INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
University Microfilms International A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor, Ml 48106-1346 USA 313/761-4700 800/521-0600 Order Number 9227231
The effect of environmental factors on the multinational enterprise’s use of external markets
Blaine, Michael James, Ph.D.
The Ohio State University, 1992
Copyright ©1992 by Blaine, Michael James. All rights reserved.
UMI 300 N. ZeebRd. Ann Arbor, MI 48106 THE EFFECT OF ENVIRONMENTAL FACTORS ON THE
MULTINATIONAL ENTERPRISE'S USE OF EXTERNAL MARKETS
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate
School of the Ohio State University
By
Michael Jam es Blaine, B.S., M.A., M.B.A., M.F.A.
The Ohio State University
1992
Dissertation Committee: Approved by
Riad Ajami
Richard Baldwin
Sven Lundstedt Administration, Graduate Program Copyright by Michael James Blaine 1992 ACKNOWLEDGEMENTS
I would like to express my sincere gratitude to several individuals, without whose assistance this project would have been infinitely more difficult. First and foremost, I would like to thank Dr. Riad Ajami and Dr. Sven Lundstedt for their support and guidance throughout all phases of this research. The opportunity to work with professors of their stature has made the writing of this dissertation a far more rewarding experience. In addition, I would like to thank Dr. Richard
Baldwin for serving on my advisory committee. Thanks also goes to Raymond
Mataloni Jr., of the U.S. Department of Commerce, for his assistance in accessing the appropriate data on the operations of U.S. MNEs; Dr. Edward
Roche, of Seton Hall University, for his assistance in developing the indicator used to evaluate various Political Systems; and Stephen Preece, a fellow doctoral student, for his comments on an earlier draft of this dissertation. Last but not least, I would like to thank Molly Dodd, Grade Mansion, Mrs. Puff, and
Brenda Walsh for providing a welcome diversion from the many problems associated with the completion of this project. VITA
Sept. 13,1952...... Born - Columbus, Ohio
1974 ...... B.S.; Antioch College; Yellow Springs, Ohio
1978 ...... M.F.A.; Florida State University; Tallahassee, Florida
1986 ...... M.B.A.; New York University; New York, New York
1989-Present ...... Teaching Assistant; Ohio State University; Columbus, Ohio
PUBLICATIONS
Blaine, M.; "Markets Reconsidered: The Role of Free Market Structures in Economic Development”; Proceedings: 27th Annual Meeting. Midwest Business Administration Association: MBAA; 1991
Blaine, M.; “Changing Dynamics in the Global Economy: The Effects of European Integration on the Pacific Rim"; Proceedings: International Symposium on Pacific Asian Business: University of Hawaii; 1991
Blaine, M.; “A Development Model for the Poorest Nations of Southeast Asia”; Proceedings: Business in Southeast Asia. A Teaching and Research Conference: Ann Arbor; University of Michigan; 1990
Ajami, R., & Blaine, M.; “America’s Ideological Quagmire in the Middle East"; Ohio State University College of Business, Working Paper Series; WPS 90-94; 1990
Ajami, R., & Blaine, M.; “Beyond Ideology: Economic Realism in the Emerging Global Economy"; Ohio State University College of Business, Working Papers Series; WPS 90-76; 1990 iv FIELDS OF STUDY
Major Field: Business Administration
Specialization: International Business Strategic Management Public Policy
v TABLE OF CONTENTS
ACKNOWLEDGEMENTS...... iii
VITA...... iv
LIST OF TABLES ...... x
LIST OF FIGURES...... xi
LIST OF ABBREVIATIONS...... xii CHAPTER...... PAGE
I. INTRODUCTION...... 1
Overview...... 1 Background ...... 3 I. Theory of the Firm...... 7 II. Organizational Structure ...... 5 III. The Organization and the Environment...... 9 IV. The Multinational Enterprise...... 11 V. MNE Structure ...... 14 VI. The MNE and the Environment...... 15 VII. The Changing International Environment...... 17 VIII. Organizational Responses...... 19 IX. "The Coming of the New Organization”...... 22 Statement of the Problem...... 24 Objectives ...... 25 Scope of the R esearch...... 26 Methodology...... 27 Research Hypotheses...... 29 I. Introduction ...... 29 II. Independent Variables ...... 29 III. Dependent Variables ...... 31 IV. Research Hypotheses...... 31 Limitations of the Study...... 32 Contribution of the Study ...... 35
vi Organization of the Dissertation...... 36 N otes...... 37
II. REVIEW OF LITERATURE...... 40
Introduction...... 40 MNE Theory...... 41 I. Introduction...... 41 II. The Hymer-Kindleberger Tradition ...... 42 III. The Product Life Cycle Hypotheses...... 45 IV. Internalization Theory...... 47 V. Diversification and Internationalization...... 50 VI. Location Theory...... 51 VII. General Theories of the MNE...... 52 VIII. Transaction Cost Analysis...... 57 IX. Economies of Scale...... 60 X. Another Classification Scheme...... 62 XI. Conclusions ...... : 63 Organizational Theory...... 67 I. Introduction ...... 67 II. The Stopford and Wells Model...... 70 III. Other Strategy and Structure Studies ...... 75 IV. Contingency Theory...... 77 V. Other Contingency Frameworks...... 86 VI. Other Organizational Theories...... 88 VII. Conclusions...... 88 Strategic Management...... 92 I. Introduction...... 92 II. Cooperation and National Competitiveness...... 93 III. Inter-Firm Cooperation...... 98 IV. Contractual Versus Direct Control...... 102 V. Joint Ventures, Licensing, and Strategic Alliances 107 VI. A New Model of the Firm...... 113 VII. Conclusions...... 116 Summary ...... 120 N otes...... 123
III. METHODOLOGY...... 126
Introduction...... 126 Regression and Correlation Analysis...... 127 I. Introduction...... 127 II. Correlation...... 129 vii III. Regression...... 130 Some Previous Studies ...... 135 Some Methodological Concerns...... 139 The Research Question...... 143 The Current Study ...... 144 I. Overview...... 144 II. Research Design...... 145 III. Subject Selection ...... 145 IV. Data Collection...... 148 V. Data Analysis...... 150 VI. Threats to Validity...... 152 Independent Variables ...... 153 I. Introduction...... 153 II. Economic Variables...... 155 lla PCGNP (Per capita GNP)...... 157 lib. GDPGRO (Average annual GDP growth rate) 157 lie. TOTRADE(% total trade to GDP)...... 158 lid. GOVEXP (% government expenditures to GNP).... 159 lie. INVEST (% gross domestic investment to G D P ) 160 III. Social Variables...... 160 Ilia PQLI (Physical quality of life index) ...... 161 lllb. LIT (Literacy rate) ...... 162 lllc. DOCTOR (1000 people per doctor)...... 163 Hid. %AGR (% of population in agricultural sector) 163 IV. Political Variables...... 164 IVa. POL (Political system )...... 164 IVb. REST (Restrictions on foreign business and FDI)... 168 IVc. PRI (Political risk index) ...... 169 Dependent Variables ...... 171 I. Introduction...... 171 II. %INTFIRM...... 172 III. %AFSALES...... 173 IV. %MOFA...... 174 V. %WORK...... 174 Research Hypotheses...... 175 N otes...... 176
IV. DATA ANALYSIS AND INTERPRETATION...... 179
Overview...... 179 Comparison of the External Activities of U.S. MNEs in 1966 an 1989 ...... 180
viii Correlation Analysis...... 186 I. Introduction...... 186 II. “Average" Country Profile...... 192 III. Correlation Between the Study’s Independent Variables ...... 193 IV. Correlation Between the Independent and Dependent Variables ...... 199 Regression Analysis...... 207 I. introduction...... 207 II. %INTFIRM...... 210 III. %AFSALES...... 213 IV. %MOFA...... 216 V. %WORK...... 218 VI. Conclusions...... 221 Summary of the Study’s Major Findings ...... 223 N otes...... 226
V. CONCLUSIONS...... 228
Introduction...... 228 Discussion of the Empirical Study ...... 231 I. Introduction...... 231 II. Some Possible Explanations...... 233 III. The Impact of Environmental Factors...... 239 IV. Public Policy Implications...... 240 Implications for the Study’s Basic Thesis...... 241
LIST OF REFERENCES...... 245
ix LIST OF TABLES
TABLE PAGE
1. Hypothesized Effect of the Study’s Independent Variables 33
2. The Study’s 52 Subject Countries ...... 147
3. Foreign Affiliates of U.S. MNEs (1966,1989)...... 181
4. Data Table ...... 187
5. Descriptive Statistics for the Study’s I Vs and DVs ...... 191
6. Correlation Matrix...... 194
7. Hypothesized (HYP) and Empirical Directions of the Relationships Between IVs and DVs...... 200
8. Step-wise Regression of %INTFIRM on all 12 IVs...... 2T1
9 Regression of %INTFIRM on REST, TOTRADE, DOCTOR 211
10. Step-Wise Regression of %AFSALES on all 12 IVs...... 214
11. Regression of %AFSALES on TOTRADE, PRI, GDPGRO, DOCTOR...... 214
12. Step-wise Regression of %MOFA on all 12 IVs...... 217
13. Regression of %MOFA on REST, PRI, PCGNP, DOCTOR 217
14. Step-wise Regression of %WORK on all 12 IVs...... 219
15. Repression of %WORK on GDPGRO. TOTRADE. PRI...... 219
x LIST OF FIGURES
FIGURES PAGE
1. The Unitary and Multi-divisional Organizational Structures 8
xi LIST OF ABBREVIATIONS
DOCTOR...... An Independent Vatiable
DV(s)...... Dependent Variable(s)
FDI...... Foreign Direct Investment
GDPGRO...... An Independent Variable
GOVEXP...... An Independent Variable
GSP ...... Global Strategic Partnership
ID...... International Division
INVEST...... An Independent Variable
IV(s)...... Independent Variable(s)
JV ...... Joint Venture
LIT...... An Independent Variable
MNE...... Multinational Enterprise
MOFA...... Majority-Owned Foreign Affliiate
NFI...... New Forms of Investment
PCGNP...... An Independent Variable
POL...... An Independent Variable
PQ U ...... An Independent Variable
PRI...... An Independent Variable
REST...... An Independent Variable
TCA...... Transaction Cost Analysis
xii TOTRADE...... An Independent Variable
SA ...... Strategic Alliance
VAP...... Value-Adding Partnership
WF...... Worldwide Functional
WOS...... Wholly Owned Subsidiary
W P ...... Worldwide Product
%AFSALES...... A Dependent Variable
%AGR...... An Independent Variable
%INTFIRM...... A Dependent Variable
%MOFA...... A Dependent Variable
%WORK...... A Dependent Variable
xiii CHAPTER I
INTRODUCTION
Overview This study attempts to address a fairly simple proposition; namely, that recent changes in the international economic environment have increased the efficiency of international markets and should, ceteris paribus, lead to corresponding changes in the organizational structure of multinational enterprises (MNEs). Specifically, it will be argued that as the cost of using international markets declines, MNEs are encouraged to shift a greater portion of their activities from internal to external markets. This shift will promote,inter alia, the “externalization" of certain foreign operations and a corresponding move from equity to contractual control of the firm’s value-added chain.1 At this point it is unnecessary to specify the structural changes that might accompany these events; but in any case, it is clear that some modification of MNE structure will be inevitable.
Unfortunately, while this proposition is straightforward, its conceptual basis is not. In order to provide the logical antecedents for this argument, it will be necessary to examine three very large and diverse bodies of literature
(Organizational Theory, the Theory of the Multinational, and International
Strategic Management) and integrate them in a manner which few writers have had the interest or inclination to do. In addition, while a number of authors have
1 2 discussed the “coming of the new organization”,2 few have attempted to develop the theoretical support necessary to associate specific organizational innovations with their corresponding environmental causes.
Therefore, the purpose of this brief section is to present, in the simplest manner possible, the logical structure of the argument that will be developed in the remainder of this text. Toward this end, the following statements may be viewed as a conceptual “map” which, it is hoped, will greatly facilitate the reader’s understanding of this study.
1) Markets and Firms represent alternative means of organizing economic activities.
2) The choice between these two alternatives depends on the cost of conducting transactions in the market on the one hand, and the cost of organizing them within the firm on the other.
3) To a large extent, these costs are a function of ever-changing environmental conditions which affect the efficiency of both external (ie. the market) and internal (ie. the firm) markets.
4) Firms can and do adjust their organizational structures to reflect changing environmental conditions; and, to a large extent, the fit between a firm’s structure and the environment determines its performance.
5) Since a firm’s performance ultimately determines its long term survival, firms can be expected to adopt certain structural changes in response to major environmental changes.
6) The MNE has emerged as a primary means of organizing international economic activities.
7) This has occurred in response to a specific set of conditions in the international environment which favored internal over external markets.
8) Recent changes in the international economic environment have dramatically altered the costs and benefits of using the MNE to coordinate international activities. 3 9) As a result, MNEs may begin to shift some of the activities currently conducted within the firm to external markets.
10) As MNEs conduct more of their activities in external markets, the rationale for controlling a large, international network of subsidiaries through equity ownership will be replaced by a need to coordinate an international network of contractual relationships along the firm’s value-added chain.
11) This shift from equity to contractual control will have a significant effect on the organizational structure of the MNE.
While the theoretical basis for these statements will be presented in detail in Chapter II, the following section provides a further development of these ideas.
Background
I. Theory of the Firm
In his 1937 article ‘The Nature of the Firm”, R. H. Coase made a powerful observation. Attempting to explain why the firm emerges in a specialized exchange economy, he concluded that “The main reason...would seem to be that there is a cost to using the price mechanism.” (Coase, 1937, pg. 390).
Traditional economic analysis suggests that prices serve as market signals which promote the efficient allocation of resources; however, this occurs only if prices reflect all relevant information and are costlessly available to all market participants. When these restrictions are violated, traditional theory has little to say about the possible outcome.
To Coase, the identification of a set of costs associated with market- mediated exchange provided a compelling reason for the creation of the firm.
He noted that the cost of discovering relevant prices, costs associated with contracting, and costs due to uncertainty all arise when transactions are 4 conducted in the marketplace; and argued that, in some cases, they could be great enough to justify the "supersession of the price mechanism” (1937, pg.
789) and the coordination of economic activities by the firm instead.3 He also reasoned that, given these costs, Ha firm will tend to expand until the costs of organizing an extra transaction within the firm becomes equal to the costs of carrying out the same transaction by means of an exchange in the open market...” (1937, pg. 394). While some have characterized Coase’s analysis as
“tautological”,4 Alchian and Demsetz (1972) note that “[its] penetrating insight is to make more of the fact that markets do not operate costlessly...” (1972, pg.
783). But perhaps even more important is Coase's persuasive assertion that the
Market and the Firm represent alternative means of organizing economic activities.
This latter point is explored in detail in Williamson’s classic study Markets and Hierarchies: Analysis and Antitrust Implications (Williamson. 1975). Unlike
Coase, Williamson was interested in explaining the firm’s behavior under various market conditions rather than its raison d'etre. Using the transaction as his basic unit of analysis,5 Williamson develops the “organizational failures framework” to evaluate the relative efficiency of Markets and Hierarchies as competing methods of governing economic activities. He argues that the interaction of a specific group of “transaction costs” (i.e. bounded rationality, opportunism, uncertainty and small numbers of market participants) often creates serious market failures which encourage the substitution of the market mechanism with the hierarchical structure of the firm. Williamson’s basic argument can be summarized as follows:
“(1) Markets and firms are alternative instruments for completing a related set of transactions; (2) whether a set of transactions ought to be executed across 5 markets or within a firm depends on the relative efficiency of each mode; (3) the costs of writing and executing complex contracts across a marketvary with the characteristics of the human decision makers who are involved with the transaction on the one hand, and the objective properties of the market on the other, and (4) although the human and environmental factors that impede exchanges between firms (across a market) manifest themselves somewhat differently within the firm, the same set of factors apply to both.” (1975, pg. 8)
In terms of the present study, the insight that the relative efficiency of the market and the firm depends on a specific set of human and environmental
“transaction costs” is of paramount importance; but equally important is
Williamson’s analysis of the effects these factors have on the organizational structure of the firm. A significant portion of Markets and Hierarchies is devoted to this topic, and Williamson concludes that “Transaction costs thus explain both the decision to shift transactions from the market into the firm and, within the firm, what organization form will be chosen.” (1975, pg. 84).
II. Organizational Structure
Williamson, however, was not the first writer to investigate the relationship between organizational structure and environmental conditions;
Alfred Chandler’s historical analysis of American business had long since established a tie between the environment, a firm’s strategy, and its corresponding structure (Chandler, 1962, 1968). Chandler noted that the archetypical structure of the “modern” industrial corporation was the direct outgrowth of the rapidly changing economic environment of the late 19th
Century. According to Chandler, the creation of a national system of railroads and canals between 1830-1870 was responsible for a revolution in management practices that “moved business activity away from organizations run by entrepreneurs...to corporations with a systematic, bureaucratic management.” (1968, pg. 231). Since existing administrative techniques were 6 unable to contend with the size and complexity of such mammoth organizations,
the growth of the railroads fostered a number of structural innovations. These
included the functional department, the geograraphic division, and the holding
company; but the railroads also precipitated the separation of long range,
strategic activities from the day to day operations of the firm, thereby
establishing management as a separate activity which demanded the full
attention of specialized personnel.6
In his classic work Strategy and Structure (Chandler, 1962), Chandler
chronicles the rise of the large, vertically integrated, industrial corporations of
the early 20th Century and examines the environmental conditions which led to
their characteristic “multi-divisional" form (M-form). While the structural
innovations described above greatly enlarged the firm’s “span of control”, technological advances and the drive toward vertical integration in many
industries pushed the firm’s size and complexity beyond the limits of these
earlier designs. By clearly separating the strategic and planning functions from the firm’s day to day operations, the M-form allowed greater decentralization of decision-making and enabled firms to coordinate and control vastly expanded
networks of vertically and/or horizontally integrated activities.
Williamson also examined the evolution of the M-form, but from a much different perspective. Beginning with the premise that failures in certain key
markets and the cost of writing and enforcing contracts greatly favored
hierarchical governance structures, Williamson viewed integration as a natural
response to market imperfections. But while firms clearly had an incentive to grow through vertical and horizontal integration, they soon reached the structural limits of existing management techniques, and began to incur 7 substantial "transactional diseconomies”.7 According to Williamson, the multi divisional structure provided a means of reducing these internal transaction costs by economizing on bounded rationality, and allowing firms to effectively gather and process the huge amounts of information required to coordinate and control their operations.
In Markets and Hierarchies. Williamson examines 3 common organizational structures (see Figrure 1): the Unitary (U) form (a basic hierarchical structure in which all decisions and planning are made by the Chief
Executive Officer); the Holding company (a diversified organization containing a number of relatively autonomous business units); and the Multi-divisional (M) form (a diversified organization containing related businesses coordinated by top management to achieve economies of scale and scope). He then advances the uM-form hypothesis” which suggests that the performance of M-form organizations will exceed the performance of either U-form organizations or
Holding companies.
Like Chandler, Williamson argued that the division of tasks inherent in the M-form allowed top management to engage in strategic planning and the coordination of related activities within the organization in a much more efficient manner than other organizational designs. As a result, companies that adopt the multi-divisional structure should achieve superior performance. Specifically, this occurs because the continuous flow of information to the top of the U-form organization exceeds bounds on rationality and leads to sub-optimal decisionmaking; while the Holding company’s "portfolio” approach to the financial coordination and control of unrelated business units eliminates the possibility of exploiting economies of scale and scope. chlel Executive
I 1 . Sales | Finance | Engineering | Manufacturing |
The Unitary (U) Form
General Office Staff I
Operating Operating Operating Division A Division B Division C
Sales Finance Engineering Manufacturing
The Multl-dlvlslonal (M) Form
F1GUBE-1 The Unitary and Multi-divisional Organizational Structures 9 An empirical test of this hypothesis was conducted by Armour and Teece
(1978) using firms in the oil industry. They concluded that firms employing the multi-divisional structure did indeed exhibit superior performance; but that performance differentials declined over time as other firms in the industry adopted this structure. Further, their results imply "that the efficacy of internal exchange is a function of organizational form (and hence that the appropriate division of economic activity between firms and markets is a function of organizational form)...” (1978, pg. 118). ill. The Organization and the Environment
Two major conclusions can be drawn from the discussion thusfar; first, that environmental factors can and do lead to corresponding changes in organizational structure; and second, that there is a relationship between performance, organizational structure, and the environment. Initially, it was assumed that a single structure or management technique could be developed that would maximize organizational performance in every situation. This was the objective of the “classical” theories of Fayol (1916), Weber (1947), and
Taylor (1911). In each case, the author presented a unique prescription for achieving organizational effectiveness,8 but it soon became clear that no one approach to structure or management was appropriate for every environmental setting.
Joan Woodward’s research at South East Essex College of Technology in the iate 1950s was perhaps the first work to directly challenge the classical approach by showing that firms with similar structures exhibited widely varying performances. Discussing her pivotal 1958 study of British manufacturing firms,
Woodward (1965) noted that, “The fact that organizational characteristics, technology, and success were linked together...suggested that not only was the 10 system of production an important variable in the determination of
organizational structure, but also that one particular form of organization was
most appropriate for each system of production.” (1965, pg. 70-71). By demonstrating that a distinct, optimal structure was associated with each production process, Woodward showed that the relationship between structure and performance was far more complex than earlier theories had indicated. In fact, studies by Woodward (1958), Burns and Stalker (1961), Chandler (1962),
Thompson (1967), Lawrence and Lorsch (1967) and others strongly suggested that the ideal organizational structure was “contingent” upon specific environmental factors.
Lawrence and Lorsch were the first to advocate a “contingency” approach to the study of organizations in their book Organizations and
Environments (1967). In general, contingency theories argue that performance is directly related to the conformity between an organization’s structure and its environment. Unlike the earlier classical theories, which attempted to identify a single, ultimate organizational structure, the contingency approach strove to isolate key environmental variables which in turn defined a ssl of appropriate organizational structures. Although researchers have examined a wide range of environmental characteristics,9 in most cases their conclusions are the same; that performance can be maximized by adopting the organizational structure best suited for the specific environment facing the firm. While the logic of this approach is compelling, the overwhelming complexity of the environment has made it difficult to identify relevant environmental variables, let alone correlate them with organizational structures. Consequently, contingency theories have been criticized for a number of reasons. 11 Lawrence addresses some of these complaints in a chapter in Van de
Van and Joyce’s recent book Perspectives on Organizational Design and
Behavior (1981). He notes that contingency theory has been criticized for being static and not dynamic; for not clearly addressing the problems of environmental uncertainty, the technological imperative or size; and for encouraging a never- ending search for contingency variables. In response, Lawrence proposes an
“adaptive” model of organizational structure which draws heavily on the biological concepts of evolution and natural selection.10 He argues that
“...existing organizations are the survivors that can be thought of as being
“selected for” by having an initial set of resources and organizational features that did not exist among the organizational casualties.” (1981, pg. 320). By underscoring the dynamic nature of the relationship between organizational structure and the environment, Lawrence makes the important point that organizations must continually adapt to their environment in order to survive.
IV. The Multinational Enterprise
Up until this point, the discussion has been limited to general organizational theories without regard for their international implications. There are 2 reasons for this: first, the vast majority of the work examined thusfar does not specifically address international issues; and second, once a general theory has been developed, it is often a simple matter to incorporate an international dimension into it. In fact, many of the basic concepts in International Business are extensions or embellishments of general theories from other disciplines.
This point will become immediately apparent in the brief discussion of the multinational enterprise (MNE)11 which follows. 12 It will come as little surprise that the multinational enterprise has emerged as a major vehicle for conducting international economic activities, particularly since the end of the Second World War. Salvatore (1987) notes that
MNEs currently account for over 20% of total World output (and an even greater portion of the output of some developing countries), while intra-firm trade12 accounts for over 25% of the World’s trade in manufactures (1987, pg. 306). In addition, MNEs control a sizable portion of the imports and exports of most industrial nations. A recent study by Helliener (1981) estimated that 48% of all
1977 U.S. imports originated in a foreign unit of a U.S. company (one in which
5% or more of the voting stock is owned) (1981, pg. 10); while Amine (1987) notes that the 9 largest Japanese trading companies (Sogo Shosha) have been responsible for 55-60% of Japan’s imports in recent years (1987, pg. 212).
Consequently, the study of the origins and operations of the MNE has become a major focus of International Business research.13 In general, it has been agrued that the MNE evolved in response to specific conditions in the international environment which encouraged firms to conduct transactions through internal rather than external markets. According to Chandler (1986), many of the same forces responsible for the development and diffusion of the multi-divisional structure also account for the rise of the MNE. Specifically, he cites the increasing scale of production resulting from new technologies and mass production which encouraged firms to expand vertically to acquire resources and/or distribution channels, and horizontally to increase their access to distant markets. As this search for resources and markets extended beyond national borders, the result was the multinational enterprise. 13 Stephen Hymer presents a somewhat different explanation of the MNE in his 1960 doctoral dissertation The International Operations of National Firms.
Hymer argued that in order for a foreign firm to overcome the innate advantages local firms derive from their knowledge of local conditions, it must possess certain firm-specific advantages. By exploiting these advantages through foreign direct investment rather than through arms-length transactions in the marketplace, the firm is able to “internalize” or supercede the market, and thereby earn monopoly rents. Thus, according to Hymer, the multinational evolved as a means of protecting and exploiting a firm’s unique capabilities in one, or several, foreign markets.
Buckley and Casson offer a similar argument in their 1976 book Ib&
Future of the Multinational. Trying to explain the phenomena of international production and vertical integration, they develop the concept of “internalization”.
Simply stated, internalization contends that in a world of imperfect competition, the markets for knowledge and intermediate products are subject to a number of failures. These failures create an incentive for firms to replace external markets with internal ones; and once a firm internalizes a market across national borders it becomes an MNE. The authors suggest, “that prior to the Second
World War multinationality was a by-product of the internalisation of intermediate-product m arkets in multistage production processes, [while] post war it is a by-product of the internalisation of markets in knowledge.” (1976, pg.
59).
Buckley and Casson also argue that a firm’s decision to internalize a failing market involves 4 factors: industry-specific factors, such as the nature of the product and the structure of external markets; region-specific factors, such as the “social distance” between the home and host countries; nation-specific 14 factors, such as the political relations between home and host countries; and firm-specific factors, such as management expertise (1976, pg. 45). Dunning
incorporates many of these ideas into his “Eclectic Paradigm” which also focuses on the phenomenon of international production.
In one of the many versions of the eclectic paradigm, Dunning (1979) argues that a firm will engage in foreign production (and hence become an
MNE) if 3 conditions are satisfied:
“(1) It possesses net ownership advantages vis a vis firms of other nationalities in serving particular markets... (2) ...it must be more beneficial...to internalize its advantages through an extension of its own activities rather than externalize them through licensing and similar contracts with independent firms. (3) ...it must be profitable for the enterprise to utilize these advantages in conjunction with at least some factor inputs (including natural resources) outside its home country..." (1979, pg. 275)
V. MNE Structure
Thus, the MNE is the result of blatent failures in certain key markets; the presence of location-specific advantages such as natural resources, labor, and
market size; and the existence of firm-specific advantages such as proprietary technology, unique management and/or production techniques, and the ability to segment markets through advertising and product offerings. In response to these unique environmental conditions, the MNE typically develops an international network of subsidiaries which allows it to engage in a wide variety of foreign activities. According to Robinson (1981), the MNE is characterized by its “equity-based control of a globally integrated...production and marketing system, with control lodged in a headquarters corporation essentially managed and owned by the nationals of one country and domiciled in that country.”
(1981, pg. 15). Robinson also notes that the structural properties of MNEs 15 domiciled in different countries are virtually identical.
One of the first studies to extensively investigate the structure of the MNE was Stopford and Wells' book Managing the Multinational Enterprise (1972).
Using a large survey of US MNEs, the authors describe a 3 phase process of multinationalization. They note:
"...firms generally go through an initial [phase] when all their new foreign subsidiaries, in the manner of a holding company, are tied to the parent firm by loose financial links...The second phase is a period of organizational consolidation when an international division is developed. The international division is typically considered an independent part of the enterprise and not subject to the same strategic planning that guides the domestic activities. In the third phase, strategic planning is carried out on a consistent and worldwide basis and the structure of the foreign activities is altered to provide closer links with the rest of the structure.” (1972, pg. 19)
Although Stopford and Wells’ approach has been criticized as only representing the evolution of U.S. MNEs at a specific point in time, it is interesting to note that this 3 phase pattern was observed in the vast majority of the firms they studied. In addition, Dymsza (1984) has traced the spread of the multinational enterprise first to Western Europe, then Japan, and finally to the
Third World, noting that foreign corporations often follow the same pattern of multinationalization described above. In fact, Kobayashi (1988) recently developed a 5 step process of internationalization based on his study of firms in the international auto industry. Although Kobayashi’s model is not as structurally explicit as Stopford and Wells’ model, the similarities between these two approaches are striking.
VI. The MNE and the Environment
One implication of this finding is that firms face similar challenges as they expand internationally, and adopt similar structural and managerial solutions to 16 address these problems. Thus, in much the same way that vertical integration
led to the multi-divisional structure, multinationalization encouraged the
construction of vast international networks of subsidiaries, that are usually
wholly owned, and often controlled and coordinated from a central
headquarters. But while this structure may have been appropriate for the
international environment of the 1950s and 1960s, by the early 1970s a
fundamental dilemma had begun to appear; namely, how to coordinate the
firm’s international operations in order to simultaneously achieve the benefits of
both global integration and local responsiveness.14
One of the first structural solutions to this paradox was the umatrixn or
“grid".15 First employed by the Dow Chemical company in the late 1960s, the
matrix was an attempt to achieve both the benefits of coordinating the firm’s international activities a n d the flexibility of adapting to local market conditions.
Unfortunately, as Davis and Lawrence (1977) note, the matrix can be an extremely difficult organizational structure to manage effectively; and, as a
result, the search for a solution to the complex problem of managing the
modem multinational has continued to receive a great deal of attention.
But an even more fundamental dilemma may now face the MNE: how to deal with recent changes in the international environment. If the multinational and its characteristic network of international subsidiaries is in fact associated with a specific set of environmental and organizational factors, what happens when these conditions change dramatically? Will they precipitate a corresponding change in the nature and structure of the multinational enterprise as well? If so, what types of changes might occur? But before these questions can be answered, it is necessary to briefly examine some of the events that 17 have significantly altered the international environment.
VII. The Changing International Environment
While it is impossible to chronicle all of the changes that have occurred since the MNE gained prominence following WWII, Porter (1986) identifies a number of trends that have greatly affected the nature of international competition. Specifically, he cites the growing similarities between countries, the emergence of large and fluid international capital markets, failing trade barriers, revolutions in technology, the increased dissemination of technology and information, and the rise of new global competitors. To these major environmental changes, Porter also adds, "...some important cross currents
[that] have made the patterns of international competition different and more complex since the 1960s and early 1970s” (1986, pg. 3). These include slowing growth rates in many countries, the erosion of traditional sources of comparative advantage (such as labor costs, natural resources, and access to technology), new forms of protectionism (such as local content laws), new types of government inducement to potential foreign investors, the proliferation of coalitions among firms of different nationalities, and the growing ability of firms to tailor their products to local markets.
