinues to mount, the industry's support for Computers ( open trading relationships may diminish. Yet ( American semiconductor produCers cannot The picture is similar in the computer in- ( back too far away from a free/trade stance dustry. U.S. firms have been undisputed lead- without jeopardizing their own overseas in- ers, with subsidiaries engaged in manufactur- terests. Formulating equitable/ trade policies ing and marketing around the world. Imports will continue to be difficult, with a wide range have been at low levels, even for personal com- of interests to be balanced. A/large fraction of puters and peripheralsalthough this could

.4.mst=.....P.41NeSSESSE121E1 458 " International Competitiveness in Electronics certainly change. The consequence has been, to narrow, following that in hardware. The flu- until quite recently, a virtual absence of con- idity of market structures emphasized in chap- cern by executives of American computer ter 5 Will leave room for aggressive foreign firms tvitir U.S. trade policy except as'it sup- competitors. ports an open international trading environ- As a result of the subsidies for computer tech- ment or affects transfers of components and nology that virtually all industrialize f7ountries subassemblies among subsidiaries. American have employed, it is not hard to envision a sce- manufacturers wish to see items 806.30 and nario in which these subsidiesas well as the 807.00 of the Tariff Schedules -preserved, but preferential treatment many governments have dumping or escape clause provisions have sel- extended to local producersbecome the tar- dom attracted their attention.. gets of countervailing duty complaints. If a Yet here too the competitive picture is chang- fOreign firm benefiting from government. lar- ing. Foreign governments, viewing information gess were to establish a significant market posi- processing as vital to national interests, have tion in the United States, can there be much found a variety of methods for subsidizing local doubt that American manufacturers would firms, as well as an array of carrots and sticks seek remedies under U.S. law? After all, the for encouraging American corporations to subsidies extended to foreign computer indus- transfer technology to local computer man- triesalbeit often rationalized on national ufacturers.48 And again as in semiconductors, security groundshave been even more visi- foreign manufacturers have taken equity posi- ble than in microelectronics. What, then, might tions in American computer companies, in part be an appropriate response on the part of the to acquire technical knowledge. The technol- U.S. Government? ogy..gap between U.S. and foreign firms has The issue,raisedand repeated in semi- diminished in both hardware and software, conductors, communications equipment, and with several Japanese manufacturers beginning other high-technology products--is that the ex- to ship mainframe machines to the United isting structure of national and international States either directly or in partnership with trade laws andagreements evolved.in another American firms. era; it was not designed with current varieties What impact will such developments have of national industrial policies and subsidies in on trade flows and on U.S. trade policy? Given mind. Countervailing duties were intended to that competition depends on much more than offset export subsidies,,such as rebates or other fast, reliable hardware. it is too early to make payments contingent on sales to overseas cus- predictions. Manufacturers must be closely at- tomers. Subsidized financing via export-import tuned to user needs; foreign computer firms banks has strained the system. Domestic sub -. lag well behind American companies in their sidies with indirect effects on exports scarce- ability to seek out and satisfy customer applica- ly fit it. Indeed, U.S.-based companies maybe tions (ch. 5). This deficiency has not gone un- among those benefiting fromindustrial poli recognized; other nations are devoting substan- cies in other nations. Antidumping laws were tial efforts to software, often aided by govern- drafted to counter explicit price discrimination ment subsidies. Furthermore, countries like by foreign monopolists, often involving govern- Britain. have always been good at software and ments and/or cartels that encouraged exports may provide a resource that firmselsewhere by charging higher prices to domestic than to can bp. The U.S. lead in software seemsbound foreign customers. Until recently, antidumping legislation.was seldom called on where "Overseas governments have used investment inceniiycs to price-cutting was extended to all customers, attract AllWriCall fitans more actively in electronics and elec. domestic as well as foreign. Today, when Ch. 11U.S. Trade Policies and Their Effects 459

Boeing? Or, more subtly, is U.S. support of re- longer term. The question for trade policy search into solid-state electronics as part of becomes: As governments increase their in- military and space programs unfair---research volvement in economic affairsand indeed in that, after privately funded follow-ons, even- the actual operation of business enterprises= tually results in commercial applications? what types of trade policies and agreements Many other examples, in any number of coun- will be needed so that participants can agree tries, could be cited. that the terms of competition are reasonably fair? This will remain a central matterk for in- For computers or communications, govern- ternational trade negotiations over the current ments`have seldom proffered financial assist- decade and beyond. While the Tokyo Round nce simply to foster exports, although this has trade negotiations addressed such questions, been one motivea strong one in Japan. Rath- the substantive chang:.. in procedures embod- er, governments have targeted.the information industry as vital to a multiplicity of national ied 2n the new subsidies code are small, and interests. Subsidies have been generalized, unlikely to have much effect. directed at industrial development over the

Summary and Conclusions

U.S. trade policy has been rather consistently by American firms against competitors in the oriented toward open international trade and Far East, primarily Japan. The response of the investment over the last half-century. In the Federal Government has been marked by de- postwar period especially, the United States lays and interagency conflicts. Fifteen years took the lead in eliminating both tariff and non- after the initial antidumping actions, the situa- tariff barriersby reducing its own restrictive tion remains unresolved, duties uncollected. measures and pressing its trading partners to The uncertainty created by this long and con- do the same. Some parts of the electronics in- voluted history has made life difficult for both dustry, notably manufacturers of consumer domestic firms and importers. To considerable goods like TVs and CB radios, have suffered extent, as the shape of the industry has altered, as a result. But if import competition has hurt complaints over trading practices have become the manufacturers of such products, othersec- moot. Orderly Marketing Agreementsnegoti- tors of the U.S. economy benefited from free- ated as an upshot of escape clause proceedings dorn to export and invest overseas. U.S. trade unrelated to unfair trade practicesacceler- policy has helped American semiconductor ated what would probably have been wide- and computer firms become leaders in markets spread eventual movement by foreign firms to- all over the world. The exception has been ward assembly here. Plants owned by foreign Japan; barriers imposed by European nations interests have replaced failing domestic TV have proved far less substantial. On the whole, manufacturers. Meanwhile, the remaining U.S.. the open trading environment resulting from producers have moved some of their assembly successive rounds of multilateral negotiations operations to low-wage offshore locations, has helped the competitiveness of the U.S. elec- helped by provisions of items 806.30 and 807.00 tronics industry. of the U.S. Tariff Schedules. These provisions which allow tariffs on re-imports to be com- More narrowly, trade policy impacts in con- puted only on foreign value addedhave been sumer electronics have centered on longstand- a target of labor interests. But if in some cases 460 International Competitiveness in Electronics lent to the export of jobs, in others transfers not the primary er avowed intent. While direct offshore have been necessary to retain any impacts on international trade tend to be small, domestic jobs. the visibility of programs such as Japan's joint R&D efforts, or West Germany's spending on As tariff walls in many parts of the world computer technology, draws frequent attacks have_ slowly come down, the attention of both by businessmen and political leaders in other private firms and governments has turned to parts of the world. The nations mounting these indirect and nontariff barriers. Ranging from programs consider them vital for economic de- implicit subsidies for domestic firms to un- velopment; they will not disappear. As has cooperative customs inspectors, such barriers been the case with complaints over unfair trade; pose much more complex subjects for interna- practices in consumer electronics, negotiations tional negotiationsas the Tokyo Round dem- concerning indirect supports and subsidies are onstrated. Some progress. has been made, but being overtaken by events; subsidies may in it is too early to tell how trade in electronics some respects function like other nontariff and may be affected over the longer term. indirect barriers to trade, but governments Among nontariff and indirect measures, sub- seldom institute them for such purposes. Nor sidies for economic and industrial develop- do they view them as elements of trade policy. ment ostensibly aimed at domestic objectives They are seen as vital tools of industrial pol- are perhaps the most difficult case, along with icypolicies developed in response to an eco- government pi ocurement. As discussed in the nomic environment in which domestic and in- previous chapter, many nations have used both ternational dimensions can seldom be isolated. types of measures consistently and aggressively as elements of industrial policyespecially in Negotiations aimed at reducing such subsi- electronics. While progress has begun in open- dies make slow progress at bestindeed, ing up government procurements, subsidies although necessary, they may finally be rather will remain thorny issues for years. More and beside the point. So long as governments re- more governments, for example, are resorting gard high-technology industries like electronics to R&D incentives to support local electronics as essential to industrial development and eco- manufacturers; inevitably, these function to nomic growth, supports and subsidies seem some extent as export subsidies, even if this is more likely to, increase than decrease. CHAPTER 12 Federal Policies Affecting Electronics: Options for the United States Contents