Of all the changes mentioned above, the growing similarities between national markets and the diffusion of technology are perhaps the most important. In his now classic 1983 article, Levitt notes that, "A powerful force drives the world toward converging commonality, and that force is technology....The result is a new commercial reality-the emergence of global markets for standardized consumer products on a previously unimagined scale of magnitude...” (1983, pg. 92). What makes this observation so important is that, in a world of standardization, many firm-specific advantages become 18 obsolete, and with them one of the MNE’s most important assets.
Peter Drucker (1986) offers another important insight. He notes, The talk today is of the "changing world economy”. I wish to argue that the world
economy is not "changing”; it hasalready changed - in its foundation and in its structure - and in all probability the change is irreversible.” (1986, pg. 2).
Drucker cites 3 fundamental causes for this change: 1) the "uncoupling” of the
primary products economy from the industrial economy, and its accompanying diminution of the comparative advantages associated with natural resources; 2) the uncoupling of employment and production within the industrial economy itself, largely due to increased productivity gains from technology; and 3) the partial uncoupling of trade and economic growth, with capital flows replacing trade as the driving force of the world economy. If Drucker’s analysis is correct, it implies that the primary sources of location-specific advantages are now passe.
Finally, recent changes in the international political economy must be acknowledged. The growing convergence of political ideologies between the
East and West and the increased spirit of cooperation that has accompanied the end of the Cold War will have a profound (and unpredictable) effect on the world economy. In the past, government actions have seriously affected the
MNE by creating inefficiencies which encouraged the use of internal markets on the one hand, and by attempting to restrict and control the MNE’s activities on the other. Both Dunning (1979, pg. 288) and Buckley and Casson (1976, pg.38-
39) argue that government intervention in the form of tariff barriers, restrictions on capital movements and/or FDI, and tax differences can justify the MNE’s internalization of external markets. Further, a number of authors have examined the rationale for Host government restrictions on FDI and the MNE and note that these policies have often determined the pattern and nature of MNE investment. 19 Recently, however, the relationship between host governments and the
MNE has undergone a significant shift. Globerman (1988) notes that,
“...numerous countries have relaxed legislative restriction on inward foreign direct investment and/or on other aspects of multinational firm (MNE) activities.”
(1988, pg. 41). Whether these actions are related to the political events cited above or not, they are likely to have a major impact on the way MNEs structure their international operations. On one hand, the elimination or reduction of restrictions on FDI could encourage firms to seek greater ownership and control of their foreign units. On the other hand, improved international political and economic relations could reduce the market imperfections which encourage FDI in the first place. While it is clearly premature to suggest that the firm’s incentive to internalize a failing market will be totally removed, it is quite possible that the benefits associated with conducting international activities within the MNE may be greatly reduced by these occurrences.16
VIII. Organizational Responses
Taken in toto, the environmental factors discussed above present a serious challenge to the MNE. As social, economic and political developments improve the efficiency of international markets and promote the convergence of technology and consumer tastes; the MNE and its characteristic network of subsidiaries may begin to appear increasingly obsolete.17 Robinson (1981) argues that, “In such a setting,18 centralized, equity-based control over globally integrated production systems becomes ever more costly; that is, the internal multinational corporate markets-for capital, for skills, for technology, for goods- are becoming less efficient (more costly) in some cases than those markets external to the multinational corporate system.” (1981, pg. 16). 20 In addition, while one set of factors is improving the efficiency of external markets, another is reducing the benefits of hierarchical governance.
Specifically, the diffusion of technologies and management practices and the growth of international competition have substantially eroded the basis of many firm-specific advantages; while the declining importance of natural resources in the production process has diminished the need for vertical integration. In both of these cases, the traditional justification for organizing international economic activities within the MNE is brought into question. In a world of standardized products and technologies, and similar consumer tastes, two things become paramount: the ability to create and apply new technologies, and the ability to access major international markets. In either case, the MNE provides a highly suspect vehicle for achieving these goals. According to Dunning (1989),
"...it may be that the type of technological advances of the 1980s and 1990s will require different organizational forms to exploit them...lt is by no means certain that the resource endowments needed for these innovations and their development will be those which generated their predecessors, nor that the hierarchical structure of firms evolved through the 1960s and 1970s will be suitable for the remaining years of the twentieth century.” (1989, pg. 26).
But if the MNE with its emphasis on hierarchical control of an international production and distribution system is no longer suitable, what type of organization is appropriate for today’s turbulent economic environment? Root
(1984) argues that recent events
“...will orient firms more toward long-term contractual arrangem ents with other firms than in the past. "Going it alone” will become a less viable strategy; firms will move away from "internalization” and toward "externalization.” They will more freely transfer nonstrategic production processes to firms, say, in the developing countries under contract manufacture, licensing, and minority-joint venture arrangements that are based far more on cooperation than on control by the international firm.” (1984, pg. 23). 21 Oman (1984) terms these latter forms of inter-firm cooperation “New
Forms of Investment” (NFI). According to Hennart (1989), “"New Forms” usually include arrangements that fall short of majority ownership, such as various forms of contracts (licensing, franchising, management contracts, turnkey and product-in-hand contracts, production sharing-contracts, and international subcontracting) as well as joint ventures.” (1989, pg. 212). Although NFIs have recently received a great deal of attention, their analysis has often been predicated upon government restrictions to FDI. For example, a recent article by
Franko (1987) found that "the use of NFI's by US firms was primarily a function of the interaction of competitive behavior in oligopolistic industries and the ownership and trade-protection policies of a number of [developing country] governments.” (1987, pgs. 39). However, based on the discussion developed thusfar, the rationale for using these new forms of investment goes far beyond the satisfaction of government restrictions on the MNE.
As early as 1981, Contractor (1981) noted that, T he generalization that multinational firms will prefer “Internalization” via direct investment over the sale of technology via licensing is a proposition that needs to be more closely examined...” (1981, pg.81). A survey of 51 U.S. MNEs conducted by Robinson
(1983) found that, There seemed to be a rapidly increasing interest in various forms of international contracting for technical, managerial, and marketing skills.” (1983, pg. 156). In fact, Robinson found that only 5 of the 51 firms had not entered into at least one minority joint venture (1983, pg. 100), indicating that
MNEs may view NFIs as an effective means of coordinating some of their international operations. All of this has led Auster to note that the use of new forms19 has “..skyrocketed over the last decade and a half." (1987, pg. 3). 22 IX. “The Coming of the New Organization”
The growing body of literature on joint ventures20 and other contractual arrangements21 suggests that the opportunities presented by the “new” international environment do not readily lend themselves to the traditional structures of the past. The model of the MNE presented above, based on the internalization of markets, vertical and/or horizontal integration, and the central control of an international network of subsidiaries, must be replaced by a new paradigm which stresses “externalization”, cooperation with host governments and other firms, and the coordination of activities through contractual agreements. Perlmutter and Heenan (1986) suggest that “The new global modeL.is more flexible about ownership and managerial control. It encourages joint decision making, vertical and horizontal planning, and the fusion of competent allies from around the world despite cultural differences.” (1986, pg.
152). The growing use of joint ventures, strategic alliances, and other “new forms” strongly supports this view.
Many authors have attempted to envision the type of organizational structures that suits this new paradigm. For example, Drucker (1988) foresees the development of an “information-based” organization. He argues that through the use of computers and telecommunications, organizations will be able to overcome the increased complexity and bounded rationality that limited earlier designs. This, in turn, will lead to smaller, flatter structures controlled by highly trained specialists rather the middle managers.22 Ohmae (1985), on the other hand, envisions a somewhat different outcome. He notes that the increased role of capital in production and the growing costs of developing and commercializing new products make it necessary for firms to compete in all 23 major markets simultaneously. Since few (if any) firms have the resources required to achieve this goal, companies will be forced to join together through alliances and/or joint ventures with firms in other major markets.
Carried to an extreme, both visions could produce larger, more complex networks of independent organizations loosely linked through R&D, production, and distribution agreements. Such an arrangement is similar to the “value- adding” partnership (VAP) discussed by Johnston and Lawrence (1988). The
VAP is “a set of independent companies that work closely together to manage the flow of goods and services along the entire value-adding chain.” (1988, pg.
94). In many ways, the Japanese trading company (a.k.a. Sogo Shosha) represents the quintessential VAP. As Yoshino and Lifson (1986) note, the sogo shosha is an
“...organizer of other firms...[and] is active from the very earliest or “upstream” activities of raw material extraction or creation, through multiple stages of production, fabrication, and distribution, “downstream” to the end user...Although it sometimes owns a partial or controlling interest in the corporations actually performing the activities...the primary role of the sogo shosha lies in the coordination or linkage of them.” (1986, pg. 6).
Based on these observations, a new organizational form may be evolving which is much more amorphous than the equity-based structures of the past. Like the sogo shosha, the MNE of the future could coordinate its international operations through contractual agreements with host governments, other MNEs, and local firms, and thus become primarily the nexus for the exchange of capital and information within a large network of independent organizations. By externalizing the firm’s value-added chain in this way, the MNE is better able to take advantage of the threats and opportunities which accompany the turbulent international environment. While it would be 24 premature to suggest that this structure uiU become the dominant form of the
MNE in the future, it seems inevitable that the multinational will require some structural modifications if it is to remain a major vehicle for coordinating international economic activities.
Statement of the Problem
Coase (1937) argued that economic activities would be organized within the firm whenever the cost of using the marketplace was higher than the cost of creating and maintaining an internal market. Consequently, as the cost of using external markets declines, one would expect firms to conduct a greater portion of their business activities in the marketplace. Based on the arguments developed above, an interesting question can be raised; namely, ‘Have recent changes in the international economic, social, and political environment affected the way in which multinational enterprises (MNEs) organize their international operations?’.
The thesis of this dissertation is that recent events have substantially increased the efficiency of international (and national) markets, inducing a corresponding decline in the costs of market-based arrangements. If this is in fact true, it could have a profound impact on the way MNEs structure their international operations. Multinationals have traditionally favored internal markets; however, as the cost of using external markets declines, so does the incentive to build and maintain a network of international subsidiaries. Under these circumstances, MNEs may find it more efficient to coordinate their activities through a market-based system of contractual agreements (ie. licensing, value-adding partnerships, strategic alliances, and other “new forms”) and/or minority joint ventures. 25 As noted earlier, government policies and location-specific factors can strongly influence the efficiency of external markets as well as provide firms with an incentive to “internalize” a particular market. As a result, it is likely that the firm’s preferred method of coordinating its activites will vary from country to country. This may make it possible to identify a group of environmental variables which encourages a corresponding increase in market-based MNE activities.
Objectives
In many ways, this study attempts to test a fundamental premises of both
Transaction Cost Analysis and the Theory of the MNE; namely, that the choice between organizing economic transactions through the market or the firm rests on the relative efficiency of each option. While it is clear that a number of significant changes have occurred in the international economy since the end of the Second World War, it is not clear if and how these changes have affected the structure of the MNE.
It has been suggested that the current international environment may no longer favor internal markets and that, in many cases, market-based arrangements now provide the most efficient means of coordinating international activities. While there is little doubt that the “new forms of investment” have grown rapidly in recent years, it is not entirely clear what has caused this increase, or the role specific economic, social, and political factors have played in its development. Finally, it is still uncertain whether the use of these market-based mechanisms is consistent with the predictions of accepted theory, particularly Transaction Cost Analysis and the Theory of the MNE. 26 Therefore, the objective of this study is to determine whether recent changes in the international environment have had a measurable effect on the way MNEs structure their international operations. Since the MNE has traditionally relied on an extensive, well coordinated network of international subsidiaries to coordinate its activities, a substantial reduction in the costs of using external markets should produce a corresponding decline in the portion of the firm’s activities conducted through internal markets. Consequently, 3 specific objectives emerge:
1) To determine whether MNEs are conducting an increasing portion of their activities through external markets.
2) To correlate the use of external markets with specific forces in the international environment.
3) To infer the types of organizational changes that might accompany a shift from internal to external markets.
Scope of the Research
Due to the potential complexity of the problem being explored, it is imperative that the scope of this project be defined in advance. However, before doing this, it may be useful to discuss what this study will not encompass. The arguments developed thusfar suggest that recent changes in the international environment have increased the efficiency of external markets. While this is certainly a reasonable proposition, the actual proof of this statement is beyond the scope of this research. Investigating the relationship between environmental variables and market efficiency is an extremely difficult task best left to the economist. Although some evidence has been already provided to support this contention, the objective of this project is to determine whether the structure of 27 multinational enterprises has been affected by these environmental changes, not whether the efficiency of various markets has increased.
This leads to a second, and equally important point. Transaction Cost
Analysis notes that there are both costs and benefits associated with internal organization. This study makes no attempt to examine how the cost of using internal markets has been influenced by recent events. Again, this is a topic that warrants independent investigation. Further, it is important to note that MNEs can favor external markets due to an increase in the cost of using internal markets as well as a decrease in the cost of using external markets. Although this study argues in favor of the latter option, only the discussion presented above is offered as proof of this claim.
Thus, the current study purports to explore only the effect certain environmental factors have on the firm’s proclivity to use external markets to coordinate its international activities. Due to the extreme complexity of these relationships and the necessarily limited scope of this study, this project should be considered exploratory in nature. Accordingly, the results should be used to strengthen existing theories, develop new hypotheses, and guide future research.
Methodology
Not surprisingly, there are a number of different ways to investigate the problem described above; however, two methodologies have been widely used to analyze the relationship between the firm and the environment. They are: the case study, and correlation analysis. For example, Chandler (1962) used case studies and historical analysis to explore the relationship between the environment, a firm’s strategy, and its corresponding structure. Applied to the 28 current study, this approach would require the selection of a representative sample of firms from several countries in a number of key industries, and interviews with corporate executives to determine the types of organizational responses their firms are making to recent environmental changes. While this approach is undeniably valid, several problems make it inappropriate for the present study.
For one thing, it is often difficult to gain access to upper management; for another, managers may be reluctant to discuss highly proprietary information about their firm’s strategies and operations. Further, since MNEs are large, complex organizations, they are likely to respond to environmental stimuli very slowly and with some reluctance. Thus, managers may ignore important environmental signals or react in ways which favor present strategies even at the cost of future performance. All of these problems are compounded by the need to examine MNEs from several countries.
Fortunately, many of these difficulties can be avoided by looking at the aggregate behavior of firms. By examining the actions of the entire population of
MNEs operating in a specific national market, a much clearer picture of emerging trends can be gained. In addition, since the efficiency of external markets varies widely between countries, examining the behavior of MNEs in a number of national markets should make it possible to establish a relationship between key environmental variables and firm behavior.
Therefore, this study will employ correlation and regression analysis to examine the relationship between specific environmental factors and the firm’s propensity to use external markets. Using data collected by various national and international sources, a purposive sample of developed and developing countries will be used to examine and the relationship between a broad group 29 of environmental variables and several measures of the firm’s market orientation. This approach is similar to the one utilized by Stopford and Wells
(1972) to investigate the relationship between certain firm-specific variables and the MNE’s propensity to enter minority or majority joint ventures.23 Finally, it should be noted that the application of regression analysis in the current study does not preclude the use of case studies at a later date.
Research Hypotheses I. Introduction
The fundamental hypothesis of this dissertation is that environmental factors which act to improve the efficiency (lower the cost) of external markets will induce MNEs to shift activities to the marketplace. Although the hypothesis is straightforward enough, it is by no means clear how to translate this proposition into a set of testable research hypotheses. In fact, Buckley (1988) recently concluded that the general theory of ‘‘Internalization’’ cannot be tested directly, and requires the formulation of special theories which “has been shown to be difficult and problematic but not impossible.” (1988, pg. 190). The same might be said of a theory of “externalization". Therefore, the purpose of this brief section is to identify a group of environmental variables which will serve as the study’s Independent Variables (IVs), as well as a number of Dependent
Variables (DVs) which will act as measures of the firm’s market orientation.
Once this has been accomplished, a more specific set of research hypotheses can be developed.
II. Independent Variables
The international environment is a complex setting. This makes the identification of a set of external factors that affect market efficiency a 30 complicated task. This task may be simplified, however, by defining a small number of distinct types of factors, and selecting several key measures from each category for use as the study’s Independent Variables.24 Using this framework, 3 basic dimensions of the international environment can be identified: Economic, Social, and Political. Unfortunately, the rather small number of possible data points (which is limited to the number of nations that regularly report economic and social data) severely limits the total number of
IVs that can be used in this study. Therefore, it is extremely important that the most accurate, representative variables from each category be utilized.
Although arguments can be raised for or against the inclusion of a particular measure, it is believed that the following 12 variables satisfy these requirements and will serve as the study’s Independent Variables. (Please note that these variables will be measured for individual countries.)
The study’s 12 Independent Variables are:
Economic 1) PCGNP (Per capita GNP) 2) GDPGRO (Average annual GDP growth rate) 3) TOTRADE (% total trade to GDP) 4) GOVEXP (% government expenditures to GNP) 5) INVEST (% domestic investment to GDP)
Social 1) PQLI (Physical quality of life index) 2) LIT (Literacy rate) 3) DOCTOR (1000 people per doctor) 4) %AGR (% population in Agricultural sector) Political 1) POL (Political system) 2) REST (Restrictions on foreign business and FDI) 3) PRI (Political risk index) 31 III. Dependent Variables
Ideally, the study’s Dependent Variable should provide a comprehensive measure of the portion of an MNE's total activity that is conducted through external markets. Clearly, this is not a standard measure. Consequently, some proxy must be found which captures the flavor of this ratio, but is also available for a large number of countries. It is believed that the following 4 variables, when considered together, provide such measures, and have been constructed for use as the study's Dependent Variables.
The study’s 4 Dependent Variables are:
1) %INTFIRM: The portion of a country’s total trade conducted through the majority owned foreign affiliates (MOFAs) of foreign MNEs.
2) %AFSALES: The portion of local MNE affiliate sales sold to other affiliated parties through the MNE's internal market.
3) %MOFA: The portion of the total MNE direct investment in a given country found in majority (as opposed to minority) owned foreign affiliates.
4) %WORK: The portion of a country’s workforce employed by foreign MNEs that work for local majority owned foreign affiliates.
In all four cases, lower values of a Dependent Variables indicate that
MNEs are conducting a greater portion of their activities through external markets in a particular country.
IV. Research Hypotheses
Based on the arguments developed above, it is hypothesized that in countries with more efficient external markets, MNEs will conduct a greater portion of their activities through the marketplace. When this logic is applied to the Dependent Variables described above, the following general Research
Hypotheses emerge: 32 1) Environmental factors that increase the efficiency of external markets will be negatively associated with %INTFIRM.
2) Environmental factors that increase the efficiency of external markets will be negatively associated with %AFSALES.
3) Environmental factors that increase the efficiency of external markets will be negatively associated with %MOFA.
4) Environmental factors that increase the efficiency of external markets will be negatively associated with %WORK.
Table 1 provides an overview of the Independent Variables and their hypothesized relationship with each Dependent Variable. A detailed discussipn of the issues presented in this section appears in Chapter III.
Limitations of the Study Unfortunately, all empirical research is subject to errors and limitations. In this case, two major problems arise which deserve special attention. The first involves the data being used in this study. Since all of the study’s Independent and Dependent Variables are measured at the national level, this study must rely on aggregate statistics gathered by a number of different sources. Needless to say, this can create problems. For one thing, the use of a specific
Independent or Dependent Variable may hinge on the availability of accurate data on the countries of interest. Although many international institutions regularly collect data from their members, the quality of these statistics can vary widely between countries. This is particularly true of developing countries which may be unable or unwilling to release accurate information about their economic and social performance. Thus, the same statistical series collected by different sources may not be directly comparable. 33 IABLE1. Hypothesized Effect of the Study’s Independent Variables
Independent Variable Dependent Variable
1) PCGNP (+) 2) GDPGRO (-) 3) TOTRADE <+) 1)%INTF1RM (-) 4) GOVEXP (*) 2) %AFSALES (-) 5) INVEST (+) 3) %MOFA (-) 6) PQLI <+) 4) %WORK (-) 7) LIT (?) 8) DOCTOR (-) 9) %AGR (-) 10) POL <+) 11) REST (+) 12) PRI (?)
(a) The sign indicates an increase (+) or decrease (-) in the value of a particular variable. For example, it Is hypothesized that an Increase in PCGNP will cause a decrease in %INTFIRM; thus, these 2 variables should be negatively correlated. 34 Further, many international statistics are only released to the public long after they have been collected, while others are measured at long intervals or may take years to compile or calculate. Finally, some scholars have argued that several commonly used economic series are not appropriate for their intended purpose. For example, Myrdal (Myrdal, 1973) and other economists have argued quite persuasively that GNP is a largely inaccurate indicator of economic performance; and similar arguments have been raised about other aggregate statistics as well.
The second major limitation involves the use of correlation and regression analysis to investigate the study’s primary relationships. Strictly speaking, this approach cannot be used to establish causality (although researchers frequently use it to this end); and therefore, it is extremely important that a strong conceptual framework be developed to guide both the study and the interpretations of its results. This has been the purpose of the “Overview” and “Background” sections of this chapter, but other solutions to this problem can be suggested as well. For example, several International Business researchers argue that the “triangulation” of different methodologies is the best way to strengthen ones findings.25 While it is unfortunately beyond the scope of this study to employ this technique, the results of this study could be greatly enhanced by conducting a case study of selected MNEs at a later date.
Even though the use of second-hand international data and correlation analysis clearly limits the power and scope of this study, it should nonetheless be possible to gain valuable information on a highly dynamic phenomenon that is of interest to a wide and diverse audience. Given the potential benefits of this project, the aforementioned limitations are deemed to be acceptable. 35 Contribution of the Study
The knowledge gained from this study will be of interest to two diverse audiences; international managers and corporate planners, and national and international governmental bodies. Since the multinational’s role in the international political economy has been intensely debated for many years,26 the objectives of MNEs and host governments are often assumed to conflict. As a result, many governments have tried to increase their share of the benefits from local MNE activities by placing restrictions on foreign direct investment and other MNE practices.27 However, as Hennart (1989) notes, rather than increasing their share of the benefits, these actions have often “decreased the size of the available rents...” instead (1989, pg. 228).
The topic explored in this dissertation is one of the few instances when the goals of MNEs and host governments can realistically be said to converge.
The “externaiization” of a portion of the firm’s value-added chain to local partners can provide MNEs with an effective method of improving their performance, and at the same time increase economic activity in the host country. However, the success of this approach depends on realigning the relative cost of using external markets. In the past, governments have attempted to make this adjustment by raising the cost of internal organization. This study argues that the same effect can also be achieved by increasing the efficiency of external markets.
By identifying a specific group of economic or political variables capable of shifting MNE activities from internal to external markets, the needs of both corporate planners and public policy-makers will be served. On one hand, externaiization provides MNEs with an additional strategic tool capable of 36 coping with environmental uncertainty, entering hitherto restricted markets and improving overall performance. On the other, it enables host governments to increase their control over the local economy and improve competitiveness without suffering the inefficiencies associated with restrictive policies.
Organization of the Dissertation
This dissertation is divided into 5 Chapters, each containing a number of specific sub-sections. Chapter I provides an introduction to the topic, including the conceptual development of the problem and a statement of the study's objectives and hypotheses. In Chapter II, a more detailed discussion of the study’s theoretical basis is presented through a review of relevant literature.
Since research from three diverse fields is necessary to fully develop the problem, the chapter is divided into a separate section for the Theory of the
MNE, Organizational Theory, and International Strategic Management. A final section is also included which summarizes the most important aspects of this literature. Chapter III focuses on the Methodology used in this study, it provides a detailed examination of the sample, the Independent and Dependent
Variables and the statistical methods employed. Chapter IV presents the data, and the results of the statistical analysis; while Chapter V contains the conclusions and a discussion of their implications for existing theory and future research. 37 NOTES
1) The term Value-added chain” is used here to denote the stages of the production process, from raw material extraction to final distribution, in which the firm is directly involved. For a further development of this subject see Kogut (1984).
2) This is the title of a recent article by Peter Drucker (1988).
3) Strictly speaking, Coase was concerned with the role of the ‘entrepreneur” in organizing economic activities; however, the entrepreneur often used the firm as a vehicle for his activities.
4) In Markets and Hierarchies. Williamson suggests that Alchian and Demsetz (1972) make this claim (see Williamson, pg. 3); but, it is arguable whether they actually go this far. They do, however, note that Coase’s insight “is a difficult proposition to disagree with or refute.”.
5) Williamson adopts Commons’ (1934) approach of using the transaction as the basic unit of economic analysis, hence the term “transaction cost analysis".
6) Chandler develops this theme in greater detail in The Visible Hand (1977).
7) See Williamson (1975, pgs. 117-131). Among other things Williamson identifies internal procurement bias, internal expansion bias, program persistence bias, communication distortions, and employment disincentives.
8) For example, see Fayol’s “une doctrine administrative”, Weber’s theories on Bureaucracy and Leadership, and Taylor's "Scientific Management”.
9) Among the key environmental factors studied were Technology, explored by Woodward (1958, 1965); Firm Size, studied by the Aston group (Pugh and Hickson, 1976); Environmental Uncertainty studied by Thompson (1967) and Lawrence and Lorsch (1967); Decision-making examined by March and Simon (1958) and Cyert and March (1963); and numerous other structural variables such as Organizational Structure, Concentration of Power, etc..
10) The analogies between Biology and Organizational Behavior have recently generated a great deal of interest; particularly the idea that organizations are "selected” by or adapt to specific environment factors. For example, see Hirshleifer (1977) or Hannan and Freeman (1977).
11) For the purpose of this study, an MNE is defined as a company which owns or controls assets in 2 or more countries. This definition is similar to the one developed by Buckley and Casson (1976, pg. 33). 38 12) Inter-firm trade refers to trade conducted between units of a single company. Unfortunately, there is no standard for considering a unit to be part of a firm, and ownership requirements can vary from 5-25 per cent depending on the researcher and his purpose.
13) Because different authors address slightly different aspects of this problem, it is worthwhile noting that the terms “foreign production”, “foreign direct investment”, and “MNE” are often used interchangeably to describe the same phenomenon.
14) This problem arises because the strategies which maximize the operations of a globally integrated firm often do so at the expense of responding to local market needs, and vis-a-versa. For a formative discussions of this issue see Doz (1980), Bartlett (1981), and Prahalad and Doz (1981).
15) The matrix structure attempts to combine 2 distinct organizational designs into a single whole. A typical example might be uniting a structure based on product divisions with one based on geographical divisions to create a “product-area" matrix. Needless to say this is a difficult process. Matrix organizations are explored in detail in Davis and Lawrence’s book Matrix (1977).
16) It is worthwhile noting that the recent trend toward ideological convergence could actually decrease the efficency of international markets by encouraging a rise in nationalism and the further fragmentation of the international economy. However, even if countries like the USSR and Yugoslavia do splinter into a number of independent nations, the traditional structure of the MNE may not provide an effective vehicle for exploiting these opportunities, due to the heightened political risk and managerial difficulties associated with operating in a large number of distinct, small national markets.
17) For example, see Behrman (1981), Stoever (1982), Dymsza (1984), and Dunning (1989, Chpts. 2 & 12).
18) Robinson is specifically addressing the changes brought about by the “multi-actor era”. This era is characterized by the increased participation of external parties (such as host governments and international agencies) in the MNE’s decision-making process.
19) Rather than using the term “New Forms of investment”, Auster speaks of “International Corporate Linkages” (ICLs) which include such things as “joint research and development, technological exchanges and transfers, and full- fledged joint ventures.” From this description the differences between ICLs and NFIs appear minor at best. 39 20) For a further development of the subject of joint ventures see Killing (1983), Beamish (1985), Gomes-Casseres (1987), Datta (1988), Harrigan (1988), Kogut (1988), and Geringer and Herbert (1989).
21) For a discussion of these arrangements see Franko (1987), Perlmutter and Heenan (1986), Jain (1987), Morris and Hergert (1987), Ajami (1989), Ohmae (1989), Contractor (1990).
22) An example of such an organization was the British civil service in India which used an extremely small number of top managers to control the vast and diverse Indian sub-continent.
23) The results of this analysis are presented in Appendix C of Stopford and Wells’ book Managing the Multinational Enterprise (1972).
24) Although factor analysis can be used to accomplish this task, the division of the environment is used here as a conceptual tool, strictly to facilitate the identification of Independent variables.
25) “Triangulation” involves the use of 2 or more distinct and unrelated methodologies to analyze a single problem. An excellent example of “triangulation” is found in Bartlett and Ghoshal’s recent book Managing Across Borders (1989). The authors employ 3 distinct research methodologies (case study, survey, and statistical analysis) in their study of various MNE activities, which greatly enhances the strength of their findings.
26) Among many examples, see Hymer (1971), Makler, Martinelli and Smelser (eds) (1982), and Behrman (1981,1985).
27) For example, see Behrman (1981), Encarnation and Vachani (1985), and Hennart (1989). CHAPTER II
REVIEW OF LITERATURE
Introduction As noted earlier, this study draws upon 3 large and very diverse bodies
of research: the Theory of the Firm (or more precisely the Theory of the MNE),
Organizational Theory (particularly Contingency Theory and the Strategy and
Structure research), and International Strategic Management (particularly the
work on Cooperation and Collaborative Arrangements). Since it is clearly
beyond the scope of this study to review each of these topics in detail, a
somewhat different approach will be taken in this chapter. Rather than present a
“review” of literature in the classical sense, an attempt will be made to provide theoretical support for the arguments advanced in the “Overview” of Chapter I.
This will be done by critiquing relevant studies from the field of International
Business. While this approach may be criticized by some for lacking rigor, the breadth and complexity of the current subject necessitates it.
This chapter is divided into 5 major sections: an introduction, a section for each of the 3 bodies of literature noted above, and a final section containing a summary of the chapter's most important points. It is hoped that this review will provide the reader with a much deeper understanding of the motives behind and foundation for the current study.
40 41 MNE Ihflory I. Introduction
The Theory of the Multinational Enterprise (MNE) has become one of the most intensely studied areas of International Business research. As a result, a number of competing theories have emerged to explain the existence and activities of the MNE. Buckley (1981) provides one of the most complete frameworks for classifying this research in his “Critical Review of Theories of the
Multinational Enterprise”. (For the remainder of this section, all references to
Buckley are assumed to refer to this 1981 article unless otherwise noted)
According to Buckley, 6 distinct approaches to MNE Theory can be identified in the literature:
1) The Hymer-Kindleberger tradition 2) Product Life Cycle Hypothesis 3) Internalization Theory 4) Diversification and Internationalization 5) Location Theory 6) Attempts to develop a General Theory of the MNE
In addition to the six categories defined by Buckley, two other approaches have received a great deal of attention in recent years; namely,
Transaction Cost Analysis and Economies of Scale. The remainder of this section is devoted to a brief review of each approach; but before proceeding with that discussion, the basic requirements of a theory of the MNE will be explored.