( hie i'Vii!W 463 The Current Policy Environment 464 Alternative Perspectives on Industrial Policies 467 1.;usurt! dit llomestic Market Base for 11.S Industries 468 Support for rritical Industries 475. lolr&ctructural irod Adjustment Policies 485 Promoting the Clohal Competitiveness of -American Industries 491 Inthistrial Policy Centered in the Private Sector 496 SummarV and Conclusions 500

List of Tables

Nle,rwres Likely to be Emphasized roofer Alternative Apploaclms 1,1 Industrial PolitN 468 CHAPTER 12 Federal Policies Affecting Eiectronics: Options for the United States

Overview

In 1982 electronics accounted for 10 percent mined efforts, strongly supported by national of U.S. merchandise exports; manufacturers of governments, continue in countries like France electronic uroducts contributed 4 percent of and 'West Germany. Japan has moved swiftly U.S. gross domestic product and 9 percent of from a position as technological laggard to be- total goods output; the industry employedmore ing one of the leadersnot only in product de- than a million and a half AmeriCans. Beyond velopment, but in the fundamentals of elec- this, the electronics industry produces goods tronics technology. The strengths of the and knowledge that are vital for much of the '`,,,panese industry have been described in rest of the economy and for society as a whole. previous chapters: an ample supply of skilled Banks, insurance companies, and many other and motivated employees; managements that service sector enterprises could hardly conduct approach markets on a global scale and have their businesses without electronic dataproc- learned to do business effectively in countries essing. Entertainment industries like television ranging from Saudi Arabia to the United States; broadcasting developed in conjunction with an economic system in which the tradeoffs be- the manufacture of consumer electronic prod- tween competitive rivalry and cooperation aim- ucts. Before too long, prosthetic devices for the ed at advancing common goals are well-man- handicapped will routinely be built around aged; a government whose industrial policies "smart" microelectronic devices. Modern com- consistently support and encourage the private mercial aircraft depend as heavily on radar, `sector. In the language of sports, the Japanese computerized flight control systems, and-elec- electronics industry has momentum. Else- tronic navigation aids as do military planes. where in Asia, developing countries can al- Over the rest of the century, computer-aided ready make many consumer electronic prod. manufacturingshop floor management sys- ucts and compopents at less cost than japan tems, automatic warehouses, smart robots or the United States; these countries will con- will helo increase productivity and redtice tinue to move into more sophisticated goods costs throughout the Nation's manufacturing though at first continuing tVocus on consumer sector. If there is any single industry whose marketssupported by export-oriented indus- technological progress and competitiveness are trial policies. Hong Kong, for one, has already critical to the economic growth and national made the transition froin discrete transistors security of the Uni;_ed States, it is electronics. to integrated circuits (ICS")computer chips as Yet even the most dynamic portionS of the weld as thos'e for consumer products; This is U.S.. electronics industrysemiconductor pro- riot to say the Ainerican electronics industry duction, where innovation is a way of life, com- risks overnight decline. It does mean that computers, where markets seem to expand nearly petition will be difficult in the years ahead, and as rapidly in bad times as good find them- neither the technological leads that the United selves increasingly challenged by foreign States still maintains nor the size and affluence manufacturers, both here and overseas. of our domestic thorket will suffice to guaran- Although few European firms have managed tee American primacy. 464 International Cornpettlivoness in Electronics

might choose to move toward a more explicit protection of domestic markets; industrial policy of its own. If other countries a "critical industries" policy; , area adopting policies in supportof industries an orientation toward infrastructural sun like electronicsand having some success, or, portprimarily for technological develop- more tellingly,-mOving downlearning curves mentand adjustment; leading to more consistently productive ef- promotion of competitive U.S. firms on a fortsthen it makes sense to ask whether such worldwide basis; and an approach could alsowork here. Such ques- Federal withdrawal, where possible, in fa- tions need to be addressed in terms of concrete vor. of the private secto7. policy objectives and prospective mechanisms The five are intended to span the realistic alter- for achieving -themmatters to whitit this natives; they overlap Somewhatin several chapter is devote 1. cases specific policy measures would be simi- with- lar if not identical. Still, if the five options The United States could afford to live themselves are not exclusive, they are distinct. out a consciously formulatedindustrial policy Each is discussed in terms of prospective ef- in years past, but the realities ofinternational fects on electronicsand, by extension, other economic competition -changed--dramatically-- high-thchnology-sectors:- -- during the 1970's. The "informationsociety" is here; industries are no longerdefined by na- The options are discussed in terms of direc- tional new jobs are openingfor tions and objectives; they are intended to of- Americans who are computer-literate at atime fer a set of alternative signposts. All start from rwhen employment of other types is indecline. the same point: the patterns of worldwide com- 'If new approaches to industrial policy areto petition outlined in the preceding chapters; the be considered, what is the rangeof possi- increasing capital7intensity of critical sectors bilities? of the electronics industry; its continued dependence on research and development (R&D); After a brief review of the ad hoc natureof needs for skilled labor and imaginative man- past U.S. industrial policies,the chapter out- agement. But each alternative implies a dif- lines five options for a more focusedapproach: ferent ro'ite to a different destination.

The Current Policy Environment

Chapters 10 and 11 summarized U.S. indus- ing equipment to copy off the air; the fate of trial and trade policies as they affect competi- AM stereo broadcasting (where the Federal tiveness in electronics, contrasting them with Communications Commission has avoided de- policy approaches - abroad. Federal policies cisions on a standard system); regulation of have been notably ad hoc, formulated and im- hoir:e information services and data commu- plemented by many different agencies, no one nications (videotext and teletext may be slow of which has overriding authority. Tradepoli- in corning, but will eventually be integrated cies are neither very predictable nor closely into home entertainment and information sys- linked with domestic industrial and economic tems). As with regulation of cable TV and policies. rights to satellite transmissions, such matters may have only indirect consequences forman- ,Outside the trade arena, regulation of broad- ufacturers of consumer electronics products, casting and telecommunications has been a but impacts that are no less real for this. continuing influence on consumer electronics. Among the recent issues with potential impacts The U.S. semiconductor and computer inan U.S. Manufacturers, either direr,t or industries benefited in their early years from Government-funded R&D and procurements; . direct, are: the rights of owners of video-tap-

O Ch. 12 Fedeial PohcieS Affecting Electronics: Options for the United States 465

military and space programs drove develop- cal climate in which industry is reluctant to test ments in microelectronics until the 1960's, the bounds of the permissible. The two agen- while both civilian and military sides of the cies generally act as if this status quo is desir Government have been heavy purchasers of able, with their public pronouncements nice- computers and related equipment. Although ly straddling the relevant issues.' While the Federal R&D is still significant for both micro- forcement attitudes of the Reagan admini5. electronics and computers, particularly in tration differ somewhat from those of its im- terms of more basic researchand in the case mediate predecessors, the more relevant e;on- of the Defense Department's Very High-Speed cernat least from the viewpoint of Congress Integrated Circuit (VHSIC) program, process might be: Do the antitrust statutes needrecon- technology and applied research as wellit is sideration in light of changes in the character now ,,decidedly secondary to the industries' of international trade and competition since own spending. Furthermore, the declining the Clayton and Sherman acts were passed in fraction of total markets t'or both semiconduc- the early decades of the century? Where would tors and computers accounted for by Govern- antitrust fit within a more coherent U.S. in- ment has made Federal procurement a much dustrial policy? weaker force than 10 or. 1:3 ago. Beyond the intangible effects of Federal an- Indirectly, a wide variety of policies titrust_enforcement--beginning with their-force affect U.S. semiconductor and,computer firms: in restraining clearly undesirable forms of an- support for education and-training, particularly ticompetitive behaviorchapter 10 mentioned of technical professionals; regulation of data the two recent and major antitrust actions di- communications; the 1982 antitrust settlements rectly involving the electronics industrysuits with IBM and AT&T; copyright and/or' patent it IBM and AT&T, both recently settled. protection, for chip designs and software; tax In both cases, critics of U.S. antitrust enforce- policiu as they influence competition for funds ment had claimed, with some validity, that the within U.S. capital markets. On the other hand, Government was trying in the name of compe- Government policies directed specifically at tition to break up enterprises that were main- the semiconductor or computer industries- gtays of U.S. competitiveness. -either_in_thedomestic context or in terms of theil'intPr -.Itionarcompetitive positionsare Regardless of possible rewrites of communi- remarkauid by their absence, the vacuum a cations legislation by Congress, the AT&T set- striking contrast with ambitious, comprehen- tlement will change the form and function of sive, and supportive (if not necessarily very Bell Laboratories. A weakening of the basicre- cost-effective) ptiblic policies in other coun- search foundation that Bell Laboratories helped tries. A major exception to the absence of poli- 'Consider this quotation from the "Statement of William F. cy is antitrust. Baxter, Assistant Attorney General, Antitrust Division, Before The Subcommittee on Employment and Productivity, Commit- American electronics firms have often com- tee on Labor and Hunan Resources, United States Senate, Con- cerning Antitrust Poiicy and Productivity, Apr. 16, 1982," pp.. plained that antitrust enforcement hinders 12-13. "Joint ventures may foster efficiencies riot available to cooperative R &D and joint ventures in inter- individual firms, and may promote technological progress and national trade- which,ifallowed, would enhance productiiity. Such joint ventures should not be deterred. by rigid or overly-broad applications of the antitrust laws. The strengthen this country's competitive position. Department's recently published Guide Concerning Research The Department of Justiceand, to the extent Joint Ventures is intended to assist businesses considering joint that it is involved in antitrust enforcement, the ventures by clarifying our enforCement policies in this regard, For example, as the Guide indicates, an important factor is Federal Trade Commission (FTC)reply that whether a joint venture leaves a significant number .ofnon- such cooperative activities are, in fact, general- participating firms free to engage independently in research. If ly permitted. The dialog, which has been go- there are not a significant number of such non-participating firms, and the joint venture's research could be done individually ing on for years, is part of the prohloi-7,1: Justice by the participating firms, antitrust problems could arise:. Much and the FTC have helped create ..( psychologi- of the lore of antitrust resides in such pronouncement's.