In general, theories of the MNE attempt to explain the purpose and function of the Multinational Enterprise. Buckley notes that the MNE can be defined in 4 distinct ways: 1) an operating definition such as Buckley and
Casson’s (1976) definition of a firm which controls or owns income generating 42 assets in more than one country, 2) a structural definition based on the way in
which a firm is organized, 3) a performance criterion such as % foreign sales or
assets, number of foreign subsidiaries, etc., and 4) a behavioral criterion such
as the ethno-, poly-, and geo- centric management styles defined by Perlmutter
(1969).
In addition, Buckley argues that a comprehensive theory of the MNE must
address 5 critical issues: 1) the relationship between market imperfections and the MNE, 2) the concept of internalization (ie. when internal markets are
preferred to external markets), 3) the relationship between trade and FDI, 4)
domestic diversification versus international expansion, and 5) the role of
Location Theory in determining the costs and motives of FDI. Given the scope of this subject, it is not surprising that a number of distinct approaches have
emerged in the literature.
II. The Hymer-Kindieberger Tradition
Even though firms have been operating in a multinational capacity for well over a century,1 there were few serious studies of the MNE before 1960.
With the exception of Southard’s (1931) study of U.S. firms in Europe and
Dunning’s (1958) study of U.S. firms in Britian, Stephen Hymer’s doctoral dissertation fTheJnternatlonal Operations of National Firms (original 1960)] was the first major investigation into the activities of the MNE. Consequently, Hymer is generally considered to be the Father of MNE Theory and his writings have been extremely influential and extensively reviewed.2
According to Buckley, the Hymer-Kindieberger approach is based on a simple proposition advanced in Hymer’s doctoral thesis; namely, that a foreign firm must possess some specific advantage to compensate for the innate 43 advantages of local firms. Kindleberger (1969) argued that these “firm-specific” advantages proceed from 4 sources: 1) imperfections in the markets for goods caused by product differentiation, marketing skills, etc., 2) imperfections in factor markets due to proprietaty technology and management skills, 3) economies of scale, and 4) government intervention in the marketplace. Advantages arising from any of these sources can allow a foreign firm to overcome its lack of knowledge about local conditions and the costs of doing business at a distance.
However, in order to serve as a theory of the Multinational, the firm's choice of
FDI over other entry modes (such as licensing or exporting) must also be explained.
Hymer explained this choice in 2 ways: first, he noted that many types of knowledge were difficult to transfer through licensing and other forms; and second, he argued that FDI could provide the firm with quasi-monopoly power over the markets it served. Accordingly, the Hymer-Kindieberger approach views the monopolistic power gained by exploiting certain firm-specific advantages through FDI as the raison d’etre for the Multinational Enterprise.
Much of Hymer’s later work was devoted to the study of the effects this monopoly power has on the nations that host FDI. For example, in “The
Multinational Corporation and the Law of Uneven Development” (1971) Hymer argues that the MNE can be seen as an instrument which channels resources away from peripheral nations (i.e. host countries) to central nations (i.e. home countries) through its international network of subsidiaries.
Hymer’s writings have had an enormous impact on MNE Theory. For example, McClain (1983) argues that newer theories, such as Dunning’s
Eclectic Paradigm or Internalization Theory, add little to the Hymer-Kindieberger approach. In fact, Horaguchi and Toyne (1990) note that by 1968 Hymer had 44 combined Coase’s view of the firm with his own market imperfection theory,
making him the first scholar to apply Internalization and Transaction Cost
Analysis to the MNE a s well.
On the other hand, Dunning and Rugman (1985) consider Hymer’s
greatest contribution to be shifting the focus of FDI theory away from classical
trade and investment theory toward an analysis of the MNE. They note that
HymeTs interest in the theory of industrial organization (Bain, 1956) and his
realization that MNEs can reduce risk and achieve monoplistic power by
segmenting markets, removing competition, or exploiting firm-specific
advantages are the most significant aspects of his dissertation. In addition, they
acknowledge the importance of Hymer’s emphasis on the dynamic nature of
ownership advantages, and his insight that international diversification can
stabilize profits when profits are negatively correlated between countries.
However, they also argue that Hymer failed to uncover the full meaning of
market failure or to recognize that hierarchical organization can replace
imperfect markets for efficiency reasons as well as to gain monopolistic power.
Buckley associates several other theories with the Hymer-Kindieberger tradition. The most important of these are Aliber’s “A Theory of Direct
Investment” (1970) and Knickerbocker’s (1973) Theory of “Oligopolistic
Reaction”. In each case, the authors use firm-specific advantages or monopoly power to explain the activities of the MNE. For example, Aliber’s theory examines the advantages accruing to firms operating in a particular currency area. He argues that investors treat the foreign investments of a firm as if they were denominated in the firm’s domestic currency. As a result, MNEs operating in certain currencies (such as the U.S. dollar or British pound) have an inherent advantage over local competitors since they can often borrow at lower interest 45 rates due to their lower required risk premiums.
On the other hand, Knickerbocker’s theory focuses on the oligopoly power associated with certain industry structures. Based on a study of U.S. firms, Knickerbocker found that firms in industries which exhibited a high degree of concentration (ie. oligopolies) tended to invest in a given foreign marketen masse. Knickerbocker argued that this behavior was essentially a defensive move aimed at preserving the existing industry structure. Franko (1989) confirms this view in a recent study on global competition. According to Franko, the upsurge in U.S. investment in Europe in the 1960s was a defensive move to preserve the position of U.S. firms in the face of growing local and third-country competition.
III. The Product Life Cycle Hypothesis
In contrast to explanations which focus on the monopoly power of the
Multinational Enterprise, Product Life Cycle (PLC) models argue that multinationality is an outgrowth of a specific stage in the life cycle of a product or industry. Vernon (1966) was the first scholar to utilize this approach in an attempt to explain international shifts in trade and investment which contradicted neoclassical trade theory. Vernon argued that most products undergo a predictable life cycle (i.e. the New Product Stage, Maturing Product
Stage, and Standardized Product Stage) which encourages firms to seek resources and markets abroad at a certain stage. Since the activities of FDI and foreign production are by definition evidence of muitinationality, the product life cycle hypothesis also provides a theory of the MNE.
Vernon’s model is based on the premise that Knowledge is not a free good, but an independent variable in the decision to trade or invest. Since knowledge about a market can affect the ultimate outcome of a particular 46 investment, Vernon argued that producers were more likely to introduce new products in their home markets (New Product stage). During the 1960, several unique characteristics (such as consumers with high relatively average incomes, high unit labor costs and unrationed capital) made the U.S. market the location of most product innovations as well as the market for their introduction.
According to Vernon, the unstandardized nature of a new product places unique requirements on the location decision. Specifically, there is a need to maintain flexibilty in inputs and rapid communication between market participants in order to accommodate product modifications, but there is a minimal emphisis on costs due to the product’s low price elasticity of demand.
As product standardization occurs during the Maturing Product stage, however, the need to decrease production costs becomes the overriding concern. As a result, producers may be encouraged to shift production to a low cost foreign location.
Although empirical evidence supported such a shift of production within the U.S., Vernon was unable to find similar evidence in international production
(possibly because the process was too new to yield such evidence).
Nonetheless, he argued that producers were likely to export when costs plus transportation (and tariffs and taxes) were lower than production in the importing market, but would establish production facilities in other developed markets at som e point. Initially, these facilities would only satisfy local demand; but due to differences in labor costs, economies of scale, and the cost of capital, it could eventually become profitable to service third-country markets from these locations a s well.
In the final stage of the product life cycle (Standardized Product stage), knowledge about a product is readily available internationally, and competition 47 centers almost exclusively on price. As a result, the role of labor costs and other
local inputs becomes paramount and the attraction of specific foreign locations
may be greatly enhanced. Thus, Vernon used the unique pressures associated
with each stage of the product life cycle to explain the firm’s decision to service
a foreign market or to engage in foreign direct investment.
By 1979, Vernon (1979) conceded that the PLC hypothesis had lost
much of its validity. Although the model worked well in explaining the activities
of U.S. multinationals from 1945-1970, by the late 1970s, changes in the
international environment (particularly the increased use of global production
networks and the equalization of incomes between developed nations) had
greatly reduced its explanatory power. Even so, Vernon argued that the model
still applied to smaller firms and continued to explain many of the elements affecting location decisions in the new product stage. In addition, Vernon
suggested that the model could be used to explain some of the behavior of
European and Japanese firms, such as the types of innovations these countries
have spawned, and the dynamic nature of certain locational advantages.
However, as a result of these developments, most writers agree that the PLC no
longer offers a viable theory of the MNE.
IV. Internalization Theory
The basis of Internalization Theory lies in Coase’s (1937) insight that the actions of the firm can replace those of the market in certain circumstances. This occurs when the cost of conducting a transaction in the marketplace is higher than the cost of coordinating it within the firm. Although Coase developed this theory to explain the existence of the (domestic) firm, the same argument can be applied to the Multinational Enterprise, particularly since imperfections in international markets often impose a high cost on market-mediated exchanges. 48 Buckley and Casson formally developed the link between Internalization and the Theory of the MNE in their book The Future of the Multinational Enterprise
(1976). Their approach rests on 3 premises: 1) firms maximize profit in a world of imperfect competition, 2) when intermediate goods markets are imperfect there is an incentive to create an internal market, and 3) internalizing a market across national borders creates an MNE. Thus, Internalization is also a theory of the MNE.
Buckley and Casson note that the strongest incentives for internalization occur in the markets for Knowledge due to the unique problems associated with the production, transfer, and sale of this good. However, markets for perishable agricultural products, intermediate products, and raw materials also lend themselves to internalization. The reason for this being, each of these markets is particularly prone to failures of the type described by Williamson (1975) (ie. bounded rationality, opportunism, and asset specificity). For example, difficulties in pricing various types of Knowledge in the marketplace subject its sale to a high degree of buyer uncertainty and opportunism. In addition, Knowledge provides a natural monopoly which is best exploited through discriminatory pricing. As a result, it is often more efficient to transfer Knowledge and other products through internal rather than external markets. However, Buckley and
Casson note that internal markets also incur higher resource costs, higher « communication costs, and can be subject to government interference; and therefore, the firm’s decision to internalize a particular market may involve industry-, region-, nation-, and firm-specific factors as well. But whatever the reason, it is clear that Internalization provides a powerful theory of the MNE.
Rugman (1980) attempts to expand Internalization Theory into a general theory of FDI (and hence the MNE) by demonstrating that exising FDI theories 49 are only sub-versions of Internalization. Using arguments developed by Coase
(1937), Hymer (1960) and Buckley and Casson (1976), Rugman illustrates that theories which use imperfections in the market for Knowledge as the motive for
FDI [such as Johnson (1970), Caves (1971), McManus (1972) and Hirsch
(1976)], can easily be explained In terms of Internalization. In fact, Buckley and
Casson used these same imperfections to explain the firm’s incentive to internalize a particular market. Rugman also concludes that Vernon’s (1966)
Life Cycle model, Knickerbocker’s (1973) Oligopolistic Reaction Theory, and
Kojima’s (1978) Japanese model of FDI (see below) also fit within the boundaries of Internalization. In each case, markets are internalized not in response to specific imperfections, but to extend the monopoly power of the firm. This aspect of Internalization stem s from Hymer’s analysis of the firm.
Finally, Rugman looks at finance based explanations of FDI such as Aliber’s
(1970) theory of currency areas and theories based on international diversification. Once again, these can be explained in terms of Internalization, since in each case FDI occurs in response to imperfections in international financial markets.
In a more recent article, Rugman (1986) examines the link between
Internalization and Williamson’s (1975) “organizational failures framework” (i.e.
Transaction Cost Analysis). Rugman suggests that the MNE provides a means of assuaging Williamson’s 3 major transaction costs (bounded rationality, opportunism, and asset specificity) since MNEs have environmental scanning and information management systems which overcome bounded rationality, and they can use opportunism and asset specificity as entry and exit barriers.
This leads Rugman to conclude that the MNE’s internalization of a particular market and the replacement of a market by a hierarchical governance structure 50 are one in the same thing. Consequently, Rugman argues that Internalization and Transaction Cost Analysis are jargely identical.
If one accepts these arguments, Internalization must be considered a
“general” theory of FDI, and hence the MNE. However, Buckley (1988) recently noted that Internalization Theory is too. general to be tested empirically.
According to Buckley, testing the theory requires the acceptance of specific restricting assumptions which greatly limit the theory’s scope. As a result,
Buckley attempts to test Internalization conceptually, by looking for empirical evidence that contradicts the theory’s predictions. Based on this analysis, he concludes that existing research strongly supports Internalization Theory; and argues that even though the theory is quite general, it still provides a powerful tool for explaining flows of FDI (and thus the MNE).
V. Diversification and Internationalization
The Diversification approach to the MNE is based on the premise that firms act as risk averse investors and use foreign investments to stabilize their returns. In general, MNEs can diversify their operations in 2 ways: by product and markets, and financially by earning returns in more than one currency.
Hymer (1960) was one of the first to recognize the risk reducing potential of international diversification through FDI, but he did not explore this idea in detail. Instead, the further development of this concept was left to the field of
Finance. A typical example of this research is a study by Rugman (1977) which attempts to assess the relative merits of financial or portfolio investment, versus direct investment. Based on his analysis, Rugman concludes that diversification through portfolio investments does not capture all of the benefits available from diversification through direct investments. Thus, the MNE derives a clear advantage over large firms operating in a single country from its ability to 51 spread risks and stabilize returns through a portfolio of international direct investments. In terms of MNE Theory, this approach suggests that the risk reducing benefits obtained from “internationalization” can be great enough to justify the existence of the Multinational Enterprise.
VI. Location Theory
Location theories of the MNE argue that the firm provides a vehicle for transferring mobile resources such as technology, capital, and management skills, to areas with immobile complementary factors such as markets, raw materials, and labor. Consequently, location-specific endowments which promote FDI can also provide a rationale for the MNE. Common examples of this process include FDI to assure the steady supply of a location-specific raw material, FDI in offshore production facilities to take advantage of a cheap labor supply, or FDI as a means of gaining access to highly protected or fragmented local markets. In each case, the end result of an investment motivated by a location-specific asset is the Multinational Enterprise. Thus, Location Theory can provide a rationale for the MNE in some cases.
Perhaps the most interesting application of this approach is Kojima’s
(1978) Japanese model of FDI. The essence of Kojima’s argument is that
Japanese direct investment is typically based on comparative advantage, and hence acts to complement trade (trade-oriented FDI); while U.S. FDI often occurs in industries where U.S. firms already have a large comparative advantage, and thus serves as a substitute for trade (anti-trade oriented FDI).
Kojima notes that anti-trade oriented FDI tends to export employment, promote trade deficits, and stifle local production, and concludes that the Japanese model of FDI is superior. 52 Although Kojima’s analysis has been dismissed by most writers, his evaluation of U.S. FDI is not inconsistent with the work of Hymer (1960), Vernon
(1966), Kindleberger (1969), Knickerbocker’s (1973) and others. In each case,
FDI is used as a means of protecting the existing industry structure or extending the monopolistic power of the firm. In addition, Japan's almost total dependence on external sources for many factors of production (particularly raw materials and labor) makes it quite plausible that Japanese FDI often conforms to comparative advantage. In fact, it could be argued that the current explosion of
Japanese investment in Southeast Asia3 has been driven by the need to access location-specific endowments, and is therefore trade-creating.
VII. General Theories of the MNE
Originally appearing in Ohlin’s book The International Allocation of
Economic Activity (1977) under the title “Trade, Location of Economic Activity and the MNE: A Search for an Eclectic Approach”, Dunning’s “Eclectic
Paradigm” represents the most important attempt thusfar to develop a general
Theory of the MNE.4 Simply stated, Dunning’s theory argues that foreign production will occur if and only if 3 basic conditions are satisfied: 1) the firm possess some specific advantage, 2) the foreign location offers some specific opportunity, and 3) internalizing the transfer within the firm provides the most advantagous means of sen/icing the foreign market. These arguments clearly follow from the Hymer-Kindieberger approach, Location Theory, and
Internalization Theory respectively; and therefore, the Eclectic Paradigm is a
“general” theory in the sense that it incorporates and synthesizes several diverse approaches to the MNE. 53 Dunning begins his analysis by identifying 3 types of ownwership (firm- specific) advantages: 1) those which any firm possesses over another operating in the some location such as exclusive access to factors, size, or exclusive rights to intangible assets, 2) the advantages that a branch may have over a de nOvo enterprise or one breaking into a new product area, and 3) those specifically associated with multinationality such as the ability to exploit differences in factor and market conditions between countries. He argues that without one or more of these, the firm will be unable to overcome the costs of operating in a foreign market and will not engage in foreign production. But while ownership advantages provide a necessary condition for multinationality, they are not in and of themselves a sufficient condition. In addition, the foreign location must hold some particular attraction stemming from its unique factor endowments and/or market characteristics; and internalization must be seen as the best means of exploiting that opportunity.
Dunning carefully examines the forces which encourage internalization and notes that imperfections in external resource allocation mechanisms
(particularly those arising from the price system and government intervention) provide the greatest incentive to internalize a particular activity. He then distinguishes between 2 types of imperfections; structural (which occur when economic rents are earned from barriers to competition, transaction costs are high, or economies cannot be fully realized), and cognitive (which involve restrictions or costs on information), and notes that internalization can be caused by either of these. In general, Dunning considers government intervention in the marketplace, the public good aspects of technology, and differences in national economic policies to be the major source of market imperfections. 54 According to Dunning, the theories of trade and FDI based on
Technology and Knowledge, and the Monopolistic theories based on market imperfections are actually complementary aspects of an "eclectic” theory of international involvement. Without the incentive to internalize the sale and production of Knowledge, FDI would give way to licensing or contractual exchanges; while without the incentive to internalize market imperfections
(through vertical or horizontal integration) transactions would occur between independent firms. Further, the competitiveness of a particular country will be a result of the ownership endowments of its firms, its locational endowments relative to other countries, and the cost of moving goods and services from one country to another. Finally, Dunning explores 3 additional aspects of MNE activity: the MNE’s ability to serve as an integrating force leading to the more efficient use of resources in some cases; the MNE’s tendency to concentrate certain activities in specific locations (particularly the centralization of R&D in the home country) and different motives for FDI.
Many of these ideas are developed further in “Explaining Changing
Patterns of International Production: In Defence of the Eclectic Theory” (1980).
In this article, Dunnning provides a more detailed analysis of the ownership- specific, location-specific, and internalization advantages which drive international production, and notes that these advantages may change over time. He also examines 6 specific reasons why the MNE might “internalize” the market mechanism. They are: to reduce transaction costs, to lower buyer uncertainty, to gain a competitive advantage by controlling factor or market access, to exploit or protect against governement intervention, to protect ones property rights, and to achieve economies of scale. 55 In a recent article, Dunning (1988a) presents an even more elaborate version of the Eclectic Paradigm. By this time ownership, location, and internalization advantages are simply “O il” advantages, and ownership advantages had been divided into asset and transaction based. Dunning uses the theory to explain the 3 main forms of international production: market seeking (import substiting), resource seeking (supply oriented), and efficiency seeking (rationalized investment). He also argues that the Eclectic Paradigm can be used to explain different forms of International involvement (such as joint ventures and contractual arrangements), the geographic locus of decision making, the process of divestment,5 and the consequences of MNE activity on the home and host country as well. Thus, in its latest version, the Eclectic
Paradigm provides a comprehensive framework for analyzing the conditions and motives surrounding the foreign production decision and the emergence of the MNE.6
Boddewyn (1988) has recently attempted to incorporate the political aspects of MNE behavior into the Eclectic Paradigm. According to Boddewyn, political elements should be included in a discussion of firm specific, internalization, and location advantages since firms may construct market imperfections through political behavior in order to raise rents and create barriers to entry. Further, Boddewyn argues that Dunning’s theory only describes the conditions necessary for FDI, while a complete theory of MNE behaviour must also explain the motives and circumstances surrounding the
FDI decision. Within this context, it is not enough to focus on the MNE’s economic goals; its political goals must be considered as well. Accordingly,
Boddewyn notes that political means are often used to achieve economic goals, 56 and he suggests that governments, labor, consumers, interest groups and others can be targets of the MNE’s political influence. Thus, by incorporating a politico-economic and socio-cultural perspective into the paradigm, Boddewyn is able to explain non-equity modes of foreign involvement and social welfare issues as well a s FDI.
But not all scholars are as enthusiastic about the Eclectic Paradigm. For example, Itaki (1991) points out 4 weaknesses in Dunning’s theory which considerably reduce its explanatory power. First, Itaki notes that many
“ownership advantages” are largely redundant since they would not exist or could not be exploited without “internalization”. Next, he argues that “ownership advantages” and “location advantages” are often inseparable since they are determined simultaneously by location factors. He also argues that “location advantages” are subject to ambiguities, and suggests that the theory should explicitly consider foreign exchange rates and differences between nominal and real terms. Finally, he discusses some methodological problems that may arise when factor analysis is applied to the 3 main types (OLI) of advantages.
This occurs because the determinants of each advantage are often closely related and too numerous to allow the proper application of this methodology.
Many of Itaki’s arguments seem to be based on the notion that the
Eclectic Paradigm and Internalization Theory are essentially the same thing.
While it is true that internalization plays a major role in Dunning’s theory, the
Eclectic Paradigm provides a much richer conceptual framework for considering the questions of foreign production, FDI, and the MNE. In fact, the Eclectic
Paradigm has provided the basis for much of the research on the MNE over the past 15 years; and therefore, despite some potential flaws, it remains the most successful attempt to develop a general Theory of the MNE to date. 57 VIII. Transaction Cost Analysis
As noted above, Buckley did not include Transaction Cost Analysis (TCA) in his 1981 review of MNE theories. There are 2 possible reasons for this: first, at the time of Buckley’s article, TCA per se had not been widely applied to
International Business research; and second, it could be argued that TCA and
Internalization Theory are quite similar since transaction costs are often used to explain the internalization process. On the other hand, (Rugman (1980, 1986) notwithstanding) Transaction Cost Analysis offers a unique perspective on the incentives which encourage the firm to internalize a particular market; and therefore, should be considered a separate approach to the MNE.
This is clearly illustrated in a 1981 article by Teece (1981). Teece argues that the MNE represents a response to 3 fundamental incentives: circumventing taxes or controls, gaining monopoly power, and increasing efficiency; and notes that most researchers have focused on the monopoly power aspect of the MNE and ignored its efficiency aspects. Using arguments developed by Williamson
(1975), Teece concludes that MNEs typically result from the internalization of the markets for intermediate goods, Knowledge, and Capital. The internalization of intermediate goods markets often involves vertical FDI (i.e. vertical integration) which can be efficiency enhancing since it eliminates many of the difficulties associated with contracting. But vertical integration is not costless, since it also incurs additional management and control costs, and may create certain distortions such as internal procurement bias, internal expansion bias, and program persistence bias.
On the other hand, internalizing failures in the market for Knowledge often leads to horizontal FDI (i.e. horizontal integration). This occurs in response to problems arising from the contractual transfer of Knowledge, particularly 58 buyer uncertainty, disclosure, and enforcement costs. Finally, Teece notes that the internalization of international capital markets can result in either vertical or horzontal FDI. Teece also examines the anti-competitive consequences of
MNEs and argues that vertically integrated MNEs can indeed stifle competition by creating entry barriers and circumventing local taxes and tariffs through transfer pricing schemes. On the other hand, horizontally integrated MNEs may actually enhance competition when their investments represent a new entrant or do not create a monopoly position.
This analysis is extended in a 1986 article (Teece, 1986) which attempts to specify which transactions will be executed in the market and which will be internalized within the MNE. Teece notes that markets will be more efficient than firms when transactions involve a large number of buyers and sellers, and less efficient when transactions are highly uncertain, complex, or between few parties. He also discusses the similarities between Transaction Cost Analysis and Internalization Theory, and notes that while both theories view the MNE as a response to market failures, TCA views the transaction (not the firm) as the basic unit of analysis and focuses on the mode of governance which most efficiently handles a particular transaction. According to TCA, the choice between internal (i.e. the firm) and external (i.e. the market) governance depends on asset specificity and the threat of opportunism. As a result, Teece concludes that the MNE provides the most efficient governance structure when specific assets give the firm a competitive advantage that is best exploited through foreign production, and best transfered through internal markets.7
Casson (1982) applies a slightly different aspect of Transaction Cost
Analysis to the MNE. He notes that since all market making activities incur costs and involve risks, economic efficiency calls for the specialization of these risks. 59 His main point is that the needs for quality control and to monitor market exchanges may lead firms to internalize the exchange process through backward integration. Casson identifies 4 reasons why a firm might operate in more than one country: 1) to achieve economies of scale, 2) to reduce buyer uncertainty for internationally mobile customers, 3) to accommodate buyers that may want to place orders in one country for delivery in another, and 4) to exploit an internationally transferable advantage through proprietary knowhow in product differentiation. Using a transaction oriented approach, Casson argues that MNEs will be strongly represented in consumer markets where quality is crucial. He uses Dunning and McQueen’s (1982) study of the hotel industry and
Read’s (1983) study of the bananna industry to show that MNEs are indeed concentrated in the high-quality, high-price end of the market. Based on this analysis, Casson concludes that the MNE represents a rational response to the problems associated with arms-length exchanges in the marketplace.
More recently, Hill and Kim (1988) have used Transaction Cost Analysis to develop a dynamic model of the MNE. They focus on the choice between two ways of serving foreign markets: the wholly owned subsidiary (WOS), and licensing. In the case of licensing, 2 transaction costs are paramount: contracting costs, and costs arising from the dissemination of knowledge. On the other hand, Hill and Kim identify 4 costs associated with internalization: 1) the capital costs of establishing an overseas presence, 2) the cost of learning about the local market, 3) the cost of internally transferring know-how, and 4) the cost of controlling the expanded organization. The authors then develop mathematical expressions for these costs and analyze their impact in a dynamic setting. They conclude that in the case of a single product, both internalization costs and economic benefits decline over time, implying that at some point a 60 firm initially choosing a WOS may find it profitable to switch to licensing. In
addition, a firm’s monopoly in Knowledge will be eroded by competitors over
time, which may also make licensing profitable at some point. Based on this
analysis, Hill and Kim predict that marketing based MNEs will be slower to
switch from licensing to a WOS than technology based MNEs, but will switch
from a WOS to licensing more quickly than technology based MNEs. This
occurs because the costs (relative to benefits) associated with internalizing
marketing knowledge are greater than the costs of internalizing technical
knowledge.
Thus, regardless of whether Transaction Cost Analysis and
Internalization Theory offer unique approaches to the MNE, TCA clearly
provides a number of unique insights into the firm’s decision to internalize a
particular market.
IX. Economies of Scale
While most authors recognize the importance of Economies of Scale
(and scope) as a “firm-specific advantage”,8 its role as a motive for
multinationality has been largely ignored. However, in some cases economies
of scale can indeed provide an incentive to expand into foreign markets. The
logic behind this approach is simple: scale economies offer the firm a significant
competitive edge in the form of lower production costs; however, in order to take
advantage of these benefits, the firm must be able to increase its sales
accordingly. Since this often requires expanding horizontally into foreign
markets, the desire to achieve economies of scale may also furnish a rationale
for multinationality. Even though Vernon (1966), Kindleberger (1969), Dunning
(1980), Casson (1982) and many others have noted this possibility in passing, a true Theory of the MNE based entirely on economies of scale has not yet 61 emerged. Instead, economies of scale are generally considered to be a technology- or industry-spedfc factor.
For example, Zengage and Ratcliffe (1988) consider the role scale economies played in helping Japanese firms gain global dominance in the semiconductor and consumer electronics industries (1988, Chapter 2). Their analysis suggests that the ability of Japanese firms to compete successfully in a number of large national markets (i.e. the U.S., Japan, and the markets of
Western Europe) allowed them to achieve economies of scale in production and thus become low-cost producers. This effectively blocked other international firms from competing in these industries, and (coupled with economies of scope) probably accounts for the dominant position Japanese companies hold in a number of global industries. Clearly, as more and more industries and markets move toward global standardization, the competitive advantages associated with economies of scale will become even more important. This has led authors such as Hout, Porter and Rudden (1982), Levitt
(1983), and Ohmae (1986) to advocate the use of “global” strategies9 in those industries characterized by significant economies of scale. In these cases, the rising “minimum efficient scale” of production has made multinationality an essential prerequisite for successful international competition.
Further, Krugman (1990) suggests that the need to achieve scale economies provides a compelling motive for international trade as well. When one combines this approach with the concept of “internalization”, these arguments provide the basis for a Theory of the MNE. Thus, while economies of scale are often cited as a major advantage of the MNE, they may also be seen as a cause of multinationality in certain instances. 62 X. Another Classification Scheme
Since research on the MNE has been widely reviewed10 it is not
surprising that a number of schemes have been proposed to classify this
literature. However, one in particular, deserves separate mention, Calvet’s
(1981) “A Synthesis of Foreign Direct Investment Theories and Theories of the
Multinational Enterprise”. The importance of Calvet's (1981) review lies in its
attempt to synthesis theories of FDI and theories of the MNE. This is necessary
because, even though FDI and the MNE are inextricably linked, these theories
represent distinct branches in the literature which address slightly different aspects of a common problem.
Caivet begins his analysis by identifying 3 stages in the development of
FDI Theory. The first stage, beginning with Hymer, linked FDI to industrial organization; the second, originating with Harry Johnson (1970), investigated the welfare implications of FDI; while the third, beginning with McManus (1972),
Buckley and Casson (1976) and Magee (1976), focused on the MNE as the institution engaging in FDI. He then identifies 4 incentives for FDI: 1) market disequilibrium, where FDI acts to bring internationally segmented markets into equilibrium, 2) government imposed distortions, where government policies create unstable conditions which encourage FDI, 3) market structure distortions, where market imperfections are the result of monopolistic or oligopolistic market charateristics and FDI is the result of industrial organization, and 4) market failure imperfections, where resources are allocated inefficiently due to external effects, public goods, and economies of scale.
Next, Caivet examines the 3 major strands of MNE Theory. They are:
Appropriability Theory, Internalization Theory, and Diversification Theory. Since
Internalization and Diversification have already been discussed, only 63 Appropriability Theory will be examined here. This approach is best illustrated by the work of Magee (1976, 1977) and combines the Industrial Organization approach of Bain (1956) with the Neoclassical notion of the appropriability of returns. Magee focused on the importance of Knowledge in generating appropriable returns, and identified 5 stages in which MNEs create information:
1) new product discovery, 2) product development, 3) product function creation,
4) market creation, and 5) appropriating rents. He notes that MNEs are more successful at appropriating rents from sophisticated technologies transferred within the firm, than from simple technologies transferred in the marketplace.
Thus, according to this approach, the incentive to appropriate excess returns arising from the generation of various types of Knowledge serves as the rationale for the MNE.
Finally, Caivet considers the Markets versus Hierarchies approach (i.e.
Transaction Cost Analysis) and Dunning’s Eclectic Paradigm. Caivet makes the interesting point that Transaction Cost Analysis replaces the technological bias of FDI and MNE Theory with a transactional framework based on human and organizational behavior. However, the most important aspect of Calvet’s review is that it clearly distinguishes between theories of FDI and theories of the MNE and notes that both perspectives are necessary to truly understand the complex operations of the Multinational Enterprise.
XI. Conclusions
This brief review has a number of implications for the present study.
Perhaps the most important of these is the central role Internalization plays in the Theory of the MNE. Developed by Coase (1937) to explain the emergence of the firm in a market economy, Internalization was used by Buckley and
Casson (1976) to explain the activities of the MNE. According to Buckley and 64 Casson, the MNE represents a response to certain imperfections in international markets which favor the replacement of the market mechanism with the internal market of the firm. Since a firm’s internalization of a market that crosses national borders necessarily creates an MNE, internalization clearly provides a Theory of the MNE. Although most of the theories discussed above use some aspect of Internalization to explain the actions of the MNE, each approach differs considerably in its explanation of why internalization occurs. In fact, it is possible to identify 3 distinct answers to this question in the literature: one based on Environmental factors, another which focuses on the
Monopolistic advantages of firms, and a third based on Transaction Cost
Analysis.
In the first case, unique factors in the international environment provide an incentive for the firm to internalize a particular market. These factors may be location-specific such as raw materials, labor, markets, or any other factor which contributes to serious market imperfections. While most writers have concentrated on imperfections caused by government intervention, market imperfections can arise from peculiar social or cultural factors, a nation’s level of development, and a number of other forces as well.
On the other hand, a firm’s desire to gain or exploit a Monopolistic advantage can also provide an incentive for internalization. This approach is primarily associated with Hymer’s notion of “firm-specific” advantages, but it also plays a role in Knickerbocker’s Oligopolistic Reaction theory, Vernon’s
Product Life Cycle hypothesis, and the discussion of Economies of Scale. The final set of factors stems from Transaction Cost Analysis and involves costs associated with contracting, uncertainty and bounded rationality, asset specificity, opportunism, and small numbers of market participants. In each 65 case, these factors can create an incentive to replace the market mechanism
with the internal governance structure of the firm. Although these transaction
costs are not unique to international markets; they may be considerably higher
in this case due to the many difficulties involved in conducting international
transactions.
Dunning's Eclectic Paradigm is the only Theory of the MNE which
formally incorporate all 3 of these perspectives. Dunning argues that the
decision to engage in foreign production (and hence become an MNE) is based
on 3 factors: the ownership-specific advantages of firms, the location-specific
advantages of countries, and the benefits derived from internalization. These
advantages correspond to the Monopolistic, Environmental, and Transaction
Cost rationales respectively. Thus, Dunning’s paradigm integrates the 3 major
strands of MNE Theory into a comprehensive framework for analyzing the firm’s
decision to internalize a particular market.
The following is a summary of the most important aspects of this section:
1) Based on arguments developed by Coase (1937) and Williamson (1975), the Firm and the Market can be seen as alternative methods of organizing economic activities. The choice between these two alternatives depends on a set of factors which influences the relative cost and efficiency of each mode.
2) Internalization Theory, as developed by Buckley and Casson (1976), extends these arguments to an international context and applies them to the MNE. According to this theory, the MNE represents a response to certain incentives which encourage the firm to replace international markets with its own internal markets. When a firm “internalizes” a market across national boundaries it becomes an MNE.
3) It is possible to identify 3 distinct types of incentives which encourage internalization: those based on Environmental factors, those arising from the Monopolistic advantages of firms, and those based on Transaction Cost Analysis. 4) In terms of the current study, it is important to recognize that Environmental factors can and do influence the firm’s decision to use internal or external markets. Accordingly, a significant relationship can be hypothesized between certain environmental factors and the firm’s decision to internalize a particular market. 67 Organizational Theory I. Introduction
The purpose of this section is to examine the application of several concepts from the field of Organizational Design and Behavior to International
Business research. As noted earlier, the current study draws heavily upon the work of Chandler (1962) and Lawrence and Lorsch (1967), particularly their research into the relationship between the firm, the environment, and organizational performance. Therefore, studies which explore the relationship between international strategy and the structure of the MNE or which adopt a
Contingency approach to the MNE will be featured in this section. However, before proceeding with this discussion, a brief overview of the extensive literature on Organizational Theory will be presented.
Van de Ven and Joyce (1981) identify 7 basic research programs which have profoundly influenced the modern development of Organizational Theory.
They are: the Sociotechnical system, the Quality of Work Life, the Aston School,
Decision Making under Uncertainty, Organization Assessment, the Organization and Environment, and Markets and Hierarchies. Each of these programs is briefly discussed below:
1) The Sociotechnical System: This research was initiated at the Tavistock institute of Human Relations in London, and is perhaps best associated with the work of Eric Trist. Sociotechnical theory has 2 basic premises; first, that organizations consist of both a social and technical system, and that performance is a result of the relationship between these two aspects; and second, that the organization is influenced by culture, values, and other environmental factors, and it is therefore necessary to understand these forces in order to understand the organization.
2) The Quality of Work Life: Research in this area originated with Rensis Likert and the group at the Institute for Social Research at the University of Michigan. The basic purpose of this program was to conceptualize and develop instruments to measure key characteristics of organizations and the attitudes of 6 8 their members.
3) The Aston School: This program originated at the University of Aston in England and is often identified with Derek Pugh and David Hickson. It represents one of the first attempts to rigorously explore various aspects of organizational structure, and has accordingly been called the “structuralist” approach.
4) Decision Making Under Conditions of Uncertainty: This approach, often associated with the Carnegie School, is best illustrated by the work of Herbert Simon and James March, which examines the impact limitations in the decision making process (particularly the effects of bounded rationality and uncertainty) have on organizational behavior and design.
5) Organization Assessment: This research, conducted by Van de Ven and his colleagues, attempts to explain and predict organizational performance based on the design and structure of the organization in relation to the environment it faces.
6) The Organization and Environment: This research originated with Lawrence and Lorsch at Harvard and provided the basis for Contingency Theory. As noted earlier, the Contingency approach argues that "...if a firm is to be successful, its design should be contingent upon the characteristics of the environment in which it operates.” (Van de Ven and Joyce, 1981, pg. 10).
7) Markets and Hierarchies: This approach is associated with the work of Williamson, and is based on the premise that Markets and Hierarchies represent alternative means of organizing economic transactions.
In addition, Van de Ven and Joyce discuss 3 approaches which are
shaping the current direction of organizational research. These are: the Natural
Selection or Population Ecology models of Hannan and Freeman (1977) and
others, the Symbolic and Phenomenological views of Organizational Reality suggested by Weick (1979), and the Political Economy and Marxian views of
Social Conflict and Change advanced by Burrell and Morgan (1979).
Doz and Prahalad (1991) develop a somewhat different scheme for classifying organizational research. In a recent article assessing the 69 contributions of Organizational Theory to the field of International Strategic
Management, they divide the literature into 4 major categories: Economic
Theories of the Organization (which include Transaction Cost Analysis and
Agency Theory); Environmental Adaptation Theories (such as Population
Ecology, Institutional Theory, and Contingency Theory), Theories of Power
Relationships and Organizational Adaptation; and Theories of Organizational
Learning. A brief discussion of each area appears below.
1) Economic Theories of the Organization: Both Transaction-Cost Analysis and Agency Theory focus on the costs involved with organizing economic activities; however, Transaction Cost Analysis examines the external costs of using the marketplace, while Agency Theory looks at the in-ternal costs of controlling the organization. This latter approach focuses on the issues of control and compliance arising from the contractual arrangements between principals (i.e. shareholders or owners) and their agents (i.e. managers), and is often associated with the work of Jensen and Meckling (1976) and Farma and Jensen (1983).
2) Environmental Adaptation Theories: In general, these theories examine the relationship between the organization and its environment. The Population Ecology model (Hannan and Freeman, 1988) uses the concepts of evolution and natural selection to explain how and why organizations thrive or wither. According to this view, the distribution of resources in the environment creates certain “niches” which support various “populations” of organizations. As environmental conditions change, populations which possess the appropriate traits are “selected” for continued survival, while others must adapt or face extinction. Thus, the role of managerial decisions in affecting the firm’s long term performance is minimal. On the other hand, Institutional Theory (Meyer and Rowen, 1977) underscores the role of the organization in determining its own fate. According to this approach, organizations can and do adapt to environmental conditions through the internal and external processes and relationships they develop. Finally, Contingency Theory stresses the link between an organization’s relationship with the environment and its performance.
3) Power Relationships and Organizational Adaptation: Based on the premise that the organization is a collection of individuals who form networks of power relationships (Crozier, 1964), this approach stresses the importance of individual actors in determining the actions of organizations. Accordingly, organizations are not seen as rational entities formed by deliberate actions, but 70 as the consequence of the power of certain individuals, acting within an accepted set of rules, to shape organizational outcomes. On the other hand, Resource Dependency Theory (Pfeffer and Salancik, 1978) focuses on the organization’s external relationships with the environment. By analyzing the organization’s need for external resources, this approach attempts to manage the organization’s dependencies in a way which guarantees its continued survival.
4) Theories of Organizational Learning: This approach stems from the work of Argyris and Schon (1978) and stresses the adaptive capabilities of organizations; in particular, their ability to “learn” from or respond to environmental treats and opportunities by creating or modifying behavioral and/or structural processes. Since learning can take place on both the macro (organizational) and micro (individual) level, different theories have identified a wide range of organizational processes which affect learning.
II. The Stopford and Wells Model
Based on the above review, it is possible to distinguish 2 general approaches to Organizational Theory: one that focuses on the structural or inanimate properties of organizations, and another that emphasizes the behavioral characteristics of the organization’s constituents. In terms of the current study, research that utilizes the former approach is of particular interest.
Chandler (1962, 1968) was one of the first scholars to investigate the relationship between the firm and the environment through his historical analysis of the evolution of the modern corporation. He found that, not only was there a clear relationship between environmental forces and a firm’s strategy, but there was also a relationship between a firm’s strategy and its organizational structure. Chandler captures this insight in the famous dictum,
“structure follows strategy..." (1962, pg. 14).
Stopford and Wells (1972) were perhaps the first to apply Chandler’s analysis to the Multinational Enterprise. Their study attempts to answer 2 major questions: first, how have MNEs altered their structure in order to implement 71 complex strategies; and second, what factors influence the firm’s decision to include local partners in their foreign operations. Based on an investigation of
187 U.S. MNEs, the authors conclude that there is a clear relationship between an organization’s strategy, and its structure and ownership preference.
However, unlike Chandler, Stopford and Wells note that a firm’s stategy can both influence and be influenced by its structure. More importantly, the authors found that firms tend to follow a similar pattern of structural change as they expand into foreign markets.
Based on Chandler’s (1962) discovery that firms move from a “functional” to “divisional” (M-form) structure as the size and complexity of their operations increase, Stopford and Wells attempt to describe the structural changes that accompany an increase in foreign activities. Much to their surprise, they found that firms often give their first few foreign subsidiaries a great deal of autonomy rather than incorporating them into a formal organizational structure. They attributed this to the defensive nature of many foreign investments and to the firm’s initial lack of a clear cut foreign strategy. However, as the number of foreign subsidiaries increased, the authors found that firms typically added an
International Division to their existing divisional structure. This enables the firm to coordinate its numerous foreign subsidiaries, but may become increasingly ineffective as the level and complexity of the firm’s international activities increases. Consequently, the International Division is often replaced by a true
"global” structure at some point.
Stopford and Wells identify 3 common global structures (Worldwide
Product, Area, and Mixed) and note that each of these provides the benefits of integrating and coordinating activities on a worldwide scale.11 In general,
Worldwide Product structures appear in firms with highly diversified foreign 72 product lines and represent an attempt to coordinate the firm’s foreign
operations with its domestic operations. On the other hand, Area structures are
typically associated with low product diversity and/or mature products and allow
activities to be integrated on a regional basis. Unfortunately, this focus on
regional integration often comes at the expense of global coordination. Thus,
while some firms have succeeded in rationalizing production on a regional
basis, this usually requires major structural adjustments to balance the
conflicting objectives of centralized production and decentralized marketing.
Mixed structures incorporate elements of several other organizational forms and
often represent a transitional phase between 2 identifiable structures, or a
response to some novel strategy. Finally, the authors examine the Grid (i.e.
Matrix) structure which represents an attempt to combine the benefits of worldwide product coordination and regional concentration into a single organizational form.
In general, Stopford and Wells found that structural changes are designed to eliminate discrepancies between a firm’s strategy and its existing structure. They also note that appropriate structures are typically associated with superior performance. However, their most important findings are contained in a model which describes the structural changes that accompany an increase in foreign activities. This model suggests that firms generally move from an International Division (ID), to a Worldwide Product (WP), Area, and then
Matrix structure as the firm’s level of product diversity and foreign involvement increase. Thus, Stopford and Weils were able to establish a strong relationship between specific aspects of international strategy (product diversity and the ratio of foreign to total sales) and the structure of the MNE.
Stopford and Wells’ study has inspired numerous subsequent 73 investigations into the relationship between strategy and structure in the
Multinational Enterprise. For example, Daniels, Pitts and Tretter (1984) explore
the relationship between several key elements of international strategy
(including product diversity, level of foreign involvement, reliance on R&D,
marketing and capital intensity, and ownership and control of foreign
operations) and organizational structure in a sample of 256 U.S. MNEs.
Although Daniels’ findings generally support the Stopford and Wells model,
there are a number of important differences. For example, while Stopford and
Wells argued that firms with high product diversity move from an International
Division to a Worldwide Product structure as their foreign involvement
increases, Daniels, et. al. found that at high diversity levels, dependence on
foreign operations did not significantly differentiate organizational structures.
The variable that best predicted whether a high diversity firm will use the
WP or the ID structure was its mode of entry. Diversified firms that expanded
through acquisition tended to use the Worldwide Product structure, while high
diversity firms with high R&D intensity used the international Division. In
general, the authors found that the International Division was a special structure
used by companies that organize their domestic operations along product
divisions. They also found that firms using an ID had a higher dependence on
foreign sales than those using WP structures, contrary to Stopford and Wells’
findings; and that functional structures did not give way as a firm’s foreign
involvement increased, as Chandler (1975) had hypothesized.
The authors use this information to update the Stopford and Wells model.
According to Daniels, et.al., as long as foreign sales are low, most companies
handle their foreign operations as an appendage to existing product or functional divisions. Whether a firm with low foreign sales adopts a Worldwide 74 Functional (WF) or Worldwide Product structure depends primarily on the level of product diversity. Low and medium diversity firms choose the WF structure while medium and high diversity firms use the WP. Finally, as a firm’s dependence on foreign sales increases, it may induce a shift from WP to ID, contrary to Stopford and Wells’ prediction. The authors argue that the centralization of resources may explain why firms adopt an ID structure; and note that the ID may promote foreign expansion more effectively than the WP structure, by placing a spokesman for geographic interests at the same level as spokesmen for preduct and functional interest in the firm's hierarchy.
A recent study by Egeihoff (1988) also attempts to revise the Stopford and Wells model. Based on a survey of 50 U.S. and European MNEs, Egeihoff concludes that a third variable (the level of foreign manufacturing) must be incorporated into the model in order to explain the MNE’s choice of an organizational structure. According to Egeihoff, when a firm’s foreign sales are low, the International Division will be used when foreign product diversity is low, and the Worldwide Product structure will appear when product diversity is high.
On the other hand, when foreign sales reperesent a high percentage of a firms activities, both product diversity aod. the percentage of foreign manufacturing are necessary to identify the appropriate structure. When a firm has high levels of both product diversity and foreign manufacturing, a Product-Area Matrix is required; while the WP is appropriate for high diversity firms with low foreign manufacturing, and the Area structure is used when foreign manufacturing is high, but product diversity is low.
Thus, based on these 3 studies, 4 elements of international strategy emerge as crucial determinants of organizational structure: product diversity, percentage of foreign sales, mode of entry into foreign markets, and the 75 percentage of foreign manufacturing.
III. Other Strategy and Structure Studies
Egeihoff has been one of the few International researchers to consistently explore the relationship between international strategy and MNE structure. In a 1982 study, Egeihoff develops “strategic profiles” of several common MNE structures which specify the critical “fits” between various elements of strategy and structure. Based on a study of major MNEs, Egeihoff concludes that the Worldwide Functional structure is appropriate for firms with
“a narrow and highly consistent worldwide product line, a limited number of foreign subsidiaries, a low level of outside ownership and few foreign acquisitions.” (1982, pg. 453). On the other hand, the International Division is suitable for firms with reletively small foreign operations, while the Area structure fits a firm with “large foreign operations and a high percentage of foreign manufacturing.” (1982, pg. 454). Finally, Worldwide Product structures apply when firms have large foreign operations in addition to high levels of product diversity and product change.
While the preceding study focuses on the strategic characteristics of specific MNE structures, a 1984 study (Egeihoff, 1984) explores the structural properties associated with various patterns of “control”. Specifically, Egeihoff examines 2 types of control commonly used by MNEs: performance reporting systems (i.e. output control), and the assignment of parent company managers to foreign subsidiaries (i.e. behavior control). Using a sample of large MNEs from the U.S., U.K., and Europe, Egeihoff concludes that subsidiaries of U.S.
MNEs submit more performance data than their European counterparts, while
European MNEs fill more marketing and manufacturing positions with parent company nationals. Thus, U.S. firms tend to exercise higher levels of output 76 control over their foreign subidiaries, whereas European firms exercise relatively higher levels of behavior control. In addition, Egeihoff found that subsidiaries in turbulent environments (Brazil) tended to have significantly higher percentages of parent company national in key positions than subsidiaries in stable environments (Europe), while larger subsidiaries tended to have greater output control over their manufacturing operations. Finally, older subsidiaries and the subsidiaries of firms with considerable foreign experience had less financial output control.
Egeihoff notes that a number of unique structural characteristics are associated with each type of control. For example, U.S. MNEs tend to measure more quantifiable and objective aspects of a foreign subsidiary and its environment, whereas European MNEs tend to measure more qualitative aspects. Consequently, control in U.S. MNEs requires more precise plans and budgets, while control in European MNEs requires a higher level of company’ wide understanding about what constitutes appropriate behavior. In addition, control in U.S. MNEs requires larger central staffs and more centralized information processing capacity, whereas European MNEs require a larger cadre of capable expatriate managers. Finally, control in European MNEs involves greater decentralization of decision-making, and favors short, vertical reporting channels between positions in the foreign subsidiaries and in the parent.
Finally, a study by Gates and Egeihoff (1986) examines centralization
(i.e. the level at which a decision must be approved before it can be implemented) in headquarters-subsidiary relations. Using a sample of large
U.S. and European MNEs, the authors found a consistent relationship between centralization and various measures of company complexity. Specifically, they 77 found that MNEs which introduce more product lines into foreign markets or
modify products to fit local environments require more decentralization, as do
MNEs that allow substantial outside ownership of foreign operations or grow through acquisition. Further, the longer an MNE has been operating abroad, the more decentralized its parent-subsidiary relationships appear to be. On the other hand, the negative relationship often posited between size and centralization was much less apparent, with European and U.K. firms decentralizing as the relative size of their foreign operations increased, but not
U.S. MNEs. In fact, the study uncovered a strong positive relationship in U.S.
MNEs between marketing centralization and the size of the MNE, the size of the subsidiary, and the relative size of the subsidiary. Finally, although previous studies had hypothesized that subsidiaries facing frequent change in local product lines and competitive local climates would be given more autonomy,
Gates and Egeihoff found little support for this assertion.
In terms of the present study, the most important conclusion that can be drawn from this brief review is that a strong empirical relationship can be established between specific elements of a firm’s strategy and corresponding aspects of its organizational structure.
IV. Contingency Theory
In many ways, Contingency Theory represents a logical extension of the
Strategy-Structure approach by suggesting that performance is a consequence of the “fit” between an organization’s structure and the environment. Not surprisingly, this view has had a major impact on the field of International
Strategic Management. In fact, the concept that performance can be improved or maximized by aligning an organization’s strategy, structure, or management style with the environment is so deeply ingrained in the International Strategy 78 literature that few writers even acknowledge their debt to Contingency theory.
However, Doz and Prahalad (1991) recently noted that, “the contingency theory
of organizations, and its emphasis on differentiated responses to diverse
environments and integration of actions across environments, has had the most
direct impact of all strands of organization theory on [MNE] management
research...” (1991, pg. 151).
A number of studies have attempted to describe the relationship between
strategy, structure, and performance. As noted above, Stopford and Wells
(1972) concluded that appropriate organizational structures were often
associated with superior performance. A study by Rumelt (1974) also
established a strong relationship between organizational structure,
diversification strategy and firm performance.12 Using a large sample of Fortune
500 firms, Rumelt found that significant differences in performance could be
attributed to various diversification strategies, as well as to various
organizational structures. (1974, pgs.146-152)
A more recent study by Davidson (1984) examines the relationship
between administrative orientation and international performance. According to
Davidson, diversified firms can administer their international operations in 3 ways: permit a high degree of autonomy in individual units (as in a holding company), use a fixed, standard approach in all units (as in a conglomerate), or establish a unique approach for each unit (typical of diversified industrial firms).
Using a sample of large firms, he found that significant differences in
performance were associated with each of these approaches. In general,
holding companies and conglomerates exhibited poorer foreign sales and profit performance than diversified firms. Davidson also explored the relationship between centralization and performance, and found that firms with centralized 79 management had lower levels of foreign sales and profit growth, while decentralized firms exhibited the highest rates of foreign sales growth.
As a result of the empirical evidence linking performance to the relationship between the firm and the environment; a number of writers have attempted to develop strategies which maximize performance under certain conditions. As Davidson (1984) notes, “A recent trend in international management research has been the development of contingency models that identify optimal planning and management modes for international business units, given key characteristics of the product, market, and industry.” (1984, pg.
11). This is typically accomplished by identifying 2 (or more) critical dimensions which define a number of individual contextual settings and then specifying the appropriate strategy for each setting.
“Portfolio planning” models provide an excellent illustration of this approach. Wind and Douglas (1981) examine a number of these models13 and note that they offer an effective tool for allocating resources across a company’s portfolio of products, markets and businesses. One of the first important portfolio models was the Boston Consulting Group’s (BCG) Growth-Share Matrix.14 The
BCG matrix defines 4 types of products (Stars, Problem Children, Cash Cows, and Dogs) based on 2 critical factors (the product’s relative share of a particular market, and the market’s potential growth prospects). The purpose of this model is to identify the optimal allocation of resources among a firm’s product portfolio in order to generate excess cash flow for reinvestment and product development. Since cash flow is seen as a function of market share which in turn depends on investment, specific strategies can be developed for managing each of the 4 product categories based on their current cash needs and 80 potential payoffs.
Bartlett and Ghoshal (1986) propose a similar model for managing headquarters-subsidiary relations. Their model assesses the importance of a national unit along 2 dimensions: the competence of the local unit, and the strategic importance of the local environment. This defines 4 basic types of subsidiaries: the Strategic Leader, which occurs when a highly competent national unit is located in a strategically important market; the Contributor, which operates in an unimportant market but has a distinctive capability; the
Implementor, which appears in a strategically less important market and has just enough competence to carry out local operations; and the Black Hole, which operates in a strategically important market but has limited capabilities.
The authors then develop a unique strategy for managing each type of subsidiary, paying particular attention to the way in which decisions are made and implemented.
In both of these models the underlying assumption is that certain critical characteristics of the firm, product, or environment can be used to distinguished several unique contextual settings (such as a product with a high share of a high growth market, or a subsidiary in a strategically important market that has limited capabilities). Once these settings have been identified, corresponding strategies can be devloped for allocating resources among a company's portfolio of products, markets, or businesses. But the portfolio perspective has been criticized for oversimplifying complex environmental factors and providing a "cookbook” approach to strategy formation. In addition, these models are usually static in nature; and therefore, the firm’s need to identify and respond to environmental change is largely ignored. Nonetheless, Haspeslagh (1982) found that by the early 1980s, portfolio planning was widely used in large U.S. 81 firms.
While portfolio planning models represent one way in which Contingency
Theory has influenced the field of Strategic Management, Lawrence and
Lorsch’s (1967) focus on the firm’s need to achieve an appropriate degree of both differentiation (of function and task) and integration (across tasks) has had an even greater impact. Since the MNE typically operates in a number of discrete environmental contexts and may performs a wide variety of unrelated tasks, its needs for differentiation and integration are often too complex to be addressed through simple Kcontingency” models like the ones described above.
Instead, writers have analyzed this problem through 3 distinct, yet closely . related, frameworks: the need for integration versus the need for adaptation, global versus multi-domestic industries (and strategies), and the economic and political imperatives. Each of these approaches is discussed below.
A number of writers have examined the conflicting strategic pressures associated with the firm’s need to adopt to its environment and its need to integrate activities across divisions and national borders.15 For example,
Lorange (1976) argues that these pressures arise because adapting to local conditions requires the decentralization of many activities which can increase costs; while integration reduces costs by promoting economies of scale, but neglects the unique characteristics of individual markets. Thus, developing a strategy that balances the firm’s need for both adaptation and integration is a difficult task indeed. Lorange addresses this challenge by specifying the unique adaptation and integration requirements of several common organizational structures (Worldwide Product, Area and 2 Matrix structures). He then develops the optimal strategy for balancing these opposing forces in each structure,and implies that a firm can improve its performance by adopting the appropriate 82 strategy for its organizational structure.
Bartlett and Ghoshal (1988, 1989) also propose a strategy for balancing the firm’s need to adapt and integrate its operations. Based on their study of major MNEs from the U.S., Japan, and Western Europe, they note that forces which encourage integration and adaptation are working simultaneously in many industries, yet few companies have the organizational capabilities to respond to them. Specifically, firms which function as a network of independent national subsidiaries have problems responding to global forces; while companies organized as a single, integrated system can respond to global forces, but have difficulties addressing local needs. They conclude that although a company’s strategy is largely shaped by its external environment, its ability to implement that strategy is constrained by the existing configuration of its assets, the distribution of resources and responsibilities between subunits, and the firm’s historic norms, values, and management style (i.e. the firm’s
“administrative heritage). The authors propose a "transnational" solution for simultaneously achieving global coordination and national flexibility. This involves creating an interdependence of resources and responsibilities between subunits of the organization, developing a set of strong cross-unit integrating devises such as clearly defined operating systems and decision making processes, and establishing a strong corporate identity and a well defined worldwide management perspective.
The concepts of adaptation and integration are closely related to the differences between “multidomestic” and “global” strategies. Porter (1980) was among the first to recognize this distinction in his analysis of the impact of 6 critical industry factors (scale economies, transportation costs, distribution channels, comparative factor costs, the heterogeneity of market demand, and 83 government barriers to trade) on a firm’s product lines. Based on this analysis,
Porter identifies 4 basic international strategies: a global strategy based on centralization which is used when product lines face perfectly competitive markets; a decentralized, national strategy which applies when product lines are subject to economic impediments such as transportation costs or heterogeneous demand; a protected niche strategy, appropriate when governments grant preferential access to certain product lines; and a mixed strategy which combines elements of several other strategies.
These ideas are developed further in a seminal article by Hout, Porter, and Rudden (1982). Based on case studies of global competition, the authors identify 2 types of international industries: “multidomestic”, where companies pursues a different strategy in each national market; and “global”, where companies compete in terms of their integrated worldwide production and marketing systems. According to Porter (1986), examples of multidomestic industries include retailing, comsumer packaged goods, distribution, insurance, and financialal services; while global industries include aircrafts, TVs, semiconductors, autos, and watches. In general, global industries are characterized by significant economies of scale and require competitive strategies which improve a company’s worldwide cost position or its ability to differentiate products. As a result, successful global competitors manage their many national operations as a single system (not as a portfolio of businesses), and attempt to leverage positions in one market against those in other markets.
This often requires making investments with zero or negative returns, setting different financial objectives for each foreign subsidiary, and viewing national markets as /'nter-dependent rather than /7?-dependent. Unfortunately, this also places a severe strain on the company’s organizatonal structure, since product 84 and area need must be balanced with the needs for centralization and autonomy; and therefore, a global strategy Is not appropriate for every firm or industry.
The distinction between global and multidomestic industries (and strategies) has had an enormous impact on the field of International Marketing.
Following Hout, Porter, and Rudden, Levitt (1983) argued that technology was encouraging the homogenization of world markets which in turn was driving many industries toward global standardization. In such an environment, global strategies which promote economies of scale become the key to international success. Levitt’s article sparked a fierce debate concerning the role of standardization in shaping international competition. Although some writers have accepted Levitt’s hypothesis,16 many other have not,17 and continue to advocate strategies which address specific aspects of the firm, product, or market.
The final framework stems from Doz’s (1980) investigation of the economic and political imperatives. Based on this analysis, Doz identifies 3 basic strategies for resolving these conflicting forces: a worldwide integration strategy, a national responsiveness strategy, and an administrative coordination strategy. In the first case, firms integrate and rationalize their international operations in order to achieve economies of scale. This approach has the advantages of generating cost and efficiency savings, increasing the firm’s bargaining power in host countries, and simplifying the management of international operations. On the other hand, a national responsiveness strategy forgoes the benefits of integration but offers several advantages over a firm operating in a single national market. These include pooling financial risks, spreading costs over a larger sales volume, coordinating activities between 85 markets, and transfering skills between subsidiaries. The administrative
coordination strategy uses structural or administrative adjustments instead of
strategic solutions to resolve the conflicting economic and political imperatives.
In this case the strategy is to have no set strategy, and the firm trades off internal
efficiency for external flexibility.
Doz notes that integration addresses the economic imperative, national
responsiveness addresses the political imperative, and administrative
coordination can address either. The proper strategy for a particular MNE is a
function of the markets it serves, the level of international competition, and the
technology being used. For example, firms with a large share of the world
market are likely to use an integration strategy; while smaller firms are likely to
find administrative coordination the appropriate choice. On the other hand, in
industries characterized by free trade, all firms will adopt a worldwide
integration strategy; but in industries where political imperatives prevail, all firms
are likely to adopt a national responsiveness strategy. Finally Doz argues that
firms using higher technology than their competitors will strive for integration;
while lower technology firms will select national responsiveness.
Doz and Prahalad (1984, 1987) have continued to search for ways to
balance the needs for international integration and centralization of decision
making, with the needs for national responsiveness and subsidiary autonomy.
They argue that the most critical task in managing the MNE is to structure the
relationship between the headquarters and its subsidiaries. Traditional structural solutions, such as the Product, Area, or Matrix structures, are unable to address both of these needs simultaneously. This can only be accomplished by completing 3 closely related tasks: ensuring that relevent data are brought to bare on decisions, creating conditions for a consensus among key managers, 86 and managing the relative power between managers. In order to achieve these
objectives, appropriate control systems must be developed, and a structural
context must be created which provides the basis for individual decisions that
are consistent with the strategic objectives of the firm.
V. Other Contingency Frameworks
Most of the work discussed thusfar has either attempted to develop a
“contingency" model based on key characteristics of the firm and/or the
environment, or has focused on the firm’s need to simultaneously adapt and
integrate its operations. A somewhat different approach was developed by
Miles and Snow (1978). They began by noting that, in order to succeed, a firm
must “align” itself with the environment. This is done by solving 3 key problems: the entrepreneurial problem of which markets to enter and what products to
make, the engineering problem of what technologies and processes will be
used to make and deliver those products, and the administrative problem of
how to organize and manage the firm. Based on their studies of numerous organizations, the authors identify 4 strategies for solving these problems: the
Defender, the Prospector, the Analyzer, and the Reactor.