463 466 International Competitiveness in Electronics lay for so much of the U.S.and world--elec- tury. Others focus on organization, some advo- tronic-aiustry seems inevitable. The reason cating that the Department of Commerce be is straiAinforward: Western Electricthe pa- transformed into a more powerful agency rent and main customer of Bell Labswill be even a "Department of International Trade and operating in a more competitive environment. Industry" modeled after Japan's MITI and re- It will iose some of its "guaranteed" markets, sponsible for coordinating and implementing as well as implicit subsidies paid for in the past policies on a sectoral basis. Another view, by all telephone subscribers. Western Electric while agreeing that new mechanisms are and 13011 Laboratories will have to learn to deal needed if the Nation's de facto industrial policy with a less predictable set of market conditions; is .to be replaced by a more systematic ap- basic research is likely to appear more in the proach, takes dispersal of responsibility for nature of a luxuryand hence be deempha- making and implementing policies to be a hall- sized. Unless alternative mechanisms for per- mark of the U.S. system, and sees a siagle cen- forming basic research and diffusing the re- tralized agency as impractical, if not danger- sults evolve and thive, one of the great sources ous. Still others, focusing on financial issues, of strength for the entire U.S. electronics in- argue that the first priority of the Federal Gov- dustry will atrophy. Th3 potential void is not ernment should be to channel investment capi- restricted to inicroelect.onics, but extends to tal to speed "reindustrialization"for exam- communications and computer Technologies as ple, through a publicly operatedinvestment or well; Bell Labs spawned the Unix operating development bank. In any of these views, elec- system now so popular in small computers, as tronics would plainly be an early subject of well as the transistor. attention. Leaving aside the special case of antitrust The dominant attitude within the executive and perhaps trade policies also, where activi- branch since the election of President Reagan ty has remained at high levers if without much has run counter to the more activist positions. sense of directionFederal policies affecting The Reagan administration has held that the electronics have been marginal, indirect, often proper role of-Government is to stayfar re- simply absent. To those who regard 'Govern- moved from the affairs of industry. Those of ment involvement in the affairs of industry as such persuasion believe that businessmen, a usually unnecessary evil, this may seem a rather than Government officials, have both the blessing. But the reasons for the absence of right and the ability to make decisions affect- polioire not so much conscious decisions that ing the futures of their firmsand thus the fu- Fee- initiatives would be counterproductive tures of the industries of which they are mem- as a nAck of agreement about what Government bers, and the competitiveness of the U.S. econ- can and should do. The subject has been dis- omythat Govei nment is largely incompetent cussedat considerable length; and in contexts in these areas. The corollary is that decisions ranging from trade reorganization to the Fed- made by private interests will affect local and eral role in prodti ivity improvement and the. regional economies, as well as the interests and quality of working'fe. But the various parties livelihoods of the people,who work for or other-- disagreebeginning with the question of wise depend on private industry. whether the U.S. Government should develop Those at the extreme end of the spectrum policies aimed at affecting the competitive posi- hold that Government should minimize its ef- tion of American industries, and extending to forts at macroeconomic policymaking, claim- questions of how the electronics industry, in ing that the Keyii,Aan economistswho, many particular, might fit into a more general frame- years ago, defined a goverrirr.:.Introle in man- work for.industrial policy. aging the aggregate economy through fiscal On the one hand, some argue that the United and mono-v policies represent abankrupt States needs to search for new engines of -tradition. Ta..; reductions and other measures growth to drive the economy into the 21st cen- aimed at capital formation,. rather than de-

4 Ch. 12Federal Policies Affecting Electronics:: Options for the United States 467 mend management, have been in vogue among that Government involvements will always be the proponents of this view. Still, even ad- riddled with mistakes arid policy failures, this vocates of supply-side econe-. ics or monetar- is more a counterargument to proposals for a ism commonly maintain thatwhile Govern- strong, centralized industrial policy apparatus ment should stay out of the :.`fairs of busi-. than to the general notion of a more explicit nessit can and should ensure a climate con- and coherent indusTrial policy. ducive to economic In essence, two of the attitudes sketched There is a core of truth to the claims of those abovecentralized industrial policy versus who advocate a Government pullback from mi- Government pullbackrepresent extremes in cro-level involvement in economic affairsthis opinion concerning the form and character of troth found in the past history of public policies future industrial policies fox the United States. that have affected U.S. industries with some- Industrial policies have existed for several hun- times adverse consequences. to say nothing of dred years in this country; Federal Government the lack of relevant expertise and analytical actions will continue to exert influence over Capability in the executive branch. On the other private sector decisions. On quetstions of how hand, it is clearly possible to devise a self- the process might be changed, as on specific conscious industrial _policy that does not de-. _policy issues, there are many shades of opin- pend on direct or extensive Federal interven- ion, a variety of perspectives that: fall between tion. Thus, even if one feels that the pOlitical the extremes. These are illustrated in the re- system in the United States works in such ways mainder of the chapter.

Alternative Perspectives on Industrial Policies

Regardless of one's attitude toward Govern- judged critical on national security or ment involvement in economic affairs--when other grounds. and where appropriate, for what reasonspol- 3. Policies that will support the technological icy choices will flow in part from analyses of base and institutional iFfrastructure for the position of American industry, and the in- American industries, particularly those terpretations placed on these analyses. Each undergoing structural change. of OTA's five alternatives has as its founda- 4. Policies designed to promote the global tion a somewhat different interpretation of the competitiveness of U.S. firms and indus- competitive situation of American electronics tries. firms; while the discussion that follows takes 5. A policy that defers if possible to the pri- its context froin this industry, the policy op- vate sector when choices concerning in- tions are not specific to electronics, or ever: to dustrial development are to be made. high-technology industries as a class. The five All of these, even the laSt, accept at least im- alternativeswhich .overlap to some extent plicitly that Government involvement in indus- while representing fundamentally different try and the c:-::onomy is inevitablethe ques- viewpointsare: tions being vven, where, how, for what pur- 1. Policies intended to ensure a strong do- pose. In each case, different sets of assump- mestic rit,,Lket base for U.S. industries tions and goals underlie the policy orientation. without particular reference to the nature From the repertory of policy tools available of the industriesalong with preservation outlined in chapter 10 and summarized in table of existing jobs and job opportunities. 82each of the five would call for a different 2. Policy measures designed to protect and/or mix and emphasis. The table is schematic, but support a limited number of industries gives an idea of the types of measures that

4a1u 468 International Competitiveness in Electronics

Table 82.Measures Likely To Be Emphasized Under Alternative Approaches to Industrial Policya

Alternative Critical Infrastructure Global Minimum Protection industries & adjustment promotion Government Trade, foreign investment Tat Competition antitrust, merger) Human resources (education, retraining) Technology (R&D, innovation, diffusion) Irvestment (capital) GovernMent procurement ,..- arn,s table is :mended only to be suggestive. many POssdpilities exist under ei..ch alternative 50tinCE Of-,e of Technology Assessment would be emphasized under each of the alter- cumstances under which trade restraints might_ natives; a "critical industries" orientation, for be justifie.! to prevent or ameliorate greater instance. would entail a strong presence by harm to a iewvulnerable firms and their em- Government compared to the other four. ployees, communities and regionsat the ex- Regardless of which of the five alternatives pense of net benefits that, when spread over wiTe chosen, a new approach to industrial pol- the Nation as a whole, are small. icy for the United States would bring pitfalls Leave aside for a moment the political ques- as well as opportunities; experience in other tions, as well as the use of protection to countries shows that there is no substitute for countervail the industrial policies of other na- good judgment in selecting and implementing tions, or unfair trade practices by foreign enter- individual policy measures. Table d2 stresses prises. The question is then an internal one: that it is not so much the individual policy tools What are the impacts within the larger domes- but the way they are put togetherthe objec- tic economy of protection granted a particular tives pursuedthat matters most. The remain- industry? This is not only a matter of present- der of the chapter treats the five alternatives day costs and benefitse.g., to consumers, to in detail. owners, managers, and other employees but of the future prospects of industries granted Ensure the Domestic Market Base protection. Some such industries :nay be in for U.S. Industries temporary decline. with reversal possible others unequivocal victims of shifting com- Protectionism is a loaded word. Not only parative advantage. For example, long-term does it imply reversal of the primary thrust of prospects for specialty steel manufacturers in postwar U.S. foreign economic policy,. but the ilia United States appear brighter than for arguments in support of open international wakers of carbon steel. Not only do the tech- trade are strong and widely accepted. Protect- .-;al demands of specialty alloys favor Amer- ing domestic industries from import competi- ican firms, but the diversified, high-technology tion via tariffs, quotas, or other barriers distort?. industries of the United States provide large market mechanisms, decreases economy'3ffi- and varied markets for alloy steels. None- ciency,andby raising pricesresults in a net theless, both specialty and carbon .steel pro- kiss in standard of living.2 Hardly anyone dis- ducers face short- as well as long-term prob- putes these general tenets; the issues more lems. It is possible, to argue on the one hand commonly raised concern the' specific cir- that trade protection will benefit specialty steelmakers by permitting them to rebuild their 21'or a brief review,1. Site U.S. Inchstrial Competitiveness:. competitiveness so as to take advantage of Comparison of Steel, Electronics, and Automobiles(Washingto,.!, D.C.: Of ficr: ofTechnology Assessment, O'fA-1SC -135, July 198 :), longer term opportunities, while on the other pp. 181182. that pibtection for carbon steel producers will Ch. 12Federal Policies Affecting Electronics: Options for the l'-;ted States 469

tgal

..,._.4....._...... _t -.4 -- _._?,- rilV:: -- .....q4Lik'." ."-- '6°2' - .. -----7-- .1-----x:: _,,,,,.:._;j.-,-*.a.4%,2. '"";-..;- ''-''''.-'j..41 .to--` - '.-- 4*.,"5.atil4 A. P.------r 4 v.,10....- -4&67°'`-----

- -

Fholo c 'ode!: RCA American consumer electronic plant complex only retard the inevitable contractions. At the rnunications equipment. Similar bargaining same. time, for domesticindustries as large as has taken place with European nations export-. steel or automobiles, the adjustment problems ing steel, among other products, to the United .stemming fromi long-term shifts in competi- States, as well as between Japan and the Euro- tiveness can iso severe that strong arguments pean Communitye.g., in the case of video cas- for temporal-, trade protection can be con- sette recorders (VCRs).3 To some, this revival 5:tructed on this basis alone. This is one of the of a bilateral rather than multilateral approach reason%.. escape clause actions aresanctioned to trade is a sign of possible return to the tinder Gi.TT (the Gene.-%!greement on Tar- prewar era of widespread protectionism. Bilateral negotiations between the United States and Japan first found a prominent place Biliite;alism in Trade in U.S. trade policy during the early 1970's, Protective actions other than tariffs, involv- when market penetration by Japanese textile ingas they generally dobilateral discussions imports became severe; in one of the more re- with exporting nations, represent something of a turn away from the postwar U.S.emphasis Faced with dumping complaints and informal import rostric- on multilaterPi trade negotiations:Persistent lions in France, Japan has voluntarily agreed to limit VCR shipments to the European Community. The ceiling will be4.55 trade friction between the Utii;:ad Statesand million annually. E. J. Dionne, Jr., "Japan Video Accord Leaves Japan !"1,19 led to bilateral negotiations (:ever- Eu:op. ans Wary but Hopeful," New York Times, Feb. 22, 1983, ing ranging From cigarettes to tef,sz.om- p. D5.

472 470 Intern3tional Competitiveness in Electronics

cent cases, after a push from the U.S. side, the of the President in dealing with other nations Japanese agreed in 1981 to accelerate tariff passed the Senate unanimously. House bills reductions on semiconductors (ch. 11). Further have made less progress. negotiations led to similar concessions on data Most broadly, reciprocity is a call for equal processing equipment. Japan's continued reli- treatment, hence would entail measures to re- ance on tariff walls to protect domestic indus- strict imports originating in nations that them- trieswhether or nut they remain infantsand selves block the entry of American firmspar- on an array of slowly crumbling and largely ticularly through indirect barriers. Spokesmen informal nontariff barriers, contrasts sharply for the U.S. semiconductor industry, to take an with duties on imports into the th,,ited States example from electronics, object to unlimited that for many years have been low; as pointed imports of Japanese ;Cs at a time when they out in the previous chapter, this country has 7'ee themselves confronting a formidable array only rarely imposed tariffs to protect domestic of obstacles to doing business in Javnobsta- industries, preferring to negotiate quotas. cles ranging from uncooperative customs in- Recently, those concerned with Japanese spt;...ors to hidden controls on foreign invest- penefration into U.S. markets for products ment ' ranging from semiconductors to machine tools have beet urging a variety of essentially pro- Pros and Cons of a Protected Market tectionist responses. The mirror image of con- Base Strategy cern with imports lies in the persisting diffi- Arguments for temporary as opposed to long- culties many American firms have faced in er.term or permanent trade protection turn on exporting to Japan, or investing thereeven quite different points. Proponents of temporary when the exports are goods in which the Jape prGtection for troubled industries.often judge nese economy is uncompetitive. The paramot.a sharp upturns in import shipments, as for col- example has been agricultural products.5 The or televisions during the 1970's, to be par- perceived asymmetry has been a major force ticularly serious. Once lost, whatever the behind calls for reciprocity in trade. reasormarket share can be ;difficult to regain. While to some, trade reciprocity need not Thus, a sudden penetration of U.S. markets cart y the in.).:r:ation of sector-specific bilateral perhaps as a result of unfair trade practices concerns, tot,.rs a means just that: if nations might devastate an industry, leaving it without such as Japan discriminate against U.S. exports the ability to recover. The remedy is to protect or investm .nt,'we should retaliate swiftly and the industry. Whether the causes are lower directiy. 5tirinp 19F20 or more bills related costs for labor or other factors of production to um -pci!yand covering many abroad, unfair trade practices such as dump- shades of r r introduced, with ing, or problems internal to the U.S. indus- many re-lnli :ed in the 98th Congress.6 In trywhich could range from outdated plant fa- April I ;183, S. 14 the Trade and Investment cilities to misjudgments of the market the ob- Act of :983, intended to strengthen the hand jective of Federal policy, in this view, should be to limit import penetration with the expec- `An nt exception has been tha levies of 45 per..:ent. declining o.:er a 5-year period. placed on large japanose --- tation that, after a limited period of relief, cycles early, n. 1983 These wen: imposed as the rest'' domestic firms will again be able to compete. escape daily! action. See "President Ira tees Sharp crease cm Motovc,,cies. ;span Criticizes ...c.tion." 71.S. Import Preferred measures to achieve such goals de- Weekly, Apr. 6. 1083. p. 5. pe.on the circumstances of the import-af- 5See, forinstancy,ReporttinTrade Mission to F:..r Rast, SUb- 4.011111'01..0 4,11 Teild,!, Committee on Ways and Mc as. U.S. fectPd sector. Examples from the recent past House of Representatives. Dec. 21. Mt include tariffs, uni:uterally imposed quotas, "See A. Redman and R. Ahearn, "Reaii.rocity is t-oreign 'Friuli.," Issue Brief No. 111820.13, Congressional Rg.seach Serv- '(in the latter, see U. C. Lehner, "Iti:ian'i Av.- &al to Sening U:e. N1i.:r. 3 1083. Most of these bills won' intended to give the Cornpanis May Be Ultirnatii E, :: rirr in rade," MI' Street President authority to impose restr:ctions1,n imports: Journal, Mar. 23, 108Z, p;.38; also. S. Loh r. ulapa.)', Capital se..:iral were IF re( led largely at /rad,: in .,:er.,,icet Market lias0.S, Ne's York Times. limo 1, 1082, p. D3. Ch. 12Federal Policies Affecting Electronics: Options for the United States 471