Defenders solve the entrepreneurial problem by seeking a dominant
position in a narrow but stable market. The key to this strategy is solving the
engineering problem, which is typically done by making continual
improvements in production and distribution processes. In general, the
administrative problem is solved by using centralized decision-making and formal operating procedures. On the other hand, Prospectors attempt to find and
exploit new markets and products. This requires flexibility in both production
and management; and therefore, the administrative problem is solved by encouraging intra-organizational linkages rather than centralized 87 decision-making. Analyzers hope to gain the benefits of both stability and flexibility. They do this by combining elem ents of the Defender and Prospector strategies. Finally, Reactors fail to develop a appropriate solution to the 3 problems, and simply reacting to change in an inconsistent manner.
Consequently, Reactor firms tend to perform poorly.
Several writers have applied Miles and Snow’s model to the MNE. For example, Lemak and Bracker (1988) recently proposed a strategic contingency model of MNE structure based on 3 separate factors: domain parameters, such as product diversity and dependence on foreign operations (i.e. the entrepreneurial problem); generic strategies, including volume maximization and value added (i.e. the engineering problem); and management orientation
(i.e. the administrative problem). Lemak and Bracker argue that regardless of the performance measurement used, outcomes will be substantially affected by the match between the organization and its environment. They use their model to generate hypotheses about the relationship between a particular management orientation (or generic strategy) and the type of organizational structure appropriate for a firm with a given domain parameter and/or generic strategy.
The importance of this approach lies in its attempt to identify the critical
“fits” between the organization and the environment. While Miles and Snow were able to specify 4 basic solutions to their 3 key problems, more complicated frameworks (i.e. those which include many dimensions) make the development of such standard strategies impracticable. This is clearly illustrated in a recent model by Yip (1989). Yip’s framework examines the interaction of over 20 specific firm, product, and market characteristics in an effort to determine whether a particular firm will benefit from a global strategy. Although Yip’s 88 model is capable of providing a rich analysis of the unique strategic
environment facing the firm, the model's inherent complexity has restricted its
use to the rather mundane question of whether to employ a global or
multidomestic strategy.
VI. Other Organizational Theories
Doz and Prahalad (1991) note that, with the exception of Contingency
Theory, the application of Theories of Environmental Adaptation to the problem
of managing the MNE has been minimal. However, these “biological” theories
of organizations have influenced the field of Strategic Management by stressing
the dynamic nature of the environment and the firm’s need to adapt to
environmental change. As noted above, the Contingency approach has been
criticized for providing a static analysis of these factors; but while many writers
acknowledge the need to create more dynamic strategic planning models, few
have been able to do so. In fact, the area that has contributed the most toward the development of a dynamic perspective is the literature on organizational
“learning”. Since learning is by definition a dynamic process, writers such as
Daft and Weick (1984), Kogut (1984), Ghoshal (1987), and Hamel (1991) have
helped move the focus of strategy formation away from static Contingency
models by stressing the importance of learning as a competitive advantage. In addition, many recent studies on competitiveness (of both the firm and the
nation) underscore the dynamic nature of competition. This is clearly illustrated
in Porter’s (1990) work on the competitive advantages of nations, but it is also
evident in many recent studies on the MNE.18
VII. Conclusions
A number of important conclusion can be drawn from this brief review.
The first concerns the relationship between strategy and structure. Beginning 89 with Chandler (1962), a number of writers have examined the link between the environment, a firm’s strategy, and its organizational structure. In general, they have found that strategy is a response to specific threats and opportunities in the environment, and that organizational structure follows strategy. Stopford and Wells (1972) were among the first to apply these ideas to the study of the
MNE. They developed a model which describes the structural changes that typically accompany an expansion into foreign markets. Based on this and subsequent research, it is possible to identify a number of specific strategy variables that have a direct effect on (MNE) organizational structure. These include product diversity, the level of foreign involvement, the level of foreign manufacturing, and the mode of entering foreign markets.
Contingency Theory represents the logically extension of the
Strategy/Structure approach by suggesting that performance is determined by the relationship between strategy, structure and the environment. The
Contingency approach has had an enormous impact on the field of International
Strategic Management by encouraging writers to develop "contingency” models
(frameworks) which specify optimal strategies for firms facing particular contextual settings. These models are typically based on key characteristics of the firm (such as its products, structure, or management style), the industry or market, and/or the environment. In addition, Lawrence and Lorsch’s (1967) emphasis on the firm's need to both differentiate and integrate its operations has been widely applied to the MNE. Writers have explored 3 closely related issues in addressing the MNE’s need to both adapt to local conditions and integrate its operations across products and markets; namely, adaptation versus integration (Bartlett and Ghoshal, 1989), multidomestic versus global industries
(and strategies) (Hout, Porter, and Rudden, 1982), and the political and 90 economic imperatives (Doz, 1980).
Finally, the theories of Environmental Adaptation and Organizational
Learning have extended the Contingency approach by stressing the dynamic nature of the environment and the firm’s need to adapt to environmental change. Unfortunately, due to the complexities involved in applying the concepts of evolution, natural selection, and population ecology to the study of the MNE, the development of dynamic strategic planning models is just beginning to occur.
The most important conclusions from this section appear below:
1) in general, 2 distinct approaches to Organizational Theory can be identified: one which focuses on the structural properties of the organization; and another which examines the behavioral characteristics of the organization’s constituents.
2) Beginning with Chandler (1962), a number of studies have examined the relationship between the firm and the environment. Their findings tend to support the argument that organizational structure is related to strategy in much the same way that strategy is related to environmental forces.
3) The Contingency approach developed by Lawrence and Lorsch (1967) extends this analysis by suggesting that performance is directly affected by the relationship between a firm’s structure and the environment.
4) Contingency Theory has had an enormous impact on the field of International Strategic Management by encouraging the development of “contingency* models which specify the optimal strategy for a firm facing a given contextual setting.
5) Contingency Theory has also encouraged writers on the MNE to consider 3 closely related issues: adaptation versus integration, multidomestic versus global stragies, and the political and economic imperatives. The focus of this research has often been to devise strategies which simultaneously address the MNE’s need to adapt to local conditions and integrate operations across activities and markets.
6) Other Organizational theories have also influenced research on the MNE. For example, Theories of Environmental Adaptation stress the dynamic nature of the environment and the firm’s need to adapt to environmental environmental change, while Theories of Organizational Learning emphasize the organization’s ability to adapt to the environment by creating or modifying behavioral and/or structural processes. These ideas are just beginning to be applied to the MNE. 92 Strategic Management
I. Introduction
The preceding section dealt with one aspect of recent Strategic
Management research; namely, how to manage the relationship between the
MNE and the environment. This section will explore a somewhat different aspect of this literature, research on various forms of inter-firm Cooperation (i.e. the so called “New Forms of Investment”).19 interest in this topic represents a response to 2 recent developments in the international environment: first, the widespread use of Joint Ventures, Strategic Alliances, and other Collaborative
Arrangements; and second, the apparent decline of U.S. Competitiveness in several key industries.
Most writers attribute the increased use of contractual and collaborative arrangements to the growing cost and complexity of doing business in a global environment. For example, Ohmae (1990) argues that the costs of developing new technologies and products, and the need to simultaneously enter the
World’s major markets accounts for the popularity of these “New Forms of
Investment”. But these arrangements offer the MNE a number of additional strategic advantages as well, such as reducing the cost and risk of product development, achieving economies of scale in production, entering closed markets, avoiding local government restrictions, and learning about local business conditions. On the other hand, “New Forms" can be difficult to manage and frequently end in failure.20 Nonetheless, these contractually-based arrangements provide the MNE with an efficient and effective means of conducting many international activities, while avoiding some of the problems commonly associated with FDI. 93 The research on inter-firm Cooperation has led some writers to question the validity of the Competitive model and its emphasis on arms-length, adversarial relations between the Firm and its Suppliers, Management and
Labor, and Business and Government. In addition, the phenomenal success of many Japanese and European firms and industries seems to indicate that cooperation between Business and Government can play a critical role in determining economic success. This has generated a lively debate concerning the effect of certain types of cooperative behavior on the economic performance of the firm, and the ability of Government to promote competitiveness through specific industrial Policies. Based on this analysis, there is a growing sense that
Cooperation not Competition is the superior means of structuring economic relationships. Accordingly, this section will present recent studies from the field of International Strategy which explore the role various types of Cooperative
Behavior play in creating critical competitive advantages for both the Firm and the Nation.
II. Cooperation and National Competitiveness
The relative decline of American Competitiveness in many industries since the mid-1970s has been the cause of great interest and concern.21 Not surprisingly, a large body of research has emerged which attempts to explain this phenomenon and develop effective responses to it. Among the most commonly cited reasons for the recent loss of American competitiveness are: 1)
The decline is a natural consequence of the artificial and unsustainable share of global markets U.S. firms held immediately after WWII. 2) It is the result of a natural shift in the product life cycle of major industries which favors the lower labor costs of foreign producers; or the consequence of a structural shift in the
U.S. economy out of manufacturing industries and into service industries. 3) 94 The decline is associated with the loss of competitiveness of U.S. exports due to the excessive value of the Dollar vis-a-vis other major currencies. 4) The decline can be attributed to inefficient management practices and the inability of
U.S. firms to commercialize recent technological innovations and produce internationally competitive products. Finally, 5) The decline is the result of
Government’s inability or reluctance to design and promote an effective industrial policy.22 But while writers may disagree about the precise cause of
America’s declining industrial competitiveness, most authors agree that
Competitiveness is inextricably linked to the process of Innovation.
Until quite recently, it was generally beleived that competition between firms produced a higher rate of innovation than cooperation. This view is often associated with Schumpeter’s (1942) argument that the energy and creativity inherent in Capitalist competition are the driving forces of innovation, and that inter-firm cooperation greatly diminishes these forces. However, recent studies of Japan and other European countries seem to indicate that cooperation between firms promotes greater productivity in many research and development activities than competition. Some writers have suggested that this is due to the secrecy and rivalry that typically accompany competitive behavior. These factors can greatly increase the cost of innovation by encouraging duplication of research efforts and making it difficult to diffuse and commercialize major innovations. For example, Brooks (1984) argues that U.S. anti-trust policy has had an inhibiting effect on innovation by encouraging more duplication than is optimal and generating an excess of trivial innovations. In fact, these sorts of arguments were responsible for the passage of the Cooperative Research Act of 1984 which allows U.S. firms to engage in many types of cooperative R&D 95 without facing anti-trust restrictions.23
Many of these ideas are developed further in a recent article by Jorde and Teece (1989). They note that the traditional fear of cooperation between competing firms has produced a world in which anti-trust laws inhibit certain types of inter-firm behavior. However, in a world with uncertainty, the authors conclude that there is little reason to assume that competitive markets will produce efficient economic outcomes in all cases. Further, since uncertainty is highest, and the need to coordinate investments is greatest, in the development and commercialization of new technologies, the authors note that cooperation among firms may actually improve economic welfare.
One reason for this is that competitive markets tend to underinvest in new technologies due to the high risks associated with these investments and the limited potential to extract rents from them. Jorde and Teece argue that this creates a need to promote greater coordination between firms than the price system can achieve alone. One solution to this problem is to expand the role of
Government in allocating resources by developing a clear industrial policy; however, the authors note that various forms of inter-firm cooperation can achieve many of the same objectives as government intervention. In particular, cooperation can provide an effective means of encouraging investment in new technologies, aiding the diffusion of innovations among firms, and assuring that these discoveries are developed into commercially viable products.
While Jorde and Teece stress the role of cooperative behavior in the innovation process, Vernon (1987) examines the role of Government in creating an economic climate conducive to innovation. Vernon identifies 2 areas where government action can be a critical determinant of success. These are: assuring a consistently high level of basic education, and encouraging basic research 96 and development (R&D). In both cases the end result of appropriate government policies is a higher rate of industrial innovation and improved economic performance.
According to Vernon, there are several good reasons for implementing industrial policies which stimulate R&D. For one thing, firms tend to underinvest in these activities when they are unsure about their ability to generate and/or capture profits from them. Consequently, policies which guarantee the legal rights of the inventor to profit from his discoveries can have a positive effect on
R&D investment. In addition, the market often fails to provide reliable signals about profitable research opportunities. Here again, specific policies can be designed to address this problem, such as granting tax credits, increasing military procurement programs, and encouraging joint R&D projects. In each case, government actions can be used to provide firms with the necessary incentives to make specific R&D investments when the market fails to do so.
Studies by Allen, Utterback, et. al. (1978), Abernathy and Charkravarthy
(1979), Nelson (1985), Ergas (1987), Dunning (1989) and many others underscore the ability of properly designed industrial Policies to promote technological innovation and industrial development. However, within the context of the present discussion, policies which promote 2 types of cooperative behavior have consistently been associated with innovation and economic growth; those that allow firms to engage in joint R&D activies, and those that encourage the diffusion of innovations among firms and throughout the economy. In addition, government programs which support basic research also provide a promising means of encouraging technological development, although some writers warn that America’s focus on military research has become increasingly ineffective in stimulating economic growth due to the 97 specialized nature of this research and its limited commercial spillover.
Of all the Industrialized Countries, Japan has been the most successful in promoting industrial development through the use of Industrial Policies. This is clearly illustrated by numerous studies of key Japanese industries, including
Autos, Consumer Electronics, Machine Tools, and Semiconductors. A typical example of this research is a recent study of the Japanese Machine Tool industry by Sarathy (1989). Sarathy examines the role Industrial Policy played in guiding this industry out of decline and into Worldwide prominence.
Japanese industrial policy has long been guided by MITI’s (Ministry of
International Trade and Industry) commitment to Developmental Capitalism, where the state works closely with private enterprise to achieve specific economic objectives. MITI policies are defined in “vision statements” which provide a basic outline of the steps required to improve an industry’s performance.
In the case of machine tools, MITI policies consisted of rationalizing the industry through mergers and product divestitures in an attempt to achieve economies of scale in production, followed by a recommendation to concentrate on Numerical Control machines. In addition, export cartels were authorized to regulate export prices and facilitate joint export activities, and government funding was used for joint research activities. Although Sarathy acknowledges the importance of corporate strategy in improving firm competitiveness, he concludes that MITI policies made a critical contribution to the industry’s ultimate success by realizing the technological importance of the numerical control breakthrough and prodding the industry into incorporating this development. Thus, the symbiotic relationship between Industrial Policy and
Corporate Strategy can be a key factor in creating and maintaining the critical 98 competitive advantages which determine industrial competitiveness.
III. Inter-Firm Cooperation
Much of the research on Industrial Policy and Competitiveness seems to indicate that Cooperation is a better means of encouraging economic growth than Competition. If this is in fact truS, it could have a major impact on the way firms structure their internal and external relationships. Lodge and Walton
(1989) explore this issue in a recent study and conclude that, in order to remain competitive, U.S. frims must move away from the adversarial, arm's-length, short-term contractual relations of the past, toward arrangements that are more cooperative, flexible, intimate, and long term.
These new relationships have a number of advantages. For example, they allow the competitive unit to achieve the benefits of scale without increasing corporate size, they facilitate technology transfer, and they can shorten the time from product conception to commercialization. Thus, according to Lodge and Walton, the competitive unit of the future will not be the corporation, but a coalition of rivals cooperating with labor, suppliers, and customers. This new approach is clearly evident in the changing attitude toward supplier relations. There is a growing realization that stable, secure, long-term supplier relations are a major source of competitive advantage. As a result, many firms are beginning to view their suppliers as an integral part of the firm’s value-added chain rather than a temporary source of a particular good or service.
Lodge and Walton argue that the corporation’s long term health demands a reexamination of old assumptions about shareholders as well. They suggest that the old models of corporate purpose and governance are no longer valid due to the increasing dispersion and lack of involvement of the 99 firm’s shareholders. Instead, a new model of the firm is emerging, based on the corporation’s responsibility to serve all of its constituents, including customers, suppliers, and employees, as well as the firm’s shareholders and debt holders.
Lodge and Walton note that this will require a corresponding set of socio political attitudes that promotes cooperation, as well as a new definition of the role and function of Government, Business, and Labor. Taken together, these steps will produce a firm that represents a “strategic partnership” between shareholders, management, employees, suppliers, and customers rather than an organization forged from the adversarial relationship between self-serving constituencies.
Many of these ideas are echoed by Spekman (1988) in an article which focuses on supplier relations. Spekman notes that the traditional adversarial model attempts to minimize the price of purchased goods and services by playing a large number of suppliers against one another through arm-length, short-term contractual arrangements. While this approach may gain price concessions in the short run, it cannot establish the trust, commitment, and information sharing necessary to achieve mutual productivity and technology gains in the long run. As a result, a new attitude toward supplier relations is developing which ties the company’s ability to compete globally to its ability to create a high level of trust and cooperation among its suppliers.
Spekman notes that this sort of collaboration requires the development and maintenance of long-term trading relationships characterized by a high degree of communication and purposeful cooperation, and a conflict resolution process based on dialogue and joint problem solving rather than negotiation.
These sorts of relationships can allow the firm to coordinate activities in such a way that the specialized skills of its suppliers can be exploited and the 100 advantages of vertical integration can be gained without incurring the costs of ownership and/or bureaucratic control.
There are many similarities between this new model of the firm and the
Jap an ese Keiretsu. The Keiretsu is a large industrial complex consisting of a number of often unrelated companies tied together through a network of formal and informal business relationships. In a recent article in “Business Week”,
Blinder (1991) examines the advantages of the Keiretsu and concludes that,
“The astounding success of Japanese auto and electronics companies suggests that the long-term relationships that define a production keiretsu might be a better coordinating mechanism than either vertical integration or open markets...” (1991, pg. 32). According to Blinder, the production Keiretsu represents a third form of industrial organization somewhere between vertical integration and arm-length exchange in the marketplace. Since the Keiretsu consists of independent firms joined together through a web of long-term relationships, this structure is capable of simultaneously combining the benefits of internal organization with those of the marketplace. As a result, the Keiretsu may prove to be a useful model for the future direction of MNE organizational structure.
Yamamura (1975) examines some of the unique socio-cultural factors that have influenced the structure of Japanese industries and firms. A common feature of these cultural traits is an emphasis on promoting cooperative behavior at all levels of the firm and throughout the society. Yamamura identifies
4 key elements in the development of Japanese business practices. First, he notes the “vertical” nature of Japanese society, where vertical (hierarchical) human relationships dominate horizontal relationships. This acts to promote harmonious social interaction by prescribing appropriate behaviors based on 101 an individual’s place in the society. The second feature is the social characteristic of paternalism or groupism. The Japanese people tend to think of themselves as belonging to a larger whole and act in a manner which places group interests above those of the individual.
Taken together, these factors account for the Japanese management practices of lifelong employment and the seniority system. Specifically, the seniority system defines an individual’s position within the firm in an unambiguous manner and is a reaction to the vertical character of Japanese society. The practice of lifetime employment is designed to satisfy the individual’s need to belong to and be accepted by a larger group. Finally,
Yamamura discusses the ringi system of decision-making which strives to promote cooperation by building a consensus among all levels of the organization.
Although most writers stress the specific nature of these Japanese business practices, the economic benefits derived from cooperative behavior are by no means limited to Japan.24 In fact, many of the benefits of Cooperation can be explained through Transaction Cost Analysis. According to this approach, the firm represents a response to the costs of negotiating, monitoring, and enforcing a contract to protect against opportunistic behavior. However, as
Hill (1990) notes in a recent article, opportunism and the safeguards adopted to attenuate it can greatly reduce the rents generated by a particular investment.
Using a Game Theoretic approach, Hill demonstrates that the value of the rents inherent in a given transaction is maximized when parties to the exchange cooperate and trust one another. Based on this analysis, he concludes that over the long run, the Invisible Hand of the Market favors cooperative not competitive 102 behavior.25
In terms of the present discussion, the types of relationships described
above are specifically designed to limit the threat of opportunism, and with it, the
need to replace the market with a hierarchical form of governance, or monitor
various constituents to assure that the interests of the firm are being protected.
According to Hill, the total rents arising from a transaction based on cooperation
between independent parties actually exceeds the rents earned from an
exchange conducted through the internal market of the firm; and therefore, cooperative behavior carries with it a considerable competitive advantage.
IV. Contractual versus Direct Control
The arguments developed above suggest that, in many cases, arms-
length transactions between independent parties can be a more efficient means of organizing economic activities then “internalizing” them within the firm.
Accordingly, Contractor (1990) notes that there are now 2 very different models
of the MNE: the traditional model based on globalization and standardization, in which the need for efficiency mandates the use of hierarchical control and internal markets; and a new model in which a number of unique organizational and environmental factors create a fragmented strategy landscape which
makes the use of contractual control and external markets the preferred option.
The shift from equity to contractual control of the firm’s foreign operations is clearly evident in the growing use of “New Forms of Investment” (NFIs). These include: Licensing, Management Contracts, Strategic Alliances, Minority Joint
Ventures, and Collaborative R&D Agreements. Contractor notes that Licensing
Agreements and Joint Ventures (JVs) currently outnumber Wholly Owned
Subsidiaries (WOS) by a 4 to 1 margin. This is based on a 1982 study which found over 31,000 independent foreign firms with licensing agreements with 103 U.S. firms; and a 1985 Department of Commerce study which revealed over
16,000 JVs or foreign corporations with shared U.S. ownership (roughly 12,000 of which involved 50-50% or minority ownership). These figures compare to a total of only 8,000 foreign WOS of U.S. corporations.
Contractor uses Transaction Cost Analysis to explain this finding, and develops a model for comparing the costs of alternative methods of servicing foreign markets. Based on this analysis, he concludes that many transaction costs have declined over time, making contractual and collaborative agreements between independent firms an increasingly attractive alternative to
FDI. In addition, the high cost of R&D and the reduction of product cycles have increased the costs and risks associated with international activities. This provides MNEs with a powerful incentive to join together to share development costs, amortize expenses over a larger joint market, and even prolong product cycles by controlling competition.
Hill, Hwang, and Kim (1990) also attempt to explain the firm’s choice between alternative entry modes. Focusing on Licensing, Joint Ventures, and
Wholly Owned Subsidiaries, the authors identify 3 key factors involved in this decision; control, resource commitment, and dissemination risk (i.e. the risk of losing control over proprietary knowledge). They note that Licensing has low control and resource commitment, but high dissemination risk; while JVs have moderate control, resource commitment, and dissemination risk; and WOS have high control and resource commitment, but low dissemination risk.
Next the authors examine 3 groups of variables (Strategic,
Environmental, and Transaction) which define the firm’s unique strategic and environmental setting. Strategic variables address the unique strategic objectives and organizational capabilities of the firm, and include such things as 104 as national differences between markets, the need to achieve scale economies,
and the extent of the firm’s global concentration of a particular activity. On the
other hand, Environmental variables focus on the unique characteristics of the
market(s) being served. They include such things as country risk, the firm’s
familiarity with a particular location, local demand conditions, and the volatility
of international competition. Finally, Transaction variables address the nature of
the specific activity being conducted and include the type and value of the firm-
specific knowledge being transferred, the extent of “tacit” know-how involved,
and the amount of asset-specific investments being made.
Based on an analysis of these 3 sets of variables, Hill, et. al. are able to
explain a firm’s preference for a particular entry mode. For example, an analysis
of Strategic variables reveals that firms which pursue a multi-domestic strategy
will, ceteris paribus, favor low-control entry modes, while firms pursuing a global
strategy will prefer high-control entry modes. This follows from the overriding
importance of national differences in a multi-domestic strategy versus the need to concentrate activities and achieve economies of scale under a global
strategy. Similarly, an analysis of Environmental variables suggests that high
country risk, uncertain demand, lack of familiarity, and volatile competition will
encourage the firm to choose an entry mode which carries a low resource commitment. Finally, Transaction variables suggest that the greater the rents generated by an MNE’s proprietary know-how, the more likely the firm will be to
select an entry mode with low dissemination risk. On the other hand, the greater the tacit component of the knowledge being transferred, the more likely a high control entry mode will be used.
Rather than focus on the positive aspects of these contractual arrangements, Hennart (1989) explains the proliferation of “New Forms” in 105 terms of the negative aspects of FDI, particularly its impact on MNE-Host
Government relations. Hennart notes that since FDI implies ownership and thus control, it often creates a dilemma for Host Governments. In many cases, governments have responded by placing severe restrictions on certain types of foreign investments in an attempt to encourage Joint Ventures with local partners or the contractual transfer of goods and services.
Hennart identifies 3 phases in Host Governments’ attitudes toward FDI.
The first phase began in the mid-1960s and consisted of many Developing
Countries restricting FDI and mandating NFIs. The result of these actions was that NFIs proved to be a costly means of acquiring advanced technology and marketing skills. This led to the second phase, beginning in the mid-1970s, which was characterized by more sophisticated forms of contracting such as
Management Contracts, Turn-key, and various forms of Countertrade. Although these changes did improve the contractual efficiency of the “New Forms”, they still remained expensive alternatives to FDI. This encouraged the third phase beginning around 1980 in which many governments began to relax their restrictions on FDI. In fact, since 1980 Developing Countries have increasingly turned to FDI and market reform as a means of encouraging economic growth.
Based on an analysis of the costs and benefits of alternative forms of international governance, Hennart concludes that the "New Forms” cannot fully substitute for FDI due to the problem of opportunism. Hennart notes that although contracts conserve on external organization costs, they often fail to constrain the behavior of all parties and can incur high transaction costs. On the other hand, FDI restrains opportunistic behavior by aligning the incentives of all parties, but often requires high internal organization costs; while Joint Ventures provide some protection against opportunism, but may encourage partners to 106 “free ride”. Hennart also notes that the higher the tacit component of the knowledge or technology being transferred and the weaker the legal protection, the less efficient contractual methods will be. Thus, Hennart argues that
Licensing provides a superior alternative to FDI only for the transfer of older, non-tadt types of knowledge, and for process rather than product technologies.
Finally, he notes that contractual arrangements should not be used when external markets contain few potential buyers due to the high risk of opportunistic behavior.
Several interesting points arise from this brief discussion; the most important being that the increased use of Contractual means to coordinate the
MNE's international activities is a response to a multiplicity of factors. While it is clear that these “New Forms” offer the MNE many advantages over traditional equity-based arrangements (i.e. FDI), they also place a number of unique organizational and strategic demands on the firm. This may explain the growing number of studies that attempt to develop “contingency” models for selecting the appropriate method of servidng a particular foreign market.
The second point is the importance Transaction Cost principles play in explaining the use of these Contractual and Collaborative Arrangements. This is particularly interesting since the very same principles were used to explain the firm’s “internalization” of a given transaction earlier in this chapter. The point here is that Transaction Cost Analysis provides an extremely powerful tool for analyzing all types of exchange relationships; and therefore, it deserves all the attention it has received over the past decade.
Finally, it is important to acknowledge the possibility that recent changes in the international environment have been driving the shift to Contractual control. Many of these changes were discussed in the “Background” section of 107 Chapter I and involve the expansion of international trade and the emergence of global markets and firms. In this new competitive environment, it has become increasingly efficient to coordinate international transactions through external markets, particularly since contractual arrangements between independent firms can reduce the costs and risks associated with many activities, and at the same time avoid Host Government restrictions on FDI.
V. Joint Ventures, Licensing, and Strategic Alliances
Since the purpose of this section is to examine various types of
Cooperative Behavior and mrt the uNew Forms of Investment”per se, it is impossible to review the growing body of literature on these popular contractual arrangements in any detail. Instead, the following discussion provides a brief introduction to current research on Joint Ventures, Licensing, and Strategic
Alliances for those readers who may be unfamiliar with these arrangements. It is included for explanatory purposes only.
A recent article by Kogut (1988) provides an excellent overview of the voluminous literature on Joint Ventures (JVs). According to Kogut, a Joint
Venture is a legal organization formed by the pooled resources of 2 or more
“parent” firms. Kogut identifies 3 basic approaches to the study of JVs in the literature. The first is based on Transaction Cost Analysis and views the JV as an attempt to avoid small number bargaining situations. Since JVs provides a superior means of aligning incentives among parties and achieving an equitable division of costs and profits, this approach considers the JV to be an appropriate response to situations involving a high degree of uncertainty over specifying and monitoring performance, and with a high degree of asset specificity. 108 The second approach focuses on the Strategic Behavior of the firm and sees the Joint Venture as a means of enhancing the firm’s competitive position.
Kogut notes that this perspective addresses the influence of competitive positioning on the asset value of the firm, while Transaction Cost Analysis examines the costs associated with a particular economic exchange regardless of its strategic objective. Consequently, these 2 approaches often differ in their predictions of which firms will enter a Joint Venture, as well as in their explanations of why firms do so. The final approach is based on the JV’s ability to act as a mechanism for transfering organizational knowledge. According to this view, a JV occurs when one or both firms want to acquire the other’s organizational know-how, or when one firm wishes to maintain an organizational capability while benefiting from the other’s current knowledge or cost advantage.
Kogut also reviews the many empirical studies of JVs in an attempt to uncover common themes in this research. Based on studies by Stopford and
Wells (1972) and Farge and Wells (1982), it appears that 2 factors have a strong influence on the amount of equity each parent assumes in a given JV.
They are: the strategic importance of the R&D or marketing expenditures involved, and the firm’s product diversity. A study by Kogut and Singh (1986) found that the firm’s preference to use a JV to enter a particular market was influenced by the size of the venture, the characteristics of the industry, and the cultural characteristics of the Home and Host countries. Further, Stopford and
Weils (1972) found that the responsibilities of a particular venture were influenced by the capabilities of the Host country and of both parents, and by the potential for conflict between the venture and any of its parents. 109 A number of studies have examined performance and instability among
Joint Ventures and have reported failure rates ranging from 30-65%. Beamish
(1985) concluded that JVs in Developing Countries had higher instability rates than those in Developed Countries (although this finding is not universally accepted); but Killing (1982, 1983) found that stability was related to the presence of a dominant parent, and Kogut (1987) demonstrated that stability was influenced by the health of the industry, the cooperative incentives among partners, and the degree of competitive rivalry.
Harrigan (1988) develops a model for assessing the usefulness of
Cooperative Strategies in various competitive environments. According to
Harrigan, the difference between a Joint Venture and other forms of inter-firm cooperation is that a Joint Venture involves an equity commitment by each partner, while Cooperative Agreements do not. She identifies several key environmental characteristics which affect the form a cooperative strategy will take; including, demand uncertainty, customer traits, infrastructure development, production technology, the volatility of competitive behavior, and the nature and extent of the relationship between the JV and its parents. Harrigan argues that demand traits will determine the need to adopt a cooperative strategy while competitor traits will suggest how firms respond to that need in a particular industry. Based on this analysis, Harrigan concludes that Cooperative
Agreements will be more appropriate when demand is uncertain or business risks are high, but that Joint Ventures will be used when demand is more stable, or when there is a need to adjust production capacity and ease competitive pressures in declining industries.