negotiated Orderly. Marketing Agreements from those of firms that must live off their own (ONIAs). and voluntary restraint on the part of profits. And if not all countries have national- exporters. In industries where allegations of ized sectors the size of that in France, in many clumping have been commonconsumer elec- economies the incidence of government sub- tronics, and, though there have been no formal sidy and control is such that market signals co,-nplaints in recent years. semiconductors become disti.ictly secondary. The preceding alternatives to antidumping proceedingT. might chapter stressed the relative impotence of the be sought because legal redress has proven traditional resler of trade lawsand of inter- \slow, notably in the case of television imports. national re?,--.nations for countervailing the `fhe Trigger Price .Nlechanism for steel illus- wide range o:*.s!tpports and subsidies that some trates a more nova 1 mechanism. governments nol.v resort to. Those who see world trade as moving toward a no-win sit :a- 1.teyond temporary protection, the notion that firm for the United E.tates sometimes urge that the\ Federal. Government should provide a 1.ve shut our own borders to imports, accepting ha0.ri for tS. industry appeals to many in- the consequences in terms of reduced exports teress. mi,,,ht_be.to help the while relying on the size and diversity of the American economy grow, to protect jobs, to U.S. economy to keep productivitymore gen- maintain the prosperity of cities, States, and erally,thegross domestic producthigh regions,\,What are the arguments for more corn- enough to maintain living standards acceptable prehensiVe or longer term import restraints? to most American: Policies designed to ensure a domestic market base for American industries might be jus- A related justification for trade restraints tified on the assumption that a certain level of starts with international differences in wage sales at home provide the necessary foundation ratesa 7)oint emphasized in chapter 4 (see for international competitiveness. In essence, table 27)., .ow-wagecountries, many with huge a variant of infant industry, senescent industry, and mounting labor surpluses, can now pro- and critical mass arguMents, at root this per- duce many types of goods at costs below those speetive views in-wort penetration as inherently in advanced nations. Increasingly, this is true dangerous, hence' worthy of Government atten- over a range from primary metals to manufac- tion. Keeping out imports permits domestic tures like automobiles or the simpler electron- manufw.:turers to achieve scale economies and ics products that were mainstay: of countries to earn the profits necessary for investments industrializing earlier. Although labor produc- in new production facilities and R&D. A proL. tivity in developing countries is often very low, tected home market vuuli:. also insulate' hem if wages are also low, costs of production can from sudden and unexpected competitive be less than else,where. Eeen Japanwith pay threats. originating not in profit-seeking scales in manufacturing industries more overseas firms, but in goVernment-controlled th:.ti half those here. and labor productivities enterprises seeking to create jobs, earn foreign in some cases better faces competit'.fe diffi- exchange, build industries, that can sup,iort culties in sectors I1.ke consumer electt Allies or military adventures. In short, this strategy steel. would insulate the Nation from the disarrays of a world econoiny that is simultaneously more open to ail corners and more susceptible to How can the United States hope to compete manipulation by organizations seeking ends under - :ch circumstances? On answer is to other than those of private co.orations in offer .,ducts that are beyond .ihe tec1,:no- American mold. logical capabilities of low-wage countries, In rro,:e conventional products, it maY-6c7-possi- If indu,:trial pi:acy is too strong a term to ble to improVe labor productivtly--ilncugh (!r;crilie foreign tactics, it is n.)netheless.true automation or other advanced manufacturing ti s3nptionalizun enterprises can with technologies enough to caset e7,:isting wage 7.mpunity goals q...ite different ciaerentials.\. Advocatv,,, of trada prote,;tion

1 z72 Con-7,7erit' eness In Electronics point out that these avenues may not guarantee trade protection as a road to recovery for sec- -flough lobs to keel) the American labor force tors that have lost_ competitiveness interna- employed. The alternatives are then to let tionally (which is not to say that protection \iges in the t"nited States drop. helping to might not serve other objectives, or be a maintain competitiveness across a broader necessary if not sufficient prelude to recovery). range of production. or to keep out imports Industries ancl/er their employees may claim from low-wage countries. Given the levels to that import penetration stems from unfair trade which y'. ages !night have to falldepend- practices, dubious _management decisions, ing on lice..' swiftly developing countries can adverse effects of Government regulations, or iniprove their own labor productivitiesthe other transient problems. If so, the argument first alternative isfar front acceptable. The runs. recovery is possible, given ime. The reali- plight of unemplmed auto and steelworkers in ty is generally more tangled. Co.plaints of on- California.,:s.vbere plants in both industries have fai:- competition or adverse regulations may be been shut dowh, illustrates the difficulty. Some Ivel!-founded but nonetheless only secondary ir.not..all of the workersmany of factors; decline may result more fundamentally whom had 'been making $12 to S15 an hour ex- from long-term trends .n the world economy clusive of fringe benefitscould find employ- i.e., shifting comparative advantage. Where this ment in Silicon Valley electronics firms. How- is thc case, trade protection will be ineffective ever, unskilled or semiskilled electronics work if temporary, costly if permanent. .pays in the range of Sti per hour: and even at When a good argument can be made that re- this level. Atari, for one, is moving some 1,700 vival is possiblethat longer trends favor jobs overseas. Given such a picture, trade pro- the United States or at least do not run too tei:lion and industrial self-sufficiency begin to strongly the other waythe question remains: seem tractive. How long will protection be necessary? Where_ %Vino is the other side of this scenariothe the Government has imposed or negotiated im- argument against either temporary or longer port quotas, these have typically been for 3- or term trade restrictions? First, import restraints 4-year periodswith renewals not nliheard of. almost always result in higher prices for Amer- Fixed periods are desirable so that domestic. ican consumers. Witness the automobile price 'es well as foreign producers face a relatively increases following; Japan's voluntary limits on predictable situation. Protection granted for an exports in 1981. Indeed, under such circum- indefinite period risks de facto permanency, stances 'rice increase::, are generally intended: decreasing incentives for domestic firms to the cominon rationale is that import-affecied make new investments or alter their business tS. firms must he temporarily shielded so that strategies. they can raise ,.prices, generating increaseo To illustrate some of the factors involved in profits to he invested in restoring their coin- decisions on protective mechanisms, consider -..!ss. Of course, rising prices often rip.: the situation in early 1982 as concern mounted pie through the cconomycausing inflationary over imports of t.+4K RAM (random access pressures. To the extent that protection for memory) chips. Japanese penetration of the domestic steelmakers has raised steel prices, marketrumThIg at about 70 percent costs gone up for automobiles, consumer was the out.:;olne of a complex of factors: rapid roads, bridges and buildings, military capaci'iy expansion ID Japan facilitated,by am- hardwi.ireeveryvvhere steel is used. ple supplies of canitek fcr investment; produc- Should these higher price:-; be considered the tion pro! '.:;ins at the plants of several prospec- necessary costs for reviving import. ;Meted in- tive U.S. suppliers; pHce7eutting by both Jap- dustries? Where there is goc ' reason to expect anese and American firms striving to build revival, the anoser might bc: yes. Unfortunate- ,i,arkc,, share (acce ipanied by accusations of ly, experi:,.nci.---e.g., in the Case of color tele-. (Imp, oft !,-,,('led at the Japanese). At a time pro,fides little ecicf1..:nce it) support of rimel merchant firms--

; Cr:. 12Federal Policies Affecting Electronics: Options for the UnitedStates 473 c> Texas Instruments and :Motorola were able about the Lne Texas Instrumentsan- to 1);'utiti: 64K RAMS in large quantities, as nouncec.' it ',:as transferring all of its 64K RAM opposed to six Japanese manufacturers. the production to Japan, Hitachi. Nippon Electric, first question froin the standpoint of com- and Fujitsu revealed plans to moveor sneed petitive dynamics became: How long would it up previous timetables for movingsome of be before other suppliers entered the their own 64K RAM assembly here.8 Onemo- arena? Given the learning and scale effects tive was to dampen trade frictions; despite the ,:haracteristic of RAM production, too greata absence of constraints or even formalcom- head start might be virtually impossible toover- plaints concerning RAM shipments, the color come. On the other hand, if American com- TV case appears to prefigure that in semicon- names carne in later but with superior designs, ductors. Is this an outcome that U.S. policy- would they he able to turn the tables? These makerswhose actions accelerated onshore in- are nontraditional kinds of questions for U.S.. vestments in consumer electronics7-should poli4:ymakers, indeed difficult for governments welcome? Certainly there are major differences anywhere to deal with: as emphasized in the between the two industries. Competition in previous chapter, the fast moving events char- television manufacture is cost-driven, with acteristic of high-technology industries do not technology playing a relatively minor role. In fit verN, comfortably into the existing frame- microelectronics, moving closer to markets is wOrk of international trade policy. But effec- one way a company can capitalize on its tech- ti::7 Government action depenth; on grasping nology to meet customer demands. Although such facets of competition. decisions by Japanese semiconductor manufac- To return to the question of the costs asso- turers were spurred by concern over trade, ciateed with trade protection, note first that they see many other advantages to theirpres- regir(liess of rationaleimport restrictions ence here. For instance, they car learn from Function as implicit subsidies for proteeter I in- American technical expertise more easilyone dustri':s and their employees. The cost'-are way is to hire American engineersif they pain by other sectors of the economy i.e., by have bases in this country, especially now that the public at large. Beyond direct costs in the U.S. companies are guarding their own tech- form of higher prices, a protected industry may nology more closely. Irithe same way, te,,hnol: he able to attract re,.ources such as capital ogy acquisition has been one of the motives be- away from other parts of the economy; in at- hind efforts by U.S. firms to set up R&D and tempting to help one industry, Government manufacturing facilities in Japan. policies can harm others. When foreign firms invest in U.S. plants, These are not the only indirect effects for they employ American workersunskilledas policymakers to worry about. Foreign compet- well as skilledalthough a substantial fraction itors often pursue inward investment as a way of value added tends to remainoverseas. But aroudirade barriersthe pattern in color tele- from the perspective (:f U.S. semiconductor vision, now also taking place in industries as firms, sales by Japanese-owned competitors- different as microelectronics and automobiles. regardless of where the products are manui'lc- In 'contrast, foreign investment in U.S. steel- turedrepresent a loss to the domestic in-' making capacity has been smallno doubt be- dustry. The numerous joint venture and tech- cause overseas investors do not see long-terr nology exchange agreements that U.S. and Jap- trends favoring the production of iron and steel anese electronics firms have entered into corn- here. Direct investment is partk.ularly attrac- pP,;ate matters further. With Hewltt-Packard tive where companies feel they have competi- getting RAM technology from Nation- tive advantages that can be exploited regardless al Semiconductor sharing with Oki, thecom- of location. American semiconductor and computer firm Amdahl joined to Fujitsu, easy na- puter manufacturers invest overseas in part tional distinctions vanish. Such trends are still' because' their technological advantages are Kanaboyashi. "64K Ram Chips ilants in U.S.." Wall easily transportable. Street Journal, Mar. 2, 1982, p. 35.