In terms of competitor traits, Harrigan examines the nature of industry competition, industry infrastructure, capital intensity, technological scale, and 110 the firm's strategic posture. Based on these considerations, she argues that firms will expose less equity to the risks of JVs in volatile competitive environments, and in situations where proprietary know-how is difficult to protect. However, as projects grow larger and more risky, product life cycles shorten, technologies become more expensive, and global competition increases, firms will find it increasingly difficult to survive independently, and will use JVs to develop new markets and industries, rationalization mature industries, and enhance the firm’s competitive position.
An early study by Contractor (1981) looks at the role of Licensing in international strategy. Contractor argues that in many cases licensing is not a second best solution, but a superior alternative to FDI on a risk adjusted basis.
In addition, he notes that U.S. executives actively engaged in licensing are not as concerned about creating new competitors or losing control over proprietary knowledge as many writers believe. This may suggest that licensing is far less risky than commonly believed. Based on earlier research, Contractor identifies
12 conditions which favor licensing. They are: 1) the standardization of the product’s life cycle, 2) Host country restrictions FDI or FDI income, 3) restrictions on imports into the Host nation, 4) either very large or very small licenser firm size, 5) high research intensity in the licenser firm, 6) industries with high rates of technological turnover, 7) the transfer of product versus process technologies, 8) cases which involve a reciprocal exchange of technology, 9) the ability to create or perpetuate licensee dependency, 10) the licenser’s ability to choose competition by granting a license, 11) the ability to create auxiliary business, and 12) the ability to diversify and reorganize product lines in the licensor firm. 111 Contractor also examines a number of recent changes in the international environment that encourage the shift from “internalization” to licensing. These include: the saturation of traditional overseas markets, the emergence of effective global competition, the global proliferation of alternative sources of technology, increased government intervention, a weakening international patent system, increased protectionism, spiraling transportation costs, and the overvalued dollar exchange rate. In each case, licensing provides an inexpensive alternative to FDI which can be used to enhance the firms competitive position without assuming the costs and risks associated with direct investment.
In a more recent study, Clegg (1990) examines the determinants of licensing behavior in 5 countries (the U.S., U.K., Japan, Germany, and
Sweden). Clegg notes that licensing is affected by the nature of the industry and product, the structural characteristics of the industry, region and country specific influences, and cultural differences. Using available empirical data, Clegg concludes that licensing activity is greatest in technology intensive industries; and that the U.K. is the largest outward licenser nation, whereas the U.S. licenses the least from abroad, and Japan licenses the most. Clegg is unable to explain the licensing behavior of all 5 countries in a single model, and notes that although technology intensity emerges as an industry specific factor, country specific factors also play an important role in the decision to license. As a result, Clegg proposes 2 models to explain the observed patterns of outward licensing; the American model (which works for all countries except the U.K.) and the U.K. model.
Specifically, the U.K. pattern of licensing is less technology intensive and more managerially intensive than that of other countries, and is biased away 112 from the developed European countries toward the Developing Countries.
Clegg also discusses the role Japanese Industrial Policy played in promoting the import of research intensive technologies, and notes that both Japan and
Germany import technologies which require a high level of operative skills.
Finally, he concludes that the effect of country specific variables of all types account for many of the observed differences in licensing behavior.
A recent working paper by Ajami (1989), looks at Strategic Alliances
(SAs) and notes that traditional entry strategies based on equity ownership must be expanded to include these contractually based arrangements as well.
A Strategic Alliance is usually defined as a business venture involving 2 or more entities that is established to achieve a specific corporate goal. Thus, SAs are characterized by non-equity based, contractual linkages between firms, and have diffused ownership and often limited lifespans.
Ajami notes that these arrangements offer a number of unique advantages, including: exploiting synergies and creating value, increasing the firm’s strategic flexibility and adaptability, reducing the costs and risks of conducting international activities, and enhacing the firm’s competitive advantage. Further, SAs provide a “fully bundled” package of technology, management, and marketing services; while Licensing provides only the
“unbundled” elements, and FDI falls somewhere in between. But SAs can be extremely difficult to m anage and may suffer from cultural conflicts, confused strategies, and a loss of focus. Accordingly, Ajami argues that SAs must reflect the following in order to succeed: shared goals and objectives, a coupling of risk and responsibility, flexibility among partners, clearly defined boundries between the alliance and its members, and shared understanding and agreement about control mechanisms. 113 Ohmae (1989) provides a somewhat different perspective on Strategic
Alliances. According to Ohmae, the global convergence of consumer needs and preferences means that firms must be able to compete in every major market simultaneously. This task is complicated by the shortening of product life cycles, the global diffusion of technology, and the numerous competences required to produce and market competitive products. As a result, the key to competitive success lies in finding a way to defray the immense fixed costs associated with operating on a global scale. SAs provide an effective way of doing this. Ohmae argues that today’s competitive environment demands a new logic, one that replaces the current emphasis on variable costs and return on investment (ROI) with a focus on fixed costs and return on sales (ROS). In a variable costs environment, managers attempt to boost profits by reducing the cost of factor inputs; however, in this new fixed cost environment, boosting sales in order to maximize marginal contributions to fixed costs is the key to competitive success.
Ohmae also argues that managers must overcome their obsession with total equity control, since it is no longer a guarantee of success. This applies to JVs as well. Although JVs may be useful in some instances, they face
2 major obstacles: problems associated with contracting, and conflicts between the venture and its parents. Therefore, a selective set of non-equity arrangements (i.e. SAs) can provide globally active companies with a means of maximixing their contribution to fixed costs, while reducing the costs and risks associated with establishing globai production, marketing, distribution, logistics, and R&D facilities.
VI. A New Model of the Firm
The Strategic Alliance provides the firm with a non-equity vehicle for conducting certain International activities and represents a specific agreement 114 between 2 or more firms to achieve a limited strategic objective. On the other hand, some authors have proposed more far-reaching structures which allow the firm to manage a wide range of international activities through contractual rather than equity means. An example of such an arrangement is what
Perlmutter and Heenan (1986) term a Global Strategic Partnership (GSPs). The
GSP is an alliance in which: 1) two or more companies develop a common long term strategy aimed at world leadership as low cost suppliers, differentiated marketers, or both, 2) the relationship is reciprocal with each partner possessing specific strengths, 3) the partners efforts are global, not just concentrated on the Developed Countries, 4) the relationship is organized along horizontal not vertical lines, and 5) the participating companies retain their national and ideological identities when competing in those markets excluded from the partnership.
Not surprisingly, the scope of this type of inter-firm collaboration requires a number of special consideration. For example, when developing a GSP, the partnership's mission must be clearly defined and a workable strategy must be found to balance the forces of cooperation and competition between partners. In addition, an effective governance structure and management style must be created as well as a way to resolve potential conflicts between partners. Finally, both parties must be equally commited to the alliance. Perlmutter and Heenan note that the traditional view of the MNE associates competitive advantage with home country factors, proprietary technology, centralized decision making, vertical planning, and market dominance. On the other hand, the new global model is more flexible about ownership and managerial control, and encourages joint decision making, vertical and horizontal planning, and the fusion of competent international allies into a globally competitive unit. 115 Johnston and Lawrence (1988) propose a similar arrangement called the
Value-Adding Partnership (VAP). The VAP is a set of independent companies that work closely together to manage the flow of goods and services along a product’s entire value added chain. The authors note that lower communication and computing costs have shifted the competitive advantage away from vertical and horizontal integration, back toward coordinated partnerships of independent businesses each specializing in a single activity. The key to success in this new competitive environment is accepting that each member of the value added chain has a stake in the others' operations.
Johnston and Lawrence use the Italian textile industry as an example of a successful VAP and note that the industry’s recent restructuring along these lines has greatly increased its competitiveness. In addition, they note that the close inter-firm relationships which characterize Japanese trading companies and auto producers are typical of VAPs. On the other hand, companies involved in arms-length transactions in a competitive environment are often guarded or antagonistic, and thus have little interest in the activities of other firms. While vertical integration can solve some of these problems, it often inhibits the development of destinctive competences at each step of the integrated chain.
Therefore, VAPs offer the best of both worlds by providing the advantages of scale and coordination associated with large vertically integrated companies, while maintaining the flexibility, creativity, and low overhead of small firms.
It is interesting to note the similarities between these 2 structures and the
Keiretsu discussed above. Although the Keiretsu is not a typical Japanese business structure (there are only 9 major Keiretsu and several smaller ones), many of its unique characteristics are the result of more basic Japanese business practices. According to Kester (1991), 4 traits define the Japanese 116 system of corporate governance: 1) implicit contracting founded on trust, 2) extensive reciprocity in equity ownership and commercial trade, 3) managerial incentives aligned toward overall corporate growth and away from transfers of wealth between stakeholders, and 4) selective intervention by key stakeholder to correct problems. Kester concludes that:
“The overall effect of Japanese corporate governance is to foster tremendous efficiencies in the execution of business transactions by making it easier to build and maintain long-term business relationships. These relationships have enabled Japanese companies to function with lower degrees of asset- ownership integration...despite high degrees of interdependence in the transactions supported by those assets. (1991, pg. 54)
Thus, the new corporate model based on Cooperation and contractual relations between independent firms can provide a powerful tool for creating critical competitive advantages.
VII. Conclusions
This section has presented research on various types of Cooperative
Behavior. Interest in this subject represents a response to 2 recent developments: the declining Competitiveness of many key U.S. industries; and the widespread use of Joint Ventures, Strategic Alliances, and other
Contractual Arrangements. Based on numerous studies of Japanese and
Western European industries and firms, there is a growing recognition that cooperative behavior can play a key role in creating and maintaining critical competitive advantages for both the Firm and the Nation. For example, cooperation between Business and Government, and Industrial Policies aimed at promoting certain types of inter-firm Cooperation (particularly joint R&D activities and the diffusion of innovations among firms and throughout the economy) have often been associated with increased innovation and economic 117 growth. As a result, a growing number of scholars are beginning to question the validity of the traditional Competitive model with its emphasis on arms-iength, adversarial relations between various stakeholder groups. Instead, there is a growing sense that Cooperation not Competition provides the superior means of structuring many economic relationships.
The increased use of Cooperative Strategies has prompted Contractor
(1990) to note that there are now 2 very different models of the MNE: the traditional model based on globalization and standardization in which the needs for efficiency and integration mandate the use of hierarchical control and internal markets; and a new model in which unique environmental and strategic factors create a fragmented strategy landscape which makes the use of contractual control and external markets preferable. A shift from the traditional equity-based model to the contractual control of the firm's international activities is clearly evident in the increased use of the so-called “New Forms of
Investment” (Licensing, Management Contracts, Strategic Alliances, Minority
Joint Ventures, and Collaborative R&D Agreements). Contractor recently found that Licensing Agreements and Joint Ventures outnumber Wholly Owned
Subsidiaries by a 4 to 1 margin.
Attempts to explain this shift away from equity can be reduced to 3 basic arguments: 1) the increased use of contractual arrangements represents a response to specific environmental and strategic factors, such as the increased costs and risks of conducting international activities, shorter product and technology cycles, and the convergence of consumer tastes; 2) contractual arrangements offer the MNE a means of circumventing various Government restrictions on FDI; and 3) the Transaction Cost rationale for the firm seriously exaggerates the threat of opportunism and other transaction costs making 118 contractual control a more viable alternative then commonly suggested.
In many ways, this “new” model of the firm, with its focus on developing and maintaining long-term, contractual relationships between various stakeholders, is similar to the Japanese Keiretsu. The Keiretsu is a large industrial complex containing of a number of often unrelated firms tied together through formal and informal means. Blinder (1991) argues that the Keiretsu represents a third organizational form somewhere between the Firm and the
Market, and suggests that it may provide a useful model for the future direction of the MNE. On the other hand, some writers have proposed even more radical organizational forms such as Perlmutter and Heenan’s (1986) “Global Strategic
Partnership” or Johnston and Lawrence’s (1988) “Value-Adding Partnership”. In both cases, purely contractual arrangements between independent firm’s are used to coordinate the entire range of the firm’s activities.
The most important aspects of this section are summarized below:
1) There is a growing recognition that, in many cases, Cooperation not Competition provides the superior means of organizing economic relationships. This has encouraged researchers to examine the role of Industrial Policy and various types of inter-firm cooperation in creating and maintaining critical competitive advantages for both the Firm and the Nation.
2) According to Contractor (1990) there are now 2 very different models of the MNE: the traditional model based on globalization and standardization in which the needs for efficiency and integration mandate the use of hierarchical control and internal markets; and a new model in which unique environmental and strategic factors create a fragmented strategy landscape which makes the use of contractual control and external markets preferable.
3) A shift away from the traditional equity-based model of the MNE toward the contractual control of the firm's international activities is clearly illustrated by the increased use of the “New Forms of Investment” (Licensing Agreements, Management Contracts, Strategic Alliances, Minority Joint Ventures, and Joint R&D Projects). 119 4) Attempts to explain this shift from equity to contractual control can be reduced to 3 basic arguments: 1) it represents a response to specific environmental and strategic factors, 2) contractual arrangements allow the MNE to circumvent various Government restrictions, and 3) the threat of opportunism and other transaction costs have been grossly exaggerated, making contractual control a more viable alternative than commonly thought.
5) A number of writers have attempted to assess the structural implications of this shift to contractual control. For example, Perlmutter and Heenan’s (1986) “Global Strategic Partnership and Johnston an Lawrence’s (1988) “Value-Adding Partnership” provide a model of the firm in which contractual arrangements between independent firms coordinate a wide range of activities. In addition, the Keiretsu represents an organizational structure which blends elements of both equity and contractual control. Similar structures may become the dominant organizational forms of the future. 120 Summary
This chapter has examined selected research from 3 large and diverse areas: The Theory of the Firm, Organizational Theory, and International
Strategic Management. Since it is clearly beyond the scope of this endeavor to conduct a thorough review of each of these fields, 2 criteria have guided the selection of the research presented above: the use of studies from the field of
International Business whenever possible, and the desire to provide theoretical and empirical support for the arguments advanced in the” Overview” section of
Chapter I.
The following statements provide a concise summary of the chapter’s most important points:
1) Based on arguments developed by Coase (1937) and Williamson (1975), the Firm and the Market can be seen as alternative methods of organizing economic activities. The choice between these two alternatives depends on a set of factors which influences the relative cost and efficiency of each mode.
2) Internalization Theory, as developed by Buckley and Casson (1976), extends these arguments to an international context and applies them to'the MNE. According to this theory, the MNE represents a response to certain incentives which encourage the firm to replace international markets with its own internal markets. When a firm Internalizes” a market across national boundaries it becomes an MNE.
3) It is possible to identify 3 distinct types of incentives which encourage internalization: those based on Environmental factors, those arising from the Monopolistic advantages of firms, and those based on Transaction Cost Analysis.
4) In terms of the current study, it is important to recognize that Environmental factors can and do influence the firm’s decision to use internal or external markets. Accordingly, a significant relationship can be hypothesized between certain environmental factors and the firm’s decision to internalize a particular market. 121 5) In general, 2 distinct approaches to Organizational Theory can be identified: one which focuses on the structural properties of the organization, and another which examines the behavioral characteristics of the organization’s constituents.
6) Beginning with Chandler (1962), a number of studies have examined the relationship between the firm and the environment. Their findings tend to support the argument that organizational structure is related to strategy in much the same way that strategy is related to environmental forces.
7) The Contingency approach developed by Lawrence and Lorsch (1967) extends this analysis by suggesting that performance is directly affected by the relationship between a firm's structure and the environment.
8) Contingency Theory has had an enormous impact on the field of International Strategic Management by encouraging the development of “contingency" models which specify the optimal strategy for a firm facing a given contextual setting.
9) Contingency Theory has also encouraged writers on the MNE to consider 3 closely related issues: adaptation versus integration, multidomestic versus global strategies, and the political and economic imperatives. The focus of this research has often been to devise strategies which simultaneously address the MNE’s need to adapt to local conditions and integrate operations across activities and markets.
10) Other Organizational theories have also influenced research on the MNE. For example, Theories of Environmental Adaptation stress the dynamic nature of the environment and the firm’s need to adapt to environmental change, while Theories of Organizational Learning emphasize the organization’s ability to adapt to the environment by creating or modifying behavioral and/or structural processes. These ideas are just beginning to be applied to the MNE.
11) There is a growing recognition that, in many cases, Cooperation not Competition provides the superior means of organizing economic relationships. This has encouraged researchers to examine the role of Industrial Policy and various types of inter-firm cooperation in creating and maintaining critical competitive advantages for both the Firm and the Nation.
12) According to Contractor (1990) there are now 2 very different models of the MNE: the traditional model based on globalization and standardization in which the needs for efficiency and integration 122 mandate the use of hierarchical control and internal markets; and a new model in which unique environmental and strategic factors create a fragmented strategy landscape which makes the use of contractual control and external markets preferable.
13) A shift away from the traditional equity-based model of the MNE toward the contractual control of the firm’s international activities is clearly illustrated by the increased use of the “New Forms of Investment” (Licensing Agreements, M anagement Contracts, Strategic Alliances, Minority Joint Ventures, and Joint R&D Projects).
14) Attempts to explain this shift from equity to contractual control can be reduced to 3 basic arguments: 1) it represents a response to specific environmental and strategic factors, 2) contractual arrangements allow the MNE to circumvent various Government restrictions, and 3) the threat of opportunism and other transaction costs have been grossly exaggerated, making contractual control a more viable alternative than commonly thought.
15) A number of writers have attempted to assess the structural implications of this shift to contractual control. For example, Perlmutter and Heenan’s (1986) “Global Strategic Partnership and Johnston an Lawrence’s (1988) “Value-Adding Partnership” provide a model of the firm in which contractual arrangements between independent firms coordinate a wide range of activities. In addition, the Keiretsu represents an organizational structure which blends elements of both equity and contractual control. Similar structures may become the dominant organizational forms of the future. 123 NOTES
1) Teichova (1986) and Wilkins (1986) provide clear historical evidence of MNE activity in the 19th Century; however, Carlos and Nicholas (1988) suggest that the Chartered Trading Companies of the 16th and 17th Centuries can also be seen as MNEs, suggesting that the basic structure of the MNE is much older than generally assumed.
2) Notable reviews of Hymer’s work appear in Buckley (1981), Calvet (1981), Dunning and Rugman (1985), Clegg (1987), and Horaguchi and Toyne (1990).
3) Japanese FDI in the Pacific Rim has more than doubled in the last 5 years, and by 1989 totaled over $30 billion. This com pares to total U.S. investment in the region of roughly $20 billion. In addition, Sanger (1990) notes that Japan accounted for almost 1/2 of the 1989 inflow of foreign investment into Thialand, and a substantial portion of the investment in Indonesia, Malaysia, and Singapore as well.
4) Dunning’s Eclectic Paradigm was designed to explain the phenomenon of International Production; however, since Dunning defines the MNE as a company which undertakes production outside its country of origin, it seems clear that the Eclectic Paradigm must also be considered a Theory of the MNE.
5) Divestment is the process of disposing of a particular investment. For a more thorough discussion of this topic, see Boddewyn (1983).
6) Dunning’s recent book, Explaining International Production (1988b), consolidates over a decade of his research on the Eclectic Paradigm and provides the most complete presentation to date of both the theoretical and empirical basis of the theory.
7) It is interesting to note the similarities between Teece’s Transaction Cost model of the MNE and Dunning’s Eclectic Paradigm.
8) For example, see Kogut (1984), Porter (1986), Ghoshal (1987), Zengage and Ratcliffe (1988), and Yoffie and Milner (1989).
9) According to Hout, Porter, and Rudden (1982), a global strategy is one in which the firm competes in terms of its integrated worldwide production and marketing system. For a detailed discussion of global strategies, see Porter (1986), Prahalad and Doz (1987), Bartlett and Ghoshal (1989), and Ohmae (1990).
10) Major reviews of MNE Theory appear in Rugman (1980), Buckley (1981), Calvet (1981), Casson (1983), Clegg (1987), Dunning (1988b), and Doz and 124 Prahalad (1991).
11) For example, Doz (1980) notes that the advantages of coordinating and integrating activities on a global scale include cost and efficiency savings arising from economies of scale, increased MNE bargaining power, and simplifying the management of international operations.
12) Although Rumelt is not interested in the MNE per se, his study employs a random sample of Fortune 500 firms. Since a large portion of these firms had international operations by the late 1960s and early 1970s, it is reasonable to assume that his results are applicable to the MNE as well.
13) Wind and Douglas (1981) discuss portfolio planning models developed by the Boston Consulting Group, G.E./McKinsey, Shell International, and A.D. Little.
14) For a more detailed discussion of the BCG matrix see Henderson (1978), and Abell and Hammond (1979).
15) For a detailed discussion of the theme of adaptation versus integration see Prahalad and Doz (1987) and Bartlett and Ghoshal (1989).
16) One of the most extensive applications of Levitt’s arguments appears in Ohmae (1985,1990).
17) Among the many writers who disagree with Levitt's hypothesis are Boddewyn, Soehl, and Picard (1986), Kotler (1986) and Douglas and Wind (1987).
18) The dynamic nature of MNE competitive advantages is explored in Vernon (1966), Dunning (1980), Porter (1986), Dunning (1989), and many others.
19) In general, the “New Forms of Investment” are those inter-firm arrangements which fall short of majority ownership and include various forms of contracting and minority Joint Ventures. For a further discussion of this subject see Oman (1984), Franko (1987) and Hennart (1989).
20) Studies by Killing (1983), Beamish (1985), and Gomes-Casseres (1987) have found that between 30-50% of all Joint Ventures end in failure. In addition, Auster (1987) notes that 7 of 10 JVs fall short of their expectations or perform unsatisfactorily.
21) As an indication of the extent of America’s competitive decline Franko (1989) notes that U.S. firms accounted for over 2/3 of World production in 10 of 15 industries in 1960, but only 4 of those 15 industries in 1986. 125
22) The reasons for America’s competitive decline are explored in Thurow (1984, 1987), National Academy of Engineering (1985, 1986), and Vernon (1986,1987) among others.
23) For a detailed discussion of the Cooperative Research Act of 1984 see Lodge and Walton (1989), and Jorde and Teece (1989).
24) A recent book by Lincoln and Kalleberg (1990) suggests that the cultural specificity of Japanese business practices is largely overstated, and that firms outside Japan can derive many of the same benefits by adopting similar organizational designs and management styles.
25) Hill notes that, although opportunism may not be as widespread as commonly believed, it can still provide the rationale for hierarchical governance when outcomes are uncertain, reputations are difficult to establish, or the rewards from opportunistic behavior in the present outweigh the discounted present value of future cooperation. CHAPTER III
METHODOLOGY
Introduction
This study attempts to explore the relationship between environmental factors and the MNE’s preference for using external (rather than internal) markets. It is based on the premise that, since the end of the Second World War, the cost of using international markets has declined substantially which should, ceteris parabus, encourage MNEs to conduct a greater portion of their activities through external markets. Accordingly, the study has 3 main objectives: 1) to determine whether MNEs have in fact increased their use of external markets,
2) to correlate the MNE’s use of external markets with specific factors in the international environment, and 3) to assess the implications of these findings for both the organizational structure of the MNE and the public policies of individual nations.
By focusing on the relationship between two sets of variables (i.e. environmental characteristics and MNE activities), this study clearly falls into the broad class of reasearch known as uCorrelational” or “Ex Post Facto”. Kidder
(1981) notes that in Correlational studies, “...researchers do not manipulate one variable and look at the subsequent effect on another variable. Instead they measure both variables and look at the relationship between the two.” (1981, pg. 27). Thus, Correlational research offers an alternative to true Experimental research, and is frequently used when the investigator is unable to manipulate
126 127 the Independent Variable(s) or assign treatments to subjects, or when the
phenomenon of interest must be studied “after the fact”. Unfortunately, the lack
of direct control over many aspects of the research process subjects this
approach to a number of limitations. Perhaps the most important of these are
the difficulties in distinguishing Cause from Effect, and the problems of choosing
appropriate variables and ruling out rival hypotheses.
In general, Relational studies begin with a thorough review of previous
research in an attempt to identify relevant variables and generate hypotheses
about the relationships between them. Statistical methods are then used to
explain the variablity in the Dependent Variable (the presumed Effect) in terms
of an appropriate set of Independent Variables (the presumed Cause). In this
way, the hypotheses are tested and conclusions are drawn about the strength
and direction of the relationships between variables. However, this approach
cannot be used to establish causality; and therefore, the researcher must build
strong theoretical or empirical arguments to support his hypotheses and
strengthen his conclusions. This has been the primary objective of the
preceding chapters which have presented the theoretical basis for the current
study.
Regression and Correlation Analysis
I. Introduction
Regression and Correlation analysis are the statistical methods most often associated with Relational research. Hays (1988) notes that these tools can be used to answer 3 basic questions: u1) Does a statistical relation affording some predictability appear between the variables X and Y? 128 2) How strong is the apparent statistical relation, in the sense of the possible predictive ability the relation affords?
3) Can an (sic) simple linear rule be formulated for predicting Y and X, and if so, how good is this rule?” (1988, pg. 545)
In the current study, Regression and Correlation will be used to describe the relationship between a group of environmental factors (the Independent
Variables) and several measures of MNE activity (the Dependent Variables).
The purpose of this exercise is to identify those environmental characteristics that have a significant effect on the MNE’s use of external markets, and then use these findings to infer the types of policies nations might employ to encourage a greater use of external markets, and the organizational structures MNEs can employ to improve their performance in specific environmental settings.
Not surprisingly, there are both advantages and disadvantages to adopting this approach. On the positive side, since Regression and Correlation have been widely used by researchers in all fields, it is a fairly straightforward matter to apply these tools and interpret their results. On the other hand, because these methods provide an extremely precise mathematical expression of the relationship between 2 or more variables, their application is contingent upon the satisfaction of a number of formal assumptions that are rarely (if ever) achieved in the real world.
The remainder of this section is devoted to a brief discussion of
Regression and Correlation and the types of problems commonly associated with their use. Because the mathematical basis for these tools is quite complex, only a elementary description of each technique will be presented below.
Readers desiring more complete knowledge about these methodologies are referred to the appropriate references. 129 II. Correlation
Correlation provides a simple mathematical technique for measuring the
strength and direction of the linear relationship between 2 random variables.1
The most widely used measure of correlation is the Pearson Product Moment,
often called simply the “correlation coefficient" (r). According to Newbold (1991),
it can be shown that the value of (r) will always lie between 1 and -1, and can be
interpreted as follows: 1) a correlation of -1 implies a perfect negative linear
association between 2 variables, 2) a correlation of 1 implies a perfect positive
linear association, 3) a correlation of 0 implies no linear association, and 4) the
larger the absolute value of the correlation coefficient, the stronger the linear
association between the random variables (1991, pg. 459).
The Pearson Product Moment is used to measure the correlation
between 2 quantitative variables. However, since researchers work with many
different types of variables, several additional correlation coefficients are also of
interest.2 For example, the Point Biserial (r) is the appropriate measure of
correlation when one of the variables involved isdichotomous (i.e. can assume
only 2 values), while the Phi Coefficient describes the correlation between 2
dichotomous variables. Both of these measures are widely used in Social
Science research since yes or no responses and the absence or presence of a
trait are 2 common dichotomous variables. On the other hand, Spearman’s
Rank correlation coefficient is used to measure the correlation between 2
variables that have been “ranked” in ascending order. This is particularly useful for variables that do not follow a Normal distribution, since the Spearman
coefficient is not influenced by extreme observations. However, Cohen and
Cohen (1983) note that, “...the formulas for point biserial, phi, and rank 130 correlation are simply computational equivalents of the...general formulas for r...They are of use when computation is done by hand or desk calculator. They are of no significance when computers are used...” (1983, pg. 40). Nonetheless, it is not uncommon to see these alternative measures appear in the literature.
In general, researchers are interested in determining the correlation between 2 variables within a population; however, it is often necessary to estimate this value through a random sample drawn from that population.
Accordingly, a distinction is made between the true population correlation coefficient (symbolized by the Greek letter urho”) and the sample correlation coefficient (symbolized by the lower-case letter “r”).3 Although (r) provides an estimate of the true population correlation (rho), these 2 measures will differ in practice due to sampling and measurement error. Thus, researchers typically test the Null Hypothesis that the true population correlation coefficient (rho) is zero. This is done by using the Student’s t distribution with (n-2) degrees of freedom, where n is the number of sample observations.4
III. Regression
Unlike Correlation, which examines only the strength and direction of the association between 2 variables, Regression provides a much richer description of the linear relationship between variables.5 Specifically,
Regression analysis uses information about a variable X (the Independent
Variable or presumed Cause) to explain or predict the behavior of a variable Y
(the Dependent Variable or presumed Effect). This is done by modeling the relationship between X and Y in terms of the mathematical expression for a straight line:
(1) Y = a + bX 131 Thus, the Unear Regression Model uses the expression:
(2) Y = a + bX + e
(where e is an error term) to explain the effect of X on Y.
Equation 2, also known as the “regression equation” or “regression line”, has a number of interesting properties. For example, the “intercept” (a) gives the expected value of Y when X * o, while the “regression coefficient” (b) denotes the expected change in the value of Y corresponding to a 1 unit change in the value of X. In addition, given a specific value of the Independent Variable (X), the equation can be used to predict the corresponding value of the Dependent
Variable (Y).
Clearly, the regression equation provides a powerful analytical tool; however, as with most mathematical tools, it is subject to a number of limiting assumptions. In this case, 4 standard assumptions accompany the Linear
Regression Model. They are: 1) the values of the Independent Variable (X) are independent of the error term (e), 2) the values of the error term (e) are random variables with a mean of 0, 3) these random variables all have the same variance, and 4) these random variables are not correlated with one another.6
Even though the validity of the Regression Model requires that each of these assumptions be satisfied, minor violations are often tolerated in practice.
Since the purpose of Regression analysis is to explain or predict the influence of one variable upon another, it is often desirable to evaluate the explanatory power of the regression equation. This can be done by calculating the Coefficient of Determination (r^), which provides a simple index of the ability of a variable (X) to predict the behavior of a second variable (Y). Specifically,
(r^) is the proportion of the variance in the Dependent Variable (Y) accounted 132 for or explained by the Independent Variable (X), and can be thought of as representing the strength of the linear relationship between X and Y.7 It can easily be shown that the value of (r2) will always lie between 0 and 1, and that the higher the value, the greater the explanatory power of the regression equation. Further, there is an interesting relationship between the Coefficient of
Determination (r2) and the Correlation Coefficient (r) discussed above; in fact,
(r2) = (r)2. Thus, if the correlation between X and Y is, (r) = .5, then the coefficient of determination for the resulting regression equation involving X and
Y will be (r2) = .25. This means that 25% of the variability in Y can be explained by its linear association with X.
Thusfar, we have only considered the relationship between Y and a single Independent Variable (X). This process is called Simple Linear
Regression. However, in many cases the relationship between Y and several
Independent Variables (X-j, X2, etc.) is of interest, particularly since the explanatory power of the regression equation is often increased by the inclusion of additional Independent Variables. Multiple Linear Regression is the appropriate technique for describing the linear relationship between Y and a number of Independent Variables,8 and the general form of the Multiple
Regression Model is:
(3) Y = a + b-j X-| + b2X2 + ... + e where Y is the Dependent Variable, Xi and X2 are Independent Variables, and e is an error term.
In most cases, the meaning of the “multiple” regression equation” is identical to that of the “simple” regression equation discussed earlier. For 133 example, the “intercept” (a) gives the value of Y when all the Independent
Variables are zero; while the values of the “partial regression coefficients” (bi,
b2, etc.) provide a measure of the influence of a particular Independent Variable on Y when the remaining Independent Variables are held constant. In addition, the Coefficient of Multiple Determination (R2) is used to evaluate the explanatory power of the multiple regression equation, but now (R2) measures the proportion of the variance in Y explained by its linear relationship with all of the Independent Variables.
Although the coefficient of multiple determination is an extremely important index, it is often desirable to assess the separate impact of a particular Independent Variable on Y. Unfortunately, this is not a simple task since the Independent Variables used in a multiple regression model are often correlated with each other. As a result, 2 additional measures, the Partial
Correlation Coefficient (prj) and the Semi-Partial Correlation Coefficient (sq), must be used to assess the strength of the linear relationship between Y and a specific Independent Variable (X j). These measures are then used to calculate the Squared Partial and Squared Semi-Partial Correlation [(pr2) and (sr2) respectively] for each Independent Variable in much the same way the correlation coefficient (r) is used to calculate (r2) in Simple Regression.9 In practice, the squared semi-partial correlation (sq2) is the more common measure, and can be thought of as the incremental increase in the total variance in Y explained by the addition of the Independent Variable Xj to a regression model already containing the other Independent Variables. On the other hand, the squared partial correlation (pq2) represents that portion of the 134 variance in Y uoi explained by the other Independent Variables that is uniquely explained by the Variable Xj.
In Regression, as in Correlation, a distinction is made between the population regression equation and the sample regression equation. In general, the researcher is interested in determining the regression equation for a population; however, in practice he typically employs a random sample of observations drawn from the population to estimate this expression. Once again, Greek letters are used to denote the coefficients of the population regression equation, while lower case letters are used for sample coefficients.
In addition, researchers use the sample regression coefficients to test the Null
Hypotheses that the population coefficients (i.e. alpha, beta-j, beta2, etc.) are zero. In the case of Simple Regression (or when a single coefficient is being tested in Multiple Regression), this is done with the Student’s t distribution. In
Multiple Regression, however, the F distribution is used to test the Null
Hypothesis that all the regression coefficients are zero.10 Further, theF distribution can be used to test the significance of the coefficient of multiple determination (R^), and the Null Hypotheses that the various population partial and semi-partial correlations are zero.11
Finally, it must be noted that a number of common problems are associated with Regression analysis. Hays (1988) discusses many of these in detail, but in terms of the present study it important to recognize that: 1) the technique is only as good as the data being used, 2) Regression analysis measures only the linear relationship between variables, and cannot be used when the variables of interest are related in a non-linear way, 3) the use of 2 or more Independent Variables that are highly correlated (multicollinearity) should 135 be avoided in Multiple Regression, 4) the use of a large number of Independent
Variables relative to the number of observation in the data set should also be avoided, and finally 5) when analyzing the results of a multiple regression, the effect of a specific Independent Variable must be interpreted relative to the other variables employed.12
Some Previous Studies
Before proceeding with a detailed description of the present study, it may
be useful to examine some previous studies that have employed Regression and Correlation analysis. Both of these techniques have been widely used in
International Business research to investigate a wide range of international problems. But while these tools have been widely used, they have not always been properly used. For example, due to the technical problems discussed above and difficulties in interpreting and generalizing results, studies which utilize Regression analysis are sometimes more confusing than illuminating.
Therefore, this brief section will examine 4 relevant studies which demonstrate both the strengths and weaknesses of these statistical tools.
Dunning (1979) provides an ambitious illustration of the use of
Regression analysis in his empirical test of the “Eclectic Paradigm”. The study attempts to evaluate the ability of certain ownership and location-specific factors to explain the industrial and geographic distribution of U.S. affiliates in 14 industries in 7 countries. By focusing on the relationship between these factors, the study is clearly Relational in nature, and therefore suitable for Regression analysis. As noted above, Relational studies often begin with a review of previous research in an attempt to identify relevant variables and generate research hypotheses. In this case, the Independent Variables were selected 136 from 3 broad "classes” of variables (ownership-specific, location-specific, and general performance indicators), while various measures of foreign participation served as the study’s Dependent Variables. Dunning uses the
Eclectic Paradigm to develop specific research hypotheses, and then applies
Multiple Regression analysis to test their validity. He concludes that the main advantages of U.S. firms are revealed by one location variable (market size) and one ownership variable (the ratio of skilled to unskilled labor).
Although Dunnning’s study provides an excellent example of the power of Regression analysis, it also demonstrates many of the difficulties associated with it. One of the most important of these is the operationalization of the study’s research hypotheses. As noted above, Regression analysis requires that a number of technical and conceptual conditions be fulfilled. In addition, particularly in studies involving Multiple Regression, a relatively large number of observation is necessary in order to generate reliable and meaningful results.
Unless both of these limitations are addressed, the internal and external validity of the study may suffer accordingly. Unfortunately, due to the extremely complex nature of the relationships being explored and the relatively small amount of empirical data available, it is often difficult to interpret the results of Dunning’s study. But perhaps even more importantly, it is extremely difficult to use such a precise statistical tool as Regression analysis to test a theory as general as the
Eclectic Paradigm.
Dunning acknowledges many of these limitations in a brief review of prior empirical studies of international production. He notes that:
“...these strands of research have much in common. Each uses, with varying degrees of sophistication, multiple regression analysis to test explanations about the relationship between various measures of international involvement and a variety of explanatory variables. Each, too, is beset with the same kind of 137 methodological and statistical problems, notably the establishment of operationally testable hypotheses, data limitations, and multicollinearity between the individual variables.” (1979, pg. 12).
Thus, Regression analysis can be a very powerful methodology but only when applied properly.
Grant’s (1987) study of the relationship between “multinationality” and performance avoids many of these pitfalls by focusing on a very limited aspect of a well defined problem. Specifically, Grant uses Regression analysis to describe the relationship between a firm's multinationality (as measured by the ratio of sales revenue from overseas operating subsidiaries to total firm sales) and several measures of corporate performance (Return on Net Assets, ROE, and ROS). Because these relationships are so clearly defined, Grant has little trouble building the theoretical arguments necessary to justify the selection of variables and the interpretation of the study’s results. In addition, by including data on a large number of individual companies, Grant greatly increases the likelihood of identifying significant relationships between variables.
Consequently, by investigating a large sample of British manufacturing companies, Grant concludes that multinationality is positively associated with performance. Thus, by restricting the scope of his study to accommodate the limitations of Regression analysis, Grant greatly improves the strength of his findings.
Compared to Regression, Correlation is a somewhat less powerful technique. Although both methods can be used to describe the relationship between variables, the information contained in the regression equation is much more detailed than that obtained from a correlation matrix. On the other hand, when only the strength and direction of the relationship between 2 138 variables is necessary, Correlation provides an appropriate statistical tool. This is clearly illustrated in a recent study by Egelhoff (1984).
Egelhoff uses Correlation to describe the relationship between the nationality of an MNE and the type of managerial control it exercises. By assessing the correlation between various control variables and certain firm- specific characteristics (such as the age and size of the subsidiary, the firm’s total number of subsidiaries, and the number of years the firm has operated abroad), Egelhoff comes to the interesting conclusion that the nationality of a firm is strongly associated with a certain type of managerial control. Specifically, he notes that U.S. MNEs exert control over their foreign subsidiaries by requiring reported data, while European MNEs exercise control by assigning parent country nationals to key positions in foreign subsidiaries. Thus, the simple technique of Correlation can be used to identify extremely complex relationships when applied properly.
On the other hand, Gates and Egelhoff (1986) employ both Correlation and ANOVA (Analysis of Variance)13 to explore the relationship between the centralization of decision-making and a large group of firm-specific factors. In this study, 2 relatively simple statistical tools are used to test a number of highly complex research hypotheses. Given the nature of the relationships being explored, it could be argued that Regression would have been the more appropriate statistical tool, since it provides more detailed information about the relationships being examined and can be used to describe the relationship between several Independent Variables and a given Dependent Variable. Even though the authors arrive at a number of interesting conclusions, there is little doubt that Regression analysis could have greatly enriched their findings. Thus, the researcher's choice of a statistical methodology is an extremely important 139 one that can greatly affect the strength and validity of his conclusions.
Some Methodological Concerns
As noted earlier, this study will employ Regression and Correlation analysis to explore the relationship between a number of key environmental factors and the MNE’s use of external markets. However, as the previous sections so clearly illustrate, in order to properly apply these statistical tools, it is imperative that the specific limitations of each technique be recognized and addressed. Within this context, a number of concerns may be raised about the nature of the relationships being investigated and the quality of the data being employed in this study.
Perhaps the most fundamental limitation of Regression and Correlation analysis is that these tools can only be used to describe the linear association between variables. Unfortunately, many potentially interesting relationships are non-linear and must be analyzed through other, more sophisticated methodologies. However, in practice, non-linear techniques are extremely difficult to apply and interpret; and consequently, researchers often assume that a linear relationship is present until proven otherwise. The assumption of linearity” will also be made in the current study; but it is quite possible that some of the relationships being examined are not linear, and that Regression and Correlation will give a misleading indication of their true significance. Thus, it is important to remember that variables which appear to be unrelated or only slightly related using these linear techniques, may instead be related in a non linear fashion.
Although other technical concerns are associated with the use of
Regression and Correlation, they are potentially less worrisome than the 140 practical concerns associated with data collection. Since a statistical methodology is only as good as the data it employs, the problems of data quality and availability are serious indeed. This is particularly true since there is a dearth of accurate and reliable information on many types of MNE activities.
For example, detailed information on the operations of U.S. Multinationals has only been collected on a regular basis since the early 1980s, and even these figures do not specify the extent of many “New Forms of Investment”. For countries other than the United States, data on the international operations of domestic MNEs is difficult to find; and when it does exist, it is often not comparable to analogous data from other nations. Therefore, in developing the current study, a number of difficult decisions had to be made in order to develop research questions and hypotheses that could be realistically tested through available data resources.
Within this context, four major problems emerge: 1) the lack of reliable data over a substantial period of time eliminates the possibility of time-series analysis, 2) the lack of data on the operations of MNEs from a large number of source countries makes it necessary to restrict the scope of the study, 3) the format of existing data resources requires the creation of several unique
Independent and Dependent variables, and 4) the use of archival or “second hand” data creates a number of additional concerns.
Clearly, because this study looks at changes over time, the use of time- series analysis or at least the comparison of MNE activities at several specific points in time is desirable. But while limited information on the operations of
U.S. MNEs since the mid-1960s is available, this data is not detailed enough to serve as the study’s primary data source. As a result, the examination of changes in the international environment over time must be replaced by an 141 exploration of differences between national environments at a point in time.
Accordingly, the study’s basic research question must be restated to reflect this modification. Instead of asking whether certain changes in the international environment have influenced the MNE’s preference for external markets, the new research question becomes, ‘Do MNEs make greater use of external markets in countries with certain environmental characteristics?’.
Even though this change is subtle, it is by no means insignificant. On one hand, it is far from clear whether the effect of specific environmental variables is the same across countries as it is across time. Even if it is, it is not entirely clear how to relate differences between the external environments of various countries to changes in the international environment since WWII. On the other hand, the study of the effects of national differences on local MNE activity is an interesting and important problem in its own right. Further, it may be possible to identify similarities between specific national environments and conditions that existed in the international environment at various points in time, and thereby infer the study’s desired relationships. However, regardless of the consequences, this change must be made in order to employ existing data sources in a meaningful way.
The second problem stems from the lack of accurate, comparable data on the international operations of MNEs domiciled in different countries. For example, the United Nations Center for Transnational Corporations and the
OECD collect information on the Worldwide activities of MNEs, but the type of data required for the present study is not available from these sources. Further, the use of information from various foreign government agencies is also impracticable, since relatively few nations collect the required data, and data collection methods vary widely between countries. As a result, this study must 142 be further restricted to an examination of the international operations of U.S.
MNEs. Luckily, there are several practical benefits to accepting this limitation.
First, it allows the Commerce Department’s 1989 Benchmark Survey of
U.S. Direct Investment Abroad (USDC, 1991) to serve as one of the study’s
primary data sources. This document contains detailed operating and financial information on the non-bank foreign affilates of non-bank U.S. parent corporations in over 50 countries, and in all major industries, in addition, the
Commerce Department has collected this data on an annual basis since 1983, and on a sporadic basis since 1966; making these surveys the most accurate and complete source of information on (U.S.) MNE activity in the World. Further, it is not unreasonable to assume that the operations of U.S. multinationals are largely representative of the operations of all MNEs. Boyacigiller and Adler
(1991) note that U.S. MNEs dominated the international economy for a large portion of the post-war era, and as a result, U.S. management practices and theories often characterize the operations of foreign MNEs as well. Thus, it was deemed better to employ complete information on the potentially representative activities of U.S. MNEs, than to use incomplete information on the activities of a sample of all MNEs operating in a particular country.
Unfortunately (though not surprisingly), the Commerce Department does not directly measure many of the activities addressed in this study. In particular, information on “New Forms of Investment” and the MNE’s use of internal and external markets does not appear in these surveys. Consequently, 4 measures of MNE activity (one focusing on intra-firm trade, another on local direct investment, a third measuring sales to non-affiliated parties, and a fourth related to local employment) must be created from the Commerce Department’s data in order to obtain an accurate picture of the extent to which U.S. MNEs use 143 external markets to conduct their local operations. These 4 measures (which are discussed in detail below) will serve as the study’s Dependent Variables.
A final limitiation applies to all studies that employ archival (or “second
hand”) data. Kidder (1981) notes that while, “Archival records afford the opportunity to assess the impact of natural events; they also elicit particular problems of interpretation.” (1981, pg. 289). Specifically, Kidder cites the problems of ruling out alternative hypotheses and inadequacies in the initial data collection process. However, since the researcher has no control over the type of data requested or the manner in which it is collected, archival data may also encourage modifications of a study in order to accommodate a particular data source. Unfortunately, the questions being raised in the present study make the use of second-hand data unavoidable; and consequently, a number of compromises have been necessary. Fortunately, these compromises will make it possible to apply the extremely powerful tools of Correlation and
Regression to the highly detailed and reliable data collected by the U.S.
Department of Commerce. As a result, this study should provide valid, easily interpretable information on an extremely important (albeit limited) aspect of the relationship between the MNE and the environment.
The Research Question
As a result of the technical and practical limitations detailed above, it is now possible to pose the following Research Question:
“How do various environmental factors influence the way in which U.S. MNEs structure their operations in a particular country?”. Specifically: “Do certain environmental factors affect U.S. MNEs’ preference for using external markets to coordinate their local operations?”. 144 But before proceeding with a detailed description of the current study, it is worthwhile asking whether this question warrants further investigation.
Lundstedt (1968) has developed a fourfold test for evaluating research questions which asks: 1) is the question important, 2) is it closely reasoned, 3) where does it fit in the context of human survival and development, and 4) is it logically stated and fitted to accepted (research) standards (1968, pg. 229).
Applying Lundstedt’s criteria to the question advanced above it seems reasonable to conclude that it is both important and well reasoned, and that it fits the practical limitations of available data as well as accepted forms of data analysis. Therefore, even though the study’s basic research question has been operationalized in a somewhat different form than the one presented in the first chapter, it still merits further examination, and will provide valuable information on a highly complex aspect of MNE activity.
The Current Study
I. Overview
The remainder of this chapter is devoted to a detailed description of the current study. As previously noted, this study will use Regression and
Correlation to explore the relationship between a number of key environmental variables (the Independent Variables) and certain types of U.S. MNE behavior
(the Dependent Variables). (The study’s Independent and Dependent Variables are discussed in separate sections below) The purpose of this investigation is to assess the impact various “country-specific” factors have on the way U.S. MNEs structure their operations in a given country. In particular, an attempt will be made to identify environmental characteristics that encourage MNEs to conduct their local activities through external (rather than internal) markets. The results 145 of this analysis will then be used to infer the types of organizational responses
Firms might use in specific environmental settings, and the kinds of public policies Nation’s might adopt to promote economic growth.
II. Research Design
The approach utilized in this study is typical of most “Relational” or “Ex
Post Facto” research, and involves the use of statistical methods to explain the variability in the study’s Dependent Variables (DVs) in terms of an hypothesized set of Independent Variables (IVs). Based on the discussion developed in
Chapter II, it has been suggested that environmental factors play a critical role in determining the manner in which MNEs enter and service a particular foreign market. Therefore, this study will attempt to explain U.S. MNEs’ preference for using external markets in a given country in terms of a small number of country- specific factors. This will be accomplished by examining the aggregate behavior of U.S. MNEs in 52 countries, and employing Regression and Correlation analysis to describe the relationship between 4 measures of external
Multinational activity (the DVs) and 12 key environmental characteristics (the
IVs).
III. Subject Selection
As noted above, in order to address both the technical limitations of
Regression and Correlation analysis and the practical limitations of data availability, the focus of this investigation is individual Nations, nol individual
Firms. Therefore, the potential universe of subjects is necessarily limited to the
170 or so Nations of the World. But while this group forms the study’s theoretical universe, it must be reduced considerably in order to accommodate available data sources (specifically, the U.S. Department of Commerce’s annual survey of
U.S. Direct Investment Abroad). As a result, the study’s “target” population has 146 been limited to the 58 countries which host a substantial amount of local U.S.
MNE investment.14 In general, these nations host investments by non-bank
U.S. parents which total over $1 billion, and represent a broad cross-section of
Developed and Developing Countries from every geographical area.
Unfortunately, complete information on all of the study’s Independent and
Dependent Variables was unavailable for 6 of these countries (Taiwan,
Bahamas, Barbados, Bermuda, Netherlands Antilles, and U.K. Islands),
effectively eliminating them as potential “subjects”. As a result, the 52 countries
listed in Table 2 represent the study’s final “data” sample, but for all practical
purposes this group can be considered the “target” population as well. This
reduction in sample size should have little impact on the results of this study for
2 reasons: first, only 3 of these countries (Bermuda, Netherlands Antilles, and
Taiwan) host a substantial amount of U.S. direct investment; and second, 5 of them are Caribbean Islands.
In both cases, there is little reason to beleive that the 6 excluded countries differ significantly from other subjects included in this study. This is
particularly true in the latter case since 3 Caribbean nations (Dominican
Republic, Jamaica, and Trinidad and Tobago) are already included in the study.
Further, a recent study by Beamish (1987) used a sample of Caribbean nations to represent the broader group of Developing Countries thus implying that the characteristics of these two groups are similar. Since many Developing nations are examined in the current study, it is reasonable to suggest that, at least in 5 of the 6 cases, the characteristics of the excluded countries are represented by other countries in the study. Finally, and perhaps most importantly, there is little that can be done to correct this problem since comparable data for these 6 countries is simply not available. 147 IABLL2 The Study’s 52 Subject Countries
1) Argentina 27) Jam aica 2) Australia 28) Japan 3) Austria 29) Korea (Rep.) 4) Belgium 30) Luxembourg 5) Brazil 31) Malaysia 6) C anada 32) Mexico 7) Chile 33) Netherlands 8) China 34) New Zealand 9) Columbia 35) Nigeria 10) Costa Rica 36) Norway 11) Denmark 37) Panam a 12) Dominican Republic 38) Peru 13) Ecuador 39) Philippines 14) Egypt 40) Portugal 15) Finland 41) Saudi Arabia 16) France 42) Singapore 17) Germany (Fed. Rep.) 43) South Africa 18) G reece 44) Spain 19) Guatemala 45) Sw eden 20) Honduras 46) Switzerland 21) Hong Kong 47) Thailand 22) India 48) Trinidad and Tobago 23) Indonesia 49) Turkey 24) Ireland 50) United Arab Emirates 25) Israel 51) United Kingdom 26) Italy 52) Venezuela 148 There is a major advantage to defining the data sample as the target population, since it is now possible to examine the entire population rather than a random or purposive sample thereof. This is important because it means that all of the study’s statistical tests will yield true population parameters rather than estimates of those parameters. This makes it unnecessary to use inferential statistics (t-tests, confidence intervals, etc.), which not only considerably reduces the number of statistical tests required in this study, but also greatly increases the strength and validity of the findings.
IV. Data Collection
Due to the nature of this investigation, it will be necessary to rely on outside sources for all of the study’s basic data. While there are clearly advantages to adopting this approach, there are also a number of unique problems associated with it. In particular, since the researcher has no direct control over the type of data collected or the manner in which it is acquired, every effort must be made to insure that all of the raw data is accurate and reliable. Several steps have been taken to guarantee this eventuality. For example, many of the study’s Independent Variables have been constructed from major Economic and Social indicators compiled by the World Bank, the
International Monetary Fund, and the United Nations. As a result, reliable recent data is available for most of the countries in the World. In addition, in almost every case, all of the data for a particular variable comes from a single source, thus reducing the problems of consistency and comparability.
However, as the World Bank (1991) notes in its recent World Bfi.v.9lQpment R eport (1991):
“Considerable effort has been made to standardize the data; nevertheless, statistical methods, coverage, practices, and definitions differ widely among countries. In addition, the statistical systems in many developing economies are 149 still weak, and this affects the availability and reliability of the data. Moreover, cross-country and cross-time comparisons always involve complex technical problems, which cannot be fully and unequivocally resolved. The data are drawn from the sources thought to be most authoritative, but many of them are subject to considerable margins of error.” (1991, pg. 270)
On the other hand, preliminary results from the Commerce Department’s recent Benchmark Survey of U.S. Direct Investment Abroad will provide information on the study’s Dependent Variables. Since all U.S. companies with foreign assets or activities over $3 million are required by law to report certain financial and operating data to the Commerce Department, this survey is not subject to many of the limitations described above. In fact, according to the
Bureau of Economic Analysis which is responsible for the survey:
"Benchmark surveys are the most comprehensive surveys, in terms of both coverage of companies and subject matter, of U.S. direct investment abroad conducted by the Bureau of Economic Analysis (BEA). The 1989 survey covered all foreign affiliates of U.S. direct investors (foreign companies owned 10 percent or more by a U.S. person) that had assets, sales, or net income of more than $3 million.” (USCD, 1991; pg iii)
For all practical purposes, this source provides accurate information on the activities of all foreign non-bank affiliates of non-bank U.S. parents for a given year, and represents an extremely powerful database. In addition, for the first time the preliminary results (usually published between 12-18 months before the final results) include estimates of data from reports that had not been received or processed in time for publication. This makes it possible to use data from the 1989 survey rather than the previous benchmark survey conducted in
1982.
Thus, the current study will use archival data compiled by major national and international agencies for the most recent available year, 1989. Every 150 attempt has been made to find and use the most authoritative resource for each of the study’s Independent and Dependent Variables; and while some caution may be warranted in accepting this approach, given the nature and scope of this investigation these concerns must be considered minor.
One final note on missing data is in order. Even under the best of circumstances, data points are occasionally missing or unusable; and therefore, some procedure must be developed to address this problem. The 3 most common methods are: 1) using the variable’s average value as the missing value, 2) using the variable’s value for a similar country as the missing value, and 3) estimating the missing value using some other method. In the current study, the following procedures will be used when a missing data point occurs:
First, if the 1989 value for a particular variable is not available for a given country, the country’s 1988 value will be used, and if possible, adjusted to reflect the variable’s trend over the last 3 years. If 1988 data is also missing,
"similar country” data will be used instead. This involves identifying 1-3 countries with a similar economic and political profile in the same geographic region, and using the variable’s average value for these countries as the missing value. Finally, if neither of these methods is feasible, the variable’s average value for all the countries with data will be used. In any case, whenever a missing value occurs in this study the precise method of estimating its value will be given in a footnote.
V. Data Analysis
The statistical methodologies used in this study are straightforward and have been discussed at length elsewhere. Basically, they involve the use of
Correlation analysis to explore the relationship between the study’s
Independent and Dependent Variables, and the use of Multiple Regression 151 analysis to explain the variability in each of study’s 4 DVs through a linear
model containing several IVs.
The first step in this process is the creation of a Correlation Matrix which
includes ail 12 of the study’s IVs and all 4 of its DVs. The correlation matrix provides an indication of the strength and direction of the linear association between variables, and can be used to identify the most appropriate
Independent Variables for each regression model. This is done by evaluating the relationship between each IV and a given DV, and then selecting the IVs with the highest correlation coefficient (r) for inclusion in the model. According to
Davis (1971): .01 < r <, .09 represents a “negligible" relationship; .10 < r < .29 indicates a “low" association; .30 £ r £ .49 indicates a “moderate" association;
.50 £ r £ .69 is a “substantial” association; and .70 £ r £ 1 indicates the presence of a “very strong” association. Whenever possible, only IVs with a correlation (r) of .40 or greater will be used in the regression model for a particular Dependent
Variable. In addition, IVs that are highly correlated with each other will not be used in the same model. These restrictions are required because the rather small number of subjects included in this study limits the total number of IVs that can be used in each model to 3 or 4.15
in order to facilitate the construction of these regression models, the
Independent Variables have been divided into 3 “sets” containing 3-5 variables each (this process is explained in greater detail below). Ideally, each regression model will contain one IV from each set; however, if no variable in a set is moderately correlated with the Dependent Variable, a more highly correlated variable from another set will be used in its place. There are 2 major advantages to adopting this approach. First, since each set contains variables with similar characteristics, the members of a set will tend to be highly 152 correlated with each other, but not with the variables in other sets. Thus, by selecting only one (or 2) variable(s) from each set, it should be possible to considerably reduce the problem of multicolinearity. Secondly, by using the most powerful variable(s) from each set it should be possible to construct a highly robust model from only 3-4 Independent Variables.
Once a suitable group of predictor variables (IVs) has been selected and the multiple regression equation for each of the DVs has been calculated, it is a simple matter to determine the amount of variance (R2) explained by each model as well as the individual contribution of each of the predictor variables. In this way, it is hoped that several aspects of external U.S. MNE behavior can be explained through the influence of a small number of environmental factors.
VI. Threats to Validity.
Every researcher must answer 2 major questions concerning a study’s
Validity: first, 'Are the findings consistent with the research hypothesis or can they be attributed to some other cause?’ (the question of Internal Validity); and second, 'Can the findings be generalized; and if so, to what group?’ (the question of External Validity). Campbell and Stanley (1963) discuss these issues in great detail and develop a number of research designs which minimize various threats to Validity. Unfortunately, none of these designs is particularly useful in the current study; and therefore, questions of Validity must be addressed in a somewhat different way. For example, several steps have been taken to increase the internal Validity of this study. These include the examination of an entire population which eliminates threats associated with sampling, and the focus on describing various relationships within that population rather than testing the effect of a specific treatment. As a result of 153 these measures, this study should indeed describe the relationship between the variables being examined within the population under investigation. The only limitations arise from the accuracy of the data itself.
But while threats to the study’s Internal Validity are relatively minor, the problem of interpreting and generalizing any of the findings is much more serious. As noted above, this study should describe the relationship between a number of country-specific factors and several measures of U.S. MNE activity; however, the hope is to apply this information to the areas of International
Strategic Management and Public Policy formulation. Therefore, extreme care must be taken when interpreting the study’s results in terms of these broader issues. For this reason, the conceptual background developed in the preceding chapters will serve as a guide to the interpretation of the study's statistical analysis. Further, an attempt will be made to limit any conclusions to the study’s target population (those countries which attract a relatively large amount of U.S.
MNE activity) and the aggregate behavior of U.S. MNEs within that population.
Independent Variables
I. Introduction
Neter, et. al. (1985) note that, “One of the most difficult problems in regression analysis often is the selection of the set of Independent Variables to be employed in the study.” (1985, pg. 417). Since the potential number of environmental factors that may affect MNE behavior is virtually limitless, this statement certainly applies to the current study. In addition, due to the relatively small size of the study’s target population, the number of Independent Variables
(IVs) that can be included in any of the regression models is severely restricted.
As a result, the selection of an appropriate group of factors for use as the study’s 154 Independent Variables is an extremely difficult and important task.
In order to facilitate this process, the universe of potential IVs has been divided into “Sets” or “Classes”. According to Cohen and Cohen (1983), “IVs are grouped into sets for reasons of their substantive content and the function they play in the logic of the research” (1983, pg. 135).16 In the current study, environmental (country-specific) factors have been conceptually divided into 3 distinct groups: Economic, Social, and Political. There are many precedents for this method of classification, particularly in the literature on Economic
Development. For example, as early as 1967, Adelman and Morris (1967) had compiled a comprehensive list of 42 Social, Political, and Economic Variables which they used to assess major aspects of a nation’s external environment.
Based on this framework, 3-5 representatives from each “Set” have been selected to serve as the study’s Independent Variables. The remainder of this section is devoted to a detailed description of each of these variables, including the rationale for including it in the study, the data source to be employed, and the hypothesized effect each variable will have on external MNE behavior. The following list represents the study’s 12 Independent Variables:
Economic Variables 1) PCGNP (Per capita GNP) 2) GDPGRO (Average annual GDP growth rate) 3) TOTRADE (% total trade to GDP) 4) GOVEXP (% government expenditures to GNP) 5) INVEST (% domestic investment to GDP)
Social Variables 1) PQLI (Physical quality of life index) 2) LIT (Literacy rate) 3) DOCTOR (1000 people per doctor) 4) %AGR (% population in Agricultural sector) 155 Political Variables 1) POL (Political system) 2) REST (Restrictions on foreign business and FDI) 3) PRI (Political risk index)
II. Economic Variables
One of the first steps in evaluating the suitability of a particular foreign country for MNE investment is to conduct a detailed examination of its Economic environment. Cavusgil (1985) provides a typical list of variables that can be used for this purpose in a recent article. They include: the Overall Level of
Development of the Host Country, its Economic Growth (GNP), the Role of
Foreign Trade in the Economy, Stability of Exchange Rates, Inflation Rate,
Balance of Payments, Per Capita Income and Distribution, and Disposable
Income and Expenditure Patterns. Once the decision to enter a foreign market has been made, the next step is to decide which mode of entry should be employed. While many factors are involved in this decision, some authors argue that the Host country’s level of development plays a critical role in the firm’s choice between using a Wholly Owned Subsidiary or Majority Owned Affiliate, or a Minority Joint Venture or "New Form” to enter and service a particular foreign market.
Rostow was one of the first writers to examine the relationship between
FDI and Development in his Stages of Economic Growth (1960). According to
Rostow, a country typically goes through 5 distinct stages of economic growth: the Traditional Society, Preconditions for Take-off, Take-off, the Drive to Maturity, and the Age of Mass Consumption. Dunning (1988b, 1989) uses Rostow’s model to create an “International Direct Investment Cycle" which explains a country’s propensity to engage in inward or outward FDI in terms of its stage in 156 the development cycle, the structure of its factor and market endowments, its economic and political systems, and the extent of its market imperfections.