0-111 n - 71 470 tiO111tc1 tit Sp.11 ;i 7;:r,iP11-W cvtlip7) s:mt10.(1:)010 P.1081.1 SpA\ 1S01111p ,s:Taspaid pp:lupip(i ,nrituud lou pa-mit-m.1cl -.10.10 snit' tio :-Ipulf :Tutu 1pfiat) spooii pue sa:mAa:5. z961 SpAN imq sfql smotfoi e Ipp.i tit nut-h7f oiu })lo,; 1,)tmin Ilutpeal uoff atp 1 S uo pue, saDfAJas II "S:1 ()11,11 11.1P110(1 alp ti:rttim snid.ins spoor 'uoffifq all; i861 scflf:uas p.-,..:0111..clupl.1 i101111(p..1r-qp pill! salps pafulai LIS sn[cDos aqi astpuetp.r:,tu 6.)tofFup -ssa.fis JE tpt-1,,; afau,v, .tit Mifpoo:)xo 0(1 fe.. sxmlufugtuf Epos se imp uaamiaq on poppp ..,:u10 all at: utr.)-,...to:mod Alplett pq per!tii-I pUP wow as -.led ft41(;r11.,:i ditl;i11S sLRP .I0J :)!I5ati;01) -1-107-)0 safeis tLa.Na--UtKitlf .11,111:111 Impatim SJOlanptioap.ups JO 7,,,11(f ;),,,f11\ 1! a0.11-i Ell .)11;,f ('" 11 ;Hi 1);11:)0(1X0 -spapoici III polo 'spun-0 1-iiittyp O1 otp sannsgo-sapciotuainu asacp peo.uf scuotied .ti;ialotitiPol-tptut ato swop s.loN,PtuA:-mfoil titic Jattuni alum -)1uSa.180e am-rmd ol2uftpou Tunic] )uatur-olpe swap-told paleala all; surmsuawtp pinoo e iJeykj uoRorAoid-la 46aleils Act slitqs UI .1,.)u-ape.11 popfiaci Jo p )StpUF?ti.:iipW -ofjap 4TM 11;1(1E1 critIfta101 7)0(1u:is:)f) aAoqu aleada P -uni sulollud LLS uong Ui 12111-Z861. suopeDoisfp a.JP, anAas UP feutstipuf fultioutup i,u1(0,,pp ..:0; pue 2upsej-Suot saop )ou ueaw imp MOH sPOp 3110 Atippualoci papt11101 110 apP11 St10110:l1saJ uopaa/0.1(1 sr .91p isag JO uaAa e algetA -Spawa.1 P 'Jtis1)tw.;1) 131.1u11I asPq uatiAN atp LInGg futtf,.1 a:11 suopuu °Aug f-11.iPp 1100A1F1(1 'r1 puu utflaloi Sat.11SilpLit 'Al suosew 2utpu.rt .10.1 uou,t woJi Sr;c1.101111 ',1111(1 110A0 ,;.11!0(1(IPSip 1 1! pip Atii:Ui Aume -cud puu f.mumn/ in? ppomio A10,1.011 `0011p1SilliTur) Ir AIM! nq oFFssoci 01 tLsfuotpat t.rado tuariAs appli s;.priod o alp JO sri Ifarlle-apu.I1 SfOuplug puu tofm Auew -.10111i103 o1 °sot otim .C.1stiput 11fln0.1111 ua11:)alo.1(1 tit atp ftuaua`ii pinotis-saidwexa asnud pinoAl aleacmpu e aauuls 1OJ 'ase;) aq1 linso.1 Ker;fut SSPlatii00011 p A:mfad Lump imputu u6pDaloid 0111 -PAOW tumpowyi io asu:)-Aq-asua suofspop Lions) se atp pa)furi A.111.11-100 pat Jaffa(' letp o.utf Alau Lrt atil opup, (nom! tiffm .fpappap maul Jot uado ape.0 UI acp auoiclaAa 1e 1Spa1 tit alp uru -sfyl poxfut .srlucl.tuf sapisaptif PUP slu)1Jut11 -aq pina.v, ltiauaq Ja2uol nutty; aJatm feuofluu.lama (It ..I.rwituutt0 tie -uf 'Afirmf.rof suafpufsaJ U0 OPeJI SA101.1 ;Meg umia :Ramp ruiest10 paittapo p.fumof .'11!..,1,)so.1(1 pal 01 A.mleftelai :samseaw UI '.3111 I 's,01;13 .A:mfad pu.! aq1 :)tisottic,f sial.tuut s)soa uo sualyitutv-,:su imp :mu pammipplo:) of tooq alp 111(11: tp"i3nof atulsas Jo ay -uofssaniaa uo0rfitula21 paau 1UU aq 1p 1:m.1 Emu snofAcio of aAutf specluif UO -op aufnuaii ;unix Afpua.uf II11S 'itificittlaffu of ;nstio s.i sisaJoulf fellow-m.1:mq suopupaau 'cum JO Sp :mison; sloiluto Jot atuos sociAi 0a Artil suofssaatioaiiue; uo sIalndwol.) L11 -xa UI ValfAllaptin stifqs a,Ntlu.luciula;) Joj SUOIS5a0U0.9 110 :11.1011M PUO 01110Pr.0 Imp sfip anurap :o;M!1111',.,pu luulsoptif sapflad idtuapu Audi aq imp lamp `duotuu ,suaquu uo U o,J. otp tuaixo imp alp tuopfos A.10:\ .inisSa:mns 1:m1)wci-S(1-)anp1mi 'sfsurf 01 aa:ui Jr misotnop 10)1.1PUI Attafe.tis siciwapu puu-a:lltu .1.101.1,0 0111 'Sal.1101111.11...:1! stual)1 10 fu.lorms asl.J put autfaap .iucu um Ji fupfsnpuf paptraluf ,)1 amasauf SaaJOJ aptS1110 -110,, pflitio-, EWA\ 111J.10,%10(1 apsaump sla)t.ww urgioa, Oj alnAon.1 tiva.fls aJ aip 'pup foil Jo llt-Htlattip.:0A0f) AIotuflui 'scunstlaw atp aJtitto walsAs JO !VUOtiPti apP.11 se If 's! jsoi\: pop. 1.1f asagf UI -PSJO papadtuf ppm.) aq Tatialeam poZtti J0-.10(111 1! AlIPIP.I!S 11!Ll1 pplOhl -um ozfsutfil atp sr- uofffso( Ui 1utiopfpu.11 ))1.1ew s1 s1(')0i ma 1;1111!1 Sputaap.ted -;minutmu 'spo0 11:1111 'S1) 11!sill!.1.1,1V110!111ill:P111 'NUM; ti1it10..1 J0).111r1r) optui..s:uni tit cuouf--saafAaas IMP .11;1.1, 'Ns'alt!s//1/ 11:).11IN TA{(iL 'V, oaf too p(i 01l futloqutwol; smait ';1--tulfclu:)io ttaiv, :11/1.1/11 flPrItitftig ):Iftp.tl!,.+A(Cif inn Jo ';uno..y.m! optim aqi -s.1-1 opal! p0fiap uo s.1,1r.,.1:, alll 0:1qop Joj Cr). r2 Federal Pcicies At fectirg E_-"riectroncs: Options for the United Stated 475

Fear ui.etniia.iion has been n very real fa; nary needs. Examples include the st; eland ma- tor in the choices made by the United States chine tool industries. concerning trade in steel. particularly import from Europe. If the United States makes The assumption underlying a critical or stra- it ton difficult for European steel to enter the tegic industries alternative is that only a subset American market. it risks restrictions on U.S. of economyperhaps relatively smallis exports of electronics products, or f.inanciai indispensable for national defense; unless the services e'..en military goods. Trade wars sel- list is kept short, this approach would differ lit- dom benefit those involved so much as those tle from the first option discussed above- Grit- isIposition to pick up the spoils. i(al industries would begin with. but not bere- stricted to, the traditional defense sector: aerospace, suppliers of armaments, military elec- Support for Critical Industries tronics firms, R&D contractorsenterprises Rather than attempting to preserve domestic that, along with large numbers of suppliers and markets in generalwlyich would in large subc(..anractors. sell to the Department of De- liWitSlIre reward those abletobuildthe fense (DOD). Beyond this, other portions of the strongest political constituenciesthe Federal electronics industry would be obvious candi- Government could decide to support and if dates for any critical industries listcomputer necessary protect only those industries hidged hardware and software, integrated circuits, critical to national security. Security might be communications equipment. Indeed, numer- broadly or narrowly defined. Either way, such ous manufacturers of computer systems and a policy would find deep histor;eal roots. Gov- semiconductor products have divisions de- ernments support transportation technologies ited exc?usively to military sales. In contrast, and systemscanals, railroads, highways, avia- consuc..ar electronics, as a sector, would have tionin part for reasons of national mobiliza- a weak case despite the fact that firms like RCA tion and defense; other examples range from and GE are major defense contractors, just as armories ind shipbuilding to telecommunica- Chrysler's tank business was largely divorced tions regnH'ions and space exploration. A na- from the automobile side of the corporation, tional sem: ,:y criterionrestricted to military so electronics suppliers that engage in military security or extended to "economic security" production generally do so through separate would narroN the focus compared to the mar- divisions or subsidiaries. ket preservation alternative, helping to control the political pressures that will always bedevil efforts at industrial policy in a country like the What is Critical? Uoited States. The difficult questions in identifying "crit- Manufacturing sectors suffering from import ica" firms and industries imolve those that do (..:ompett ion frequently argue for Govern meat not engage directly in defense-related research, remedies on the basis that their productsor or production, but whose products or R&D their plant and equipmentcontribute to na- might still have vital military applications tional security. Some clearly have better cases under some circumstances. Synthetic fibers than others. The end products of some com- like nylon and Kevlar providean example. panies and some industries consist of military Used in clothing, parachutes, body armor, and hardware: armaments, communications sys- fiber-reinforced composite materials for struc- terns. In other cases. end products may be used tures ranging from missile casings to stealth only indirectly for national defense, though no aircraft, these materials are obviously critical loss critical for that. This is true of supercom- to the defense base. But would this .have been 'inters, needed in the design of some types of predicted when synthetic fiber technology was military systems. In still other cases, the goods in its infancy? That is the nexus of the problem; produced by an industry may be vital, but only if the Federal Government is to support critical some fraction of the industry's production ca- industries--identifying those that will he vital pacity would ever be consumed in meeting mil in the future. 4 47.5 E-lec:;rcrucs

the telterie are to extficti to cini)::iice tione-conteed that the U.S. Government should ect,11()11,1t; strength. then the leaner o!nlli- take a more active role in ensuring the well- fying critical industries--without being so in- being of strateg:e industries.16 A primary strand citee'.'t,that support and protection go to in the argument is that if the United States evorvmle who asksbeeomes still more per- comes to depend too heavily on foreign prod- plieke ee One reason is simply that terms like ucts or the -Nation's defensive -economic strength" are not very meaningful. could be impairednet only in the In practice, virtualft industry threatened event of war, but even in a rapidly escalating fore42,n competit ,ould attempt to de- arms race." clare itself cI itical. alit:tiny in recogniz- As table 82 indicated, a wide variety of policy ing iedustries that will he critical in the future instruments could be used to provide for the wouli.i arise here as wellwhere it is a variant continued strength of critical industries, going of the -sunnse industry" problem. Once the well beycnel tariffs or quotas for protecting do- sunnp, dad everyone knows it, Federal pol- mestic manufacturers and beyond the well - icy may not be especially important; oppor- established relationships that already link DOD tunities will be evident, investors will be at- and the community of rnilitt.ry contractors and tracted. Alt Iv nigh Government might be able suppliers. Multiyear procurements have been to nurture the growing industry, its role could suggested as a means to strengthen the defense well be peripheralmore so in an eco nomy like industrial base. DOD is also paying a good deal that of the United States than in japan or of attention to manufacturing technologies as France. But when an industry is truly an in one way of getting more for our money, aswell fent, its prospects for the future uncertain, then as short.:ning procurement cycles.The atten- Government may be no better able to recognize tion to manufacturing will have spillover ef- its potential than the private sector (sc:ee fects in the civilian economy that could be sig- would say the primary difference will be nificant. Beyond such steps, secbral policies that GovO'rnment's time horizons need not be could provide targeted supports and subsidies constrained by the desire for quick returns on in much the same way that the American farm- investments as for private suppliers of funds. er has been given special consideration. Pro- curement could be steered to particular firms. In any event. a strategic or critical industries DOD-sponsored R&D efforts like the VHSIC approach implies that the Federal Government program and the other research and engineer- can and should identify such industries, then ing activities of the services and the Defense adopt 7eolicius to: Advanced Research Projects Agencymany of I. Ensure that the United States eittintains an them related to electronicrmight be enlarged indigenous prodteltion capaeility suffi- and broadened still further, with the aim of cient to meet direct military needs. par- strengthe!!ing the U.S. technological hese and ticularly in the event of national mobiliza- infrastructure. Manpower and education pol- tion or crisi icies could channel institutional support to- 2. Support industries and technologies that ward engineering and relevant scion,!s, fund have a substantial role in providing the un- ,°For a detailed presentation of this view, see "Statement of derlying bte-.eeither in terms of R&D or Gen. Alton 0. Slay. Former Commander, Air Force Systems productionfol. U.S. military strength: Command,' Revitali7 tint: and the U.S. rr7onomy, hearings, Part Subc.ommittel, on Economic Stabilizat, in, Committee on Bai;It. 3. Optionally, support industries and tech- ing. Finance. and I Jrbari Affairs, House of Representatives, Feb. nologies that clearly and unambiguously 25; Mar. 25, 26, 1981, pp. 258-479. contribute to economic strength. "Stich arguments were athanced by opponenU of th-.: award of contracts to Fujitsu for a Boston-Washington fiber-optic communications link After intense lobbying by the 1)00 and others. In a context of growing East-WesttelU AT&T gave the contract to its own subsidiary, Western Elec- advocates of such an approachparticularly tric. See E. Meadows, "Japan Runs Into America, Inc.," For- those who emphasize, direct military produe- tune, Mar. 22. 1982, p. sta.

7, j CS. 12Federal Policies Affecting Electronics: Options for the United States4-, 477