According to Dunning, there is little role for foreign investment in the initial stage of development; however, by the second stage (beginning when a nation's Per Capita GNP reaches around $500) foreign MNEs may find local markets attractive for some types of investments. Since markets are usually extremely imperfect in these early stages, MNEs will prefer to use internal
(rather than external) markets in these countries in order to avoid high transaction costs, and to exploit local factor advantages through rationalized production. As countries move into the third stage of development, local firms begin to generate their own advantages, and outward FDI or export activities start to occur. The fourth stage begins when a country becomes a net outward investor; while the fifth stage is characterized by 2 additional features: the
“ownership” advantages of local firms become more firm-specific and less country-specific, and location decisions by MNEs become more strategic in nature and less influenced by comparative advantage.
In terms of the current study, Dunning’s argument that 'at higher stages of development, the rationale for using licensing, minority joint ventures, and other
“New Forms" to enter a given market increases accordingly’, is extremely important since it suggests that economic factors that describe a country’s level of development and/or the efficiency of its marketplace may also explain the method U.S. MNEs use to enter that market. Based on this discussion, 5 variables have been selected to provide an indication of a nation’s economic environment (PCGNP, GDPGRO, TOTRADE, GOVEXP, and INVEST). A brief description of each of these variables is given below. 157 lla. PCGNP (Per capita GNP)
GNP, and Per Capita GNP are perhaps the most widely used measures of a country’s economic strength and stage of development. GNP measures the total foreign and domestic Hvalue-added” claimed by a nation’s citizens; however, since GNP does not account for the “size" of a given economy, it is often adjusted for population in order to facilitate direct comparisons between countries. Thus, the study’s first independent Variable is per capita GNP
(PCGNP) which is simply (the value of GNP in $U.S. of County X) divided by
(the population of Country X). The study will use 1989 values for this variable taken from Table 1 of the World Development Report (World Bank, 1991). The
World Bank notes that 2 problems may affect the comparability of these figures between countries. The first is inaccuracies in the Population and GNP estimates themselves, and the second involves the use of official exchange rates to convert local GNP figures to U.S. dollars.17 While these problems are certainly worth noting, the World Bank has made every effort to compile the most accurate available figures; and given the scope of the current study, these concerns must be considered minor. Finally, since higher levels of PCGNP tend to be associated with higher levels of economic development, this variable should be positively correlated with U.S. MNE’s use of external markets in a particular country. lib. GDPGRO (Average annual GDP growth rate)
Unlike GNP, which measures the total value of all goods and services produced in a country plus net income from foreign sources, GDP measures only the value of economic output within a given country. As “The Economist"
(Sept, 12,1991) notes, “GNP is probably more useful in comparing the relative levels of income per head in different countries, but GDP provides a better 158 guide to changes in domestic production...” (1991, pg. 33). Thus, GDPGRO provides a measure of the internal growth and vitality of a particular economy.
Specifically, the variable is (the average annual growth rate from 1980-1989 of the GDP of Country X). Values for this variable were taken from Table 2 of the
World Development Report. Unfortunately, there is no clear relationship between growth and development; in fact, during the 1980s it was the Low- income economies that had the highest levels of GDP growth, while the OECD nations (i.e. the Developed Countries) experienced much lower growth rates.18
Thus, since higher values of GDPGRO have tended to be associated with less developed economies, this variable should be neoativelv associated with the use of external markets by U.S. MNEs. lie. TOTRADE (% total trade to GDP)
Since a country’s trade pattern provides an indication of both its economic vitality and the efficiency or "openness" of local markets, some measure of trade should be included as an Independent Variable in the current study. However, examining only the level of exports or imports can give a misleading picture of the importance of trade to a particular country since these figures are affected by the size of the local economy, local factor endowments, and other similar differences. Thus, in order to allow meaningful comparisons between nations, TOTRADE has been constructed so that the size of the local economy will not affect the value of the variable. Specifically, TOTRADE is (the sum of merchandise Imports plus Exports for Country X) divided by (the GDP of
Country X). Tables 3 & 14 of the World Development Report were used to calculate this figure for the year 1989. It seems reasonable to suggest that higher values of TOTRADE will be associated with more open and efficient local markets; and therefore, this variable should be positively correlated with U.S. 159 MNEs use of external markets. lid. GOVEXP (% government expenditures to GNP)
The recent international trend toward “Free Markets" has reinforced the view that Government intervention in the economy, regardless of its form or intended outcome, promotes market inefficiencies. On the other hand, many recent studies on Industrial Policy seem to suggest that, when used properly, government spending can have a positive effect on the economy19 In either case, it seems clear that some measure of the Government’s involvement in the local economy should be included in the current study. Accordingly, GOVEXP has been selected to provide a broad indication of the "size” of Government relative to the "size” of the local economy. GOVEXP measures [the portion of a nation’s GNP accounted for by expenditures for goods and services (including expenditures on national defense) by all levels of government]. For most countries, 1989 values for this variable were taken from Table 11 of the World
Development Report. However, value for 6 countries (China, Greece, Honduras,
Jamaica, Luxembourg, and Switzerland) had to be obtained from a second source, the IMF’s International Financial Statistics Yearbook (1990), and cover various years from 1987-89.
Meltzer (1991) notes that "A problem in using spending to measure the size of government is that governments spend for different purposes. Public spending for defense and police protection are classic examples of public goods...On the other hand, in many countries that seek to reduce government spending, relatively large reductions are made in spending for infrastructure...” pg. 135. Meltzer’s point is important because spending on infrastructure can clearly influence both an MNE’s decision to invest in a particular country and the mode of entry it chooses. Nonetheless, based on the current strength of 160 “Reaganomics”, it will be hypothesized that higher values of GOVEXP will be associated with greater market inefficiencies; and therefore this variable should be negatively correlated with U.S. MNEs’ use of external markets, lie. INVEST (% gross domestic investment to GDP)
According to Todaro (1989), “In order to grow, economies must save and invest a certain proportion of their GNP. The more they can save, and therefore invest, the faster they can grow.” (1989, pg. 66). Thus, a nation’s investment rate is likely to affect both an MNE’s decision to enter a particular market and the mode of entry it uses. INVEST is simply (gross domestic investment as a percent of GDP), and 1989 values for this variable appear in Table 9 of the
World Development Report. It could be argued that higher levels of domestic investment improve the competitiveness of local firms and therefore the efficiency of local markets. If this is true, INVEST should be positively correlated with U.S. MNE’s use of external markets; however, this hypothesis should be viewed with caution. Since investment and growth tend to be highly correlated, and since Developing Countries have experienced the highest levels of growth in recent years, high levels of domestic investment may be associated with earlier stages of the development process. Based on Dunning’s observation that MNEs prefer to use internal markets in these countries, INVEST may be negatively correlated with measures of external market activity instead.
III. Social Variables
Economic variables provide information on only one aspect of a country’s local environment. In fact, many writers have discussed the problems involved with using per capita GNP and similar economic variables to assess a country’s level of development since these measures fail to include various types of 161 production (such as subsistence agriculture and housework) or to address the
issues of social welfare and income distribution.20 As a result, a number of
Social indicators have been included in this study in order to describe another dimension of a country’s external environment.
Todaro (1989) discusses the evolution of various Social Indicators in detail in his book Economic Development in the Third World 21 After reviewing
major studies by Adelman and Morris (1967), the United Nations Research
Institute on Social Development (1970), Morris (1979), and Weigal (1989) a wide range of potential Social variables can be identified. These range from
Life Expectancy and Educational indicators, to Caloric intake per person per day, Urbanization indexes, and Newspaper circulation. From this list, 4 variables were chosen for use in the present study. A brief discussion of each of these is given below.
Ilia. PQLI (Physical quality of life index)
A number of studies have attempted to devise an index capable of expressing a country’s level of development in terms of its ability to meet the basic needs of the majority of its citizens. One of the most successful of these efforts is the “Physical Quality of Life Index” (PQLI) developed by Morris (1979).
The PQLI is a composite index consisting of 3 basic Social indicators (Life
Expectancy at Age 1, Infant Mortality, and Literacy), and is calculated by ranking a country on a scale of 1-100 (worst to best) for each of the component indicators and taking the simple arithmetic average of these 3 figures.
According to Todaro (1989):
“The PQLI appears on the surface to be free of the major problems associated with using GNP as a measure of development. It aims directly at incorporating welfare considerations through measuring the ends of development in terms of the quality of human life. It also incorporates distributional considerations by 162 using three indicators that reflect distributional characteristics in the sense that countries cannot achieve high national averages of life expectancy, infant morality, and literacy unless the majority of their populations are receiving the benefits of progress in each of these areas.” (1989, pg. 111-112).
Further, the PQLI is only moderately correlated with per capita GNP. This suggests that it measures a somewhat different group of country-specific factors.
Values for this variable appear in Table A2.2 of Todaro (1989) and are for the year 1985. (It was necessary to use 1985 data because 1989 figures were unavailable for one of the component indicators) Since higher values of PQLI are associated with higher levels of development, Dunning’s “Investment Life
Cycle” suggests that this variable should be positively correlated with the use of external markets by U.S. MNEs.
Illb. LIT (Literacy rate)
Even though the Literacy Rate is a component of PQLI discussed above, there are several reasons for including it as a separate Independent Variable in this study. Perhaps the most important is the relationship between literacy and the skill-level of the local workforce. Since higher value-added activities require a literate, well trained workforce, a country’s literacy rate can influence both an
MNE’s decision to invest in a particular market and the mode of entry it uses.
Specifically, a literate local labor pool may encourage MNEs to use Majority
Owned Affiliates in a particular country, but it may also produce competitive local firms capable of acting as partners. As a result, the relationship between
LIT and the MNE’s use of external markets is not immediately clear. Values for this variable were obtained from Table A2.2 of Todaro’s EconomicDevelopment in the Third World for the year 1985, and represent (the percent of the Adult population capable of reading and writing). 163 lllc. DOCTOR (1000 people per doctor)
Since health care not only sustains life but also improves the quality of life, a nation’s ability to provide basic medical services to its people represents an important Social indicator. Accordingly, a measure of the number of health care providers in a given country has been included in the current study.
DOCTOR is (the number of people in 1000s per doctor in Country X). Values for this variable appear in Table 28 of the World Development Report for the year
1984. The World Bank notes that, "The figure for physicians, in addition to the total number of registered practitioners in the country, includes medical assistants whose medical training is less than that of qualified physicians but who nevertheless dispense similar medical services...” (1991, pg. 284). Thus,
DOCTOR gives a slightly broader measure of the number of health care providers in a particular country. In any case, higher levels of this variable are an indication that a larger portion of the population must be served by a single health care provider, and will tend to be associated with lower levels of development. Therefore, based on the arguments presented above this variable should be negatively correlated with U.S. MNEs to use external markets.
Hid. %AGR (% of population in agricultural sector)
As a country moves from a traditional Agrarian Society to a modern
Industrial Economy, the portion of its population employed in the agricultural sector drops accordingly.22 Thus, %AGR provides an important indicator of a nation’s position in the development cycle. Specifically, %AGR is (the % of the total Economically Active Population employed in Agriculture, Hunting, Forestry, and Fishing), where "economically active” includes owners, self-employed workers, employees, and unpaid family workers. Unfortunately, it was necessary to use 2 sources to obtain values for this variable, and those values 164 cover a wide range of years. For the vast majority of countries, values for %AGR were taken from the 1989-90 Year Book of Labor Statistics published by the
International Labor Organization (ILO), and cover the period from 1987-89.
However, values for 9 of the study’s 52 countries (Argentina, China, Dominican
Republic, Ecuador, India, New Zealand, Saudi Arabia, South Africa, and the
United Arab Emirates) were not available from the ILO, and had to be obtained from a second source, World Economic Data (Schumacher, 1989). In general these latter figures cover a much earlier time period (from 1980-1986), and are somewhat less reliable than the ILO data. While this may slightly reduce the strength and reliability of the resulting correlation coefficients, the importance of this variable makes it necessary to accept this limitation. Based on the arguments developed thusfar, higher values of %AGR are an indication of lower levels of economic development, and therefore this variable should be negatively correlated with the use of external markets by U.S. MNEs.
IV. Political Variables
The demise of the Centrally Planned Economies of Eastern Europe and the Soviet Union, and the almost universal acceptance of “Free Market” economics, has strengthened the belief that a nation’s political system is a critical determinant of its economic success. According to this view, open, participative Democratic systems promote free markets and strong economic growth; while closed, Communist systems lead to central planning and economic ruin. Since a nation’s political system clearly affects its attitude toward foreign business and investment, the political characteristics of the Host country also play a key role in an MNE’s decision to enter a particular market, while various government restrictions or incentives may affect the form those 165 investments take. Cavisgil's (1985) suggests that the following factors should be considered in an analysis of a country’s Political Environment: System of
Government, Political Stability and Continuity, Ideological Orientation,
Government Involvement in Business and Communications, Attitudes toward
Foreign Business (including trade and investment restrictions), and National
Economic and Development Priorities. Based on these criteria, 3 variables have been selected to provide information on the Political dimension of the external environment. A brief discussion of each of these measures is given below.
IVa. POL (Political system)
In 1987, roughly 1/2 of the countries of the World were Absolute
Monarchies, or under Military or One Party rule.23 On the other hand, the vast majority of the countries involved in this study had Presidential or Parliamentary
Democracies. Thus, it appears that a country’s political orientation plays an important role in a U.S. MNE’s decision to invest in a particular market, and may play a role in the mode of entry it chooses as well. As a result, the variable POL has been developed to describe the type of political system a nation employs.
Unfortunately, this is a difficult variable to operationalize for several reasons. For one thing, the use of a “dummy variable” which takes on 2 values (i.e.
Democratic and Other) is not feasible because most of the countries involved in this study are Constitutional Democracy of one form or another. This means that subtle distinctions between political systems cannot be accurately evaluated, which greatly reduces the strength and validity of the resulting Correlation and
Regression analysis. Further, while dividing political systems into groups like
Military Rule, Absolute Monarchy, One Party Systems, Presidential Systems, and Parliamentary Systems certainly improves the level of detail, this approach 166 is extremely difficult to operationalization and requires the use of 4 separate
dummy variables.
Therefore, POL has been constructed as a “continuous” variable which
assesses several key features of a country’s political system; namely, the
centralization of power, the degree of voter participation and eligibility, the
strength of Democratic institutions, and the number and strength of opposition
parties.24 Based on these factors, a country is assigned a value from 0-10 using the following procedure;
0 - Pure Totalitarianism 1 - Absolute Rule, no participation by general populace 2 - Absolute Rule, with an appointed Council, no general participation 3 - Absolute Rule, with an elected Legislature possessing minimal power 4 - One Party System, Leader appointed by an elected Legislature 5 - Parliamentary System, Legislature possesses minimal power 6 - Parliamentary System, Leader appointed by an elected Legislature 7 - Presidential System, President possesses broad powers 8 - Presidential System, both Leader and Legislature popularly elected 9 - Representative System, with a wide degree of general participation 10 - Pure Democracy
Although the preceding framework is clearly superior to the use of dummy variables, as described above, this scheme fails to account for such important factors as the system’s real locus of decision-making power, the length of time a system has been in place, the system’s stability, and other similar features. Delury (1987) notes that:
“Over much of the world, European government and party systems, both democratic and communistic, are often fragile and nearly irrelevant overlays on patterns of social power that are seldom clearly understood even by the participants. A professedly democratic system with regular and apparently open processes of political debate and resolution of conflict can be and often is an oligarchic system in which members of the elite only superficially represent the interests of the otherwise politically impotent.” (1983, pg. xx) 167 Therefore, it is best to view the preceding framework as a guide for assigning values for this variable rather than a hard and fast rule. Accordingly, common political systems may be assigned a range of values in order to allow the subtle distinctions between systems to be taken into consideration. For example, a country with an Absolute Monarch or Dictator, or under Military Rule could be assigned a value anywhere between 0-4; a One Party System between 2.S-5.5; a Parliamentary System between 3.5-7.5; and a Presidential
System system between 5.5-9.5. Using this less restrictive format, it is quite possible that an open, stable One Party system could receive a higher rating than a repressive, volatile Presidential system, as it undoubtedly should.
Further, it is possible to ignore the presence of a Monarch (or similar absolute entity) when he/she plays a largely symbolic role as in the U.K. and
Commonwealth nations. Although these changes are minor, they should greatly increase the validity of this variable.
Information on the political systems in place in various countries in 1989 comes from 3 sources; Political Handbook of the World (Banks ed., 1990),
Countries of the World and their Leaders (Blair ed. 1991), and World
Encyclopedia of Political Systems and Parties fDelurv ed., 1987). Since higher values of POL are associated with increasing levels of Democracy, it is hypothesized that this variable will be positively correlated with the use of external markets by U.S. MNEs. This implies that MNEs demand greater control in more restrictive political environments. While a number of writers support this view, it is equally likely that MNEs will prefer lower levels of ownership in these countries to avoid the risk of expropriation and similar governmental actions.
Therefore, this hypothesis should be viewed with caution. 168 IVb. REST (Restrictions on foreign business and FDI)
There is little doubt that Host government restrictions on trade and investment have a strong influence on both an MNE’s decision to enter a particular market and the mode of entry it uses. A classic example of this is
IBM’s decision to leave the Indian market in the mid-1970s rather than comply with a law which prohibited foreign investors from owning a majority share of a local business.25 Host governments can employ a wide range of restrictions on foreign investors, including limits on local ownership, controls on foreign exchange and profit repatriation, import and export restrictions, employment restrictions, and special taxes and duties, to mention a few. This makes it difficult to develop a variable capable of effectively measuring local market restrictions. In addition, many of the problems associated with the use of dummy variables are present in this case as well.
Therefore, REST has also been constructed as a “continuous’’ variable with a range from 0-10, which measures several features of a country’s business and investment environment. Specifically, the variable evaluates 4 common government restriction on foreign business: industries closed to foreign investment and limits on local ownership, restrictions on trade, foreign exchange controls and limits on profit repartriation, and other restrictions (such as investment review boards, licensing requirements, and similar barriers to trade and investment). A country is assigned a value between 0-2.5 (i.e. no restrictions to totally restricted) for each of these 4 catagories, and these numbers are then summed to produce the final value of REST. As a result, this variable should provide a fairly comprehensive measure of the obstacles facing an MNE as it trys to conduct activities in a particular market. 169 Information on the government restrictions extant in 1989 comes from 3 sources: Business international’s “Investing, Licensing & Trade Conditions
Abroad”; Price Waterhouse Information Guides “Doing Business in (Country X)”; and the Economist Intelligence Unit’s “Country Profiles”. Since one of the most common forms of government restrictions on foreign business is limiting the permissible amount of foreign ownership, this variable should be positively correlated with U.S. MNE’s use of external markets. Nonetheless, this hypothesis should be viewed with caution, since firms may demand greater control over their local operations in more restrictive local environments.
IVc. PRI (Political risk index)
The study’s final Independent Variable provides a measure of the
“political risk” associated with a given country. In general, political risk is defined as ‘the risk that a particular investment will be affected by unexpected government policies or actions’.26 While this concept is not new, the fall of the
Shah of Iran and similar political disturbances in the 1970s greatly increased the demand for reliable information about the political environment of particular countries. In response to this growing need, a number of private organizations developed “political risk indexes” which ranked countries along a number of key political and economic dimensions. Among the most popular of these services are, Business International’s “International Country Risk Guide”, a similar index developed by the Economist Intelligence Unit, Frost and Sullivan’s “Political
Risk Assessment Service”, and Euromoney’s “Country Risk Guide”. In most cases, a country is rated on a scale from 0-100, although some services use letter ratings (A-D) or qualitative ratings (Good-Bad, High-Low, etc.). In addition, even though each service uses a different method to arrive at its country ratings, most countries tend to receive similar ratings from each service. 170 After reviewing each of these sources, Euromoney’s “Country Risk
Guide” was chosen for use in this study. There were several reasons for this decision. First, Euromoney ranks countries on a scale from 0-100 making it ideal for use in Regression and Correlation analysis. Secondly, it provides a more comprehensive risk measure than some of the other services since it includes economic and credit indicators as well as political indicators.
Euromoney notes that, “...each factor we have used in our risk rating contains varying elements of other factors; for instance, the market is perceptive and takes political, analytical and credit indicators very seriously, while a country’s economic stability may be dependent on the market’s favor” (1989, pg. 207).
Finally, since Euromoney is a popular business magazine, its “Country Risk
Guide” is compiled by a group of economists, political analysts, and market specialists who may be somewhat more sensitive to the needs of international m anagers.
Thus, values for PRI were obtained from Euromoney's 1989 Country Risk ratings which appears in the September, 1989 issue of Euromoney, with one slight caveat. In the Euromoney rating system, 100 represents the lowest level of risk, while 0 is the highest. This approach is counter-intuitive since higher values of PRI would be associated with lower levels of country risk. Therefore,
PRI is actually (100 - the Euromoney rating for Country X). Accordingly, higher values of this variable are an indication of higher levels of country risk.
Unfortunately, there is no consensus in the literature on the direction of the relationship between Risk and MNE entry strategies.27 As Vernon (1990) notes, firms may respond to certain risks through Internalization and FDI in order to increase their control over foreign operations; but these risks can also be handled through Joint Ventures and a reduction in the MNE’s share of its local 171 affiliates. Therefore, the effect of PRI on U.S. MNE’s decision to use external
markets is unclear.
Dependent Variables
I. Introduction
The Dependent Variable used in this study must provide a comprehensive measure of the amount of U.S. MNE activity in a particular
country that is conducted through external (rather than internal) markets.
Clearly, this is an extremely difficult figure to calculate, particularly since MNEs
perform a wide range of activities and it is not altogether clear which of these should form the basis for this variable. Therefore, 4 dimensions of U.S. MNE activity (trade, investment, employment, and sales) will be examined in this study, and a separate variable has been created to measure external activity in each area. The study’s 4 Dependent Variables are:
1) %INTFIRM (% of total trade accounted for by intra-firm trade) 2) %AFSALES (% of total majority owned foreign affiliate (MOFA) sales to affiliated parties) 3) %MOFA (% MNE foreign assets in MOFA) 4) %WORK (% of locally employed workers employed in MOFA)
Before proceeding with a brief discussion of each variables, it is important to stress several points: First, all 4 of these variables will be estimated for the 52 countries involved in this study through information contained in the
Commerce Department’s 1989 Benchmark Survey U.S. Direct Investment
Abroad (USDC, 1991). For all practical purposes, this document provides information on the activities of all non-bank foreign affiliates of non-bank U.S. parents in a particular country. (For the purposes of this study, a “foreign 172 affiliate” is defined as a foreign company which is owned 10% or more by a U.S. parent. A majority owned foreign affiliate (MOFA) is a foreign company which is owned 50% or more by a U.S. parent) Second, in order to gain an accurate picture of U.S. MNE activity in a specific country, ail of these variables must be considered concurrently. It is clearly not enough to just examine the ownership, employment, trade or sales practices of U.S. MNEs separately, since these activities may vary widely between countries due to various industry- and firm- specific factors. Third, each of these variables is expressed as a percentage or ratio in order to minimize the effect of the size of a nation’s economy on the value of a given variable. This will facilitate direct comparisons between countries. Finally, each of these variables has been constructed in such a way that lower values indicate that a greater portion of local U.S. MNE activity is being conducted through external markets.
II. %INTFIRM
This variable attempts to measure the amount of intra-firm trade taking place between U.S. parents and their local majority-owned foreign affiliates in a given country. Both Casson (1986) and Korbin (1991) discuss the difficulties involved in estimating and interpreting figures on intra-firm trade. These include: the problems of data availability and accuracy, distortions due to transfer prices, and the high level of aggregation of many statistics. As a result, the measure used in the current study has been constructed in the most straightforward manner possible. Specifically, %INTFIRM is the sum of [(all U.S. merchandise
Exports shipped by U.S. parents the their MOFAs in country X) plus (all U.S. merchandise Imports shipped by U.S. MOFAs in Country X to their U.S. parents)] divided by the sum of [(all merchandise Imports from Country X to the 173 U.S.) plus (all merchandise Exports from the U.S. to Country X)]. The variable is calculated from 1989 data contained in the Direction of Trade Statistics
Yearbook (IMF. 1990), and Table 64 of the Commerce Department’s 1989
Benchmark Survey.
Since %INTFIRM provides an estimate of the portion of merchandise trade between the U.S. and a given country that is conducted through the internal markets of U.S. MNEs, higher values will occur when local MOFAs are used to assemble or manufacture goods for sale in the U.S., when vertical integration is used to guarantee access to raw materials and other factors, or to protect firm-specific assets. As a result, this variable should be neoativelv associated with country-specific factors that increase the efficiency of external markets.
III. %AFSALES
%AFSALES provides another estimate of the amount of intra-firm trade occurring between U.S. parents and their local affiliates, and is quite similar to a variable used by Korbin (1991) to measure “transnational integration”. Unlike
%INTFIRM, which only measures merchandise trade between local MOFAs and their U.S. parents, %AFSALES measures the portion of local MOFA sales of goods and services that go to all affiliated parties (this includes local and foreign sales to the U.S. parent and any party which is owned 10% or more by the parent). Accordingly, this variable provides a much broader measure of intra-firm trade and can be thought of as indicating the extent to which U.S.
MOFAs in a given country are “integrated” into the international networks of their
U.S. parents. %AFSALES is constructed from Table 40 of the Benchmark
Survey, and is simply [the total sales (of goods and services) of U.S. MOFAs in
County X to affiliated parties] divided by [the total sale (of goods and services) of 174 U.S. MOFAs in Country X]. Since higher values of this variable indicate that the bulk of local MOFA output is going to other affiliates through the internai markets of their U.S. parents, %AFSALES should be neoativelv correlated with environmental factors which increase the efficiency of external markets.
IV. %MOFA
This variable attempts to capture the extent to which U.S. firms are using
“New Forms of Investment” to enter and service a foreign market. However, since reliable data on many New Forms is not available, %MOFA focuses only on the amount of local U.S. investment held in the majority (versus minority) owned affiliates. Specifically, the variable is the ratio of (the total assets of U.S.
MOFAs in Country X) to (the total assets of all foreign affiliates of U.S. firms in that country). It is constructed from information contained in Table 1 and Table
20 of the Commerce Department’s 1989 Benchmark Survey. Once aoain higher values of the variable indicate that more U.S. firms are using internal markets to conduct their local activities; and therefore, country-specific factors which promote efficient external markets should be neoativelv associated with this variable. Note, however, that the Commerce Department’s data does not include foreign affiliates that have less than 10% U.S. ownership; and therefore,
%MOFA slightly overestimates the true portion of U.S. assets held in majority owned affiliates in a particular country.
V. %WORK
%WORK measures the portion of all workers employed by the local affiliates of U.S. firms that work in majority owned affiliates. It is simply (the number of employees in all U.S. MOFAs in Country X) divided by (the number of employees in all affiliates of U.S. firms in Country X). it is also constructed from data contained in Table 1 and Table 20 of the 1989 Benchmark Survey. Since 175 employment and production are inextricably linked, hioher values of %WORK suggest that a greater portion of the local production of U.S. MNEs is generated through majority owned affiliates rather than minority Joint Ventures or other similar arrangements. This situation is consistent with the tighter integration and control of the MOFAs in a particular country within the international networks of their U.S. parents; therefore, this variable should be negatively correlated with country-specific factors that encourage the use of external markets.
Research Hypotheses The hypothesized relationships between the study’s Independent and
Dependent Variables are summarized in Table 1 of the first chapter, and have been discussed in detail above. However, since some confusion may arise about the direction of the relationship between variables due to the way in which the Dependent Variables have been operationalized, the basic Research
Hypotheses are restated in their direct form below.
1) PCGNP should be negatively correlated with all 4 of the study’s DVs. 2) GDPGRO should be positively correlated with all 4 of the study’s DVs. 3) TOTRADE should be negatively correlated with all 4 of the study’s DVs 4) GOVEXP should be positively correlated with all 4 of the study’s DVs. 5) INVEST should be negatively correlated with all 4 of the study’s DVs. 6) PQLI should be neoativelv correlated with all 4 of the study’s DVs. 7) The direction of the relationship between LIT and the study’s 4 DVs is unclear. 8) DOCTOR should be positively correlated with all 4 of the study’s DVs. 9) %AGR should be positively correlated with all 4 of the study’s DVs. 10) POL should be negatively correlated with all 4 of the study’s DVs. 11) REST should be negatively correlated with all 4 of the study’s DVs. 12) The direction of the relationship between PRI and the study’s 4 DVs is unclear. 176 NOTES
1) For a detailed discussion of Correlation see Newbold (1991, pgs. 457-472) and Hays (1988, pgs. 555-564).
2) See Cohen and Cohen (1983, pgs. 30-40) for a full discussion of the various types of correlation coefficients and their appropriate usege.
3) The procedure used to estimate the sample correlation coefficient is explained in detail in Newbold (1991, pgs. 462-3).
4) For a detailed discussion on the procedure used to test the Null Hypothesis that the population correlation coefficient is zero see Newbold (1991, pgs. 464- 465).
5) For a detailed discussion of Regression analysis see Newbold (1991, pgs. 472-556), and Hays (1988, pgs. 544-670)
6) For a detailed discussion of the general assumptions associated with the Linear Regression Model see Newbold (1991, pgs, 481-2).
7) The coefficient of determination (R2) and the procedure used to calculate it are explained in detail in Newbold (1991, pgs. 483-486) and Hays (1988, pgs. 559-60).
8) For a detailed discussion of Multiple Linear Regression see Newbold (1991, pgs. 512-567), and Hays (1988, pgs. 608-669).
9) For more information on squared partial and squared semi-partial correlation [(pr2) and (sr2) respectively] see Hays (1988, pgs. 609-620).
10) The procedure for testing the Null Hypotheses that the population regression coefficients are zero is explained in detail in Newbold (1991, pgs. 491-496 and 533-548).
11) The F-test used to test the Null Hypothesis that the coefficient of multiple correlation (R2) is zero is given in Hays (1988, pgs. 638-653).
12) For a detailed discussion of the problems associated with Regression analysis see Hays (1988, pgs. 653-656).
13) Analysis of Variance (ANOVA) is a statistical technique used to test hypotheses about the relationship between the means of two or more populations. This is done by testing the significance of the resulting F-ratio(s). 177 For a complete discussion of the topic of ANOVA see Keppel (1983).
14) This restriction is necessary because, although the Commerce Department collects data on every country in which U.S. firms are active, it only provides detailed, disaggregated data on 58 countries. The criteria used by the Commerce Department to choose these countries is not immediately apparent; however, a review of these nations suggests that the amount of U.S. MNE investment and/or activity plays a key role in this decision.
15) A general “rule of thumb” is that at least 10-20 subjects are required for each independent Variable included in a regression model, but this not a hard and fast rule.
16) For a detailed discussion of the use of “Sets” of Independent Variables in Regression analysis see Chapter 4 of Cohen and Cohen (1983, pgs. 133-177).
17) For the technical notes accompanying the World Development Report Tables, see (World Bank; 1991; pgs. 270-289).
18) From Table 2 of the World Development Report 1991 the average annual GDP growth rate from 1980-89 was 6.2% for Low-income economies, 2.9% for Middle-income economies, 3.2% for Upper-middle