SUIdell!Sminoring iii these reward peo- ij have to be et-awnin somecases fairly ar- ple who choose to work in defense industries, bitrarilybecause in the most generalsense provide incen :es for retrainilig and continu- nearly all industries and technologiescon- ing education in advanced technical subjects. tribute in some way to defense readiness. Cor- The services have recently argued that, in the porations may produce the boots that soldiers years ahead. they may be unable to meet re- wear, the food they eat, or small computers for quirements for skilled workerselectronics battlefield command and control. When only technicians. Lf'-ospace fabricators, aircraft portion of an industry's outputgoes to the maintenance sp- 'a listsas well as ,olgtneers; militarywhether the industry be steelor semi- a strategic industries policy would airto rec., conductorshow might the Government allo- tify such probhiins. cate its support? Critical Industries for National Defense The struggles of DOD with the "militarily critical technologies list" recommended bythe l'.:ations traditionally give special attention well-known Bucy report and endorsed byCon- to economic sectors on which military strength gress in the late 1970's shed light on theprac- and secur:tv depend. Shipyards and armories tical difficulties of a defense-centered industrial are obvious examples; for many years, histor- policy. The first list of 15 militarily critical ians .and economists have probed the sNmbiosis technology categories was published in1980.12 betw.-!en military and civilian production, ex- Included were computer networking, large emplified by the evolution of precision manu- computer systems, software, design andman- facturing and interchangeable parts. If 19th ufacture of very large-scale ICs, anda number century productioa technologies were driven of others related to electronicsof the,15 cat- in part by military nee.ds, certainly the relation- egori..1..i. only 3 had little or no electronicscon- ship between military and civilian sidPs of the tent. The thrust of the exercise was to develop economy has altered greatly since. Theprva- a systematic approach to export controls; asa siveness and complexity of modern technclogy result, it was narrowly focusedon military-ap- makes identification and support of strategic plications. Despite the well-definedpurpose industries more problematicships, arms, in essence to update ands :ipplant the Com- even missiles and panes, hardly exhaust the m( .ty Control Listprogress has been pain- fequiremeo.; of modern warfare. During fthly slow. Oncc 'he 15 generalareas had been World War II, automobile. plants could be re- determined, the eilort bogged down in'details. tooled to make weapons, but in the past four Critics doubt that it will ever be possibleto decades. military and: civilian technologies agree on procedure,for reducing the case-by- have diverged, As for commercial technologies, case reviews of export licenses that are now new demands and applications come in rapid necessary. If nothing else, the continuing sequencechemical and biological warfare, debate over militarily critical teihnologies terrain-following cruise missiles, surveillance which in principle seem relatively straightfor- satellites, war in space, cryptology, computer- ward to defineindicates how difficultit ized translation of foreign latuages. One need would be to devise criteria for entire industries. not stop here. Economic w.rfare, in various After all, these industries would be rewarded forms, has a long history. Wheat, cobalt and not with export licenses that might adda few chromium supplies, energyail can be weap- percent to reverr..:eshut in Lit least somecases ons. Ultimately, a nation's military potent;.al is with substantial subsidies and °filer Govern.. a function of the size and composition of its ment favors. of ,the fraction of gross national product f. Wing to spend on defense. '1Technology and East-West Trade: e tWalthingtoil, D.C.: Office of Technology Assep.sment, May 19:33),p. 37. A de- Sooner or later, then, any policy based on a fined critical technologies list published in cla.sified farmis critical industries approach will face -a series the end of 1581 ran to 800 pages. Also FOCTechnologyand East. West Trade (Washington, Office of Technology Aase&s- Of decisions on what is really essential. Lines :nerd, OTAISC-101, November 1475). pp. 92434.

4u 473 International Competitiveness an Electronics

Looking more narrowly at the electronics in- Such is the circus for which a critical industry. consider. again the situation created by dustries policy would provide the rings. De- imports of 64K RAMs from Japan and the respite the concern generated by 641:. RAM im- sultingflurryof activity inthe Federal ports, the underlyin. national security question bureaucracy. In December 1981, the Cabinet remainsunanswereda question which is in Council on Commerce anr:1 Trade authorized fact much broader than that of RAM chips or an interagency study of high-technology indus- semiconductors in general. One way to frame tries--carried out largely by the Department of the questionin the context of microelec- Commerce." Several months later, the Depart- tronicsis as follows. As technologies become ments of Defense and Commerce began their more complex and industries expand, oppor- joint examination of the national, security con- tunitieS for different countries to specialize in sequences of 64K RAM imports, considering certain kinds of products grow; Japan's semi- the advisibility of a more formal section 232 conductor manufacturers; at the moment, are proceeding.14 As pointed out in chapter 11, this specializing in RAMs. As a consequence, U.S. section of the Trade Expansion Act of 1962 em- production might decline, with the result that powers the President to restrict imports in the the Nation could find it difficult to meet future event .of harmful implications for national defense needs, particularly in a situation call- security; the remedies available include quotas ing for rapid mobilization.18 Again, the point or higher tariffs. At the time, domestic man- is that the ongoing dynamics of international ufacturers were accusing the Japanese of competition hold one of the keys to policy dumping 64K chipsin fact, industry lobbying choice. appeared responsible for much of the concern over national security.15 Simultaneously, DOD So long as questions such as these remain was trying to convince the same group of Jap- narrowly definedconcerned with particular anese firms to transfer some of their technol- products or with classes of technologyit ogy to the United States, as well as to produce should be possible for policymakers to agree components and equipment that would help on priorities and make the necessary choices. meet American, military needs. This was also In its recommendations for the fiscal 1984 the period when Texas Instruments was 'mov- defense budget, for example, the Defense Sci- ing its 64K RAM productieri. ';',:ipan and Jap- ence Board ranked the following technologies anese firms were announcing plans to make in order of importance for future U.S. military these parts in the United States. A little later, systems:17 the Justice Department announced its price-fix- 1: Very high-speed integrated circuits, ex- ing probe of Japanese importersinvestigating emplified by the DOD R&D program prices that might be too high instead of too low (VHSIC) mentioned elsewhere. (ch. 11). Meanwhile, Congress was flooded 2. Stealth aircraft. with trade reciprocity bills, some motivated by 3..Computer software. trade friction in semiconductors. 4. Microprocessor-based teaching aids. 5. Fail-safe and fault-tolerant design meth- "An Assessment of U.S. Competitiveness in High Technology Industries (Washington, D.C.: Department of Commerce. Feb- ods for electronic systems. ruary 1983). The first paragraph of the summary states: "This study is being released as a Department of Commerce document. "Part of the reason is simply that the military market does not The methodology. findings and concluSions do not necessarily attract that many manufacturers. The 20 percent of U.S. elec- represent the views of other Executive Branch agencies" (p. iii). tronics sales that go to the military are unevenly distributed; in "C. li. Farnsworth, "Japanese Chip Sales Studied," New York some product categories, defense needs account for only a small Times, Mar. 4. 1982, p. D1. Also see ch. 11. fraction of outputElectronics, Jan. 13, 1983. pp. 128.140: In "On lobbying efforts by the industry, see "Horror Story," Elec- semiconductors, the military market is perhaps 10 percent Of tronic News, Feb. 8, 1982. p, 12. One of the reasons nothing came the total (fig. 34, ch. 5), and heavily weighted toward less sophis- of the Section 232 study was simply that 64K RAMs new prod- ticated devices: during 1982, any 64K RAMs going to DOD would ucts in the marketplacehad not yet been incorporated into any have been embodied in commercially available hardward for use U.S., weapons systems. thus, the national security implica- in offices (.r laboratories. not weapons. tion's of a supply interruption remained matters of speculation "See R. Connolly, The Big 17 Future Technologies." Elec- concerning future weapons needs and designs. tronics. May 5, 1982, p. 98. C71. 12Federal Policies Affecting Fectronics: Options for the United States 479

17. Short-wavelength lasers. As many as a dozen of these are electronics technologies, or systems for which electronics is a vital element: Given some agreement on prioritiesof which lists such as that above might form one starting pointand recognizing that prioritieS would have to be reexamined and updated more or less continuously, what policy mews- , ures,beyond decisions on R&D funding levels, I might then be called for to ensure that military \needs were met? Almost certainly, such ques- t; ns would have to be approached much as for those dealing with research prioritiesi.e., on a case-by-case basis; given past experience in trying to define critical technologies for export control, formulating general criteria for an industrial policy based on national security would seem a hopeless task, one con-f- pounded by uncertainties surrounding mool;i- zation scenarios... Furthermore, quite apart from debates over the needs of high-technology sectors like electronics versus bas::: industries like steel or machine tools, industrial policy that set defense priorities consistently above civilian needs would be politically painful. Like all sectorally based policies, Photo credit: GCA Corp such an approach is susceptible to the criticism Directstep-onwafer systein for lithographic that other industriesand economic welfare fabrication of integrated circuits as a wholewould suffer relative to sectors chosen for support. 6. Rapidly solidified materials--e.g., amor- One of the underlying questionsfor this and phous metals with high strength and re- other industrial policy alternativesbecomes: sistance to corrosion. Given the policymaking environment in the 7. Computer programs for artificial intelli- United States, would this framework contrib- gence. ute to good decisions at the level of individual 8. Supercomputers for nuclear weapons de- policy instruments, or would it simply curduse sign and computational fluid dynamics. matters further? The Nation's policymaking 9. Composite materials. system is not likely to change very quickly or 10. High-density focal-plane arrays for in- very dramatically. As a result, one of the pri- frared imaging. mary objectives of a more focused industrial 11. Radiation-hardening techniques for elec- policy for the United States can be viewed tronic systems. simply as a movement of the system toward 12. Space nuclear powerplants. better decisionmaking on the average. From 13. High-power microwave generators. such a perspective, it would seem more desir- 14. Technologies for erecting large structures able to regard national securityparticularly in space. direct military procurement and production 15. Optoelectronics, as one factor to be weighed when making in- 16. Space-based radar. dustrial policy decisions, but not the center- 482 480 Internationa! Competitiveness in Electro -CS piece. Where military security is to "pick winno7...,."' is c;.AaS5,:i4:YE!.d, second stake, DOD and the defense community ge,1 ,c.5r- and third poin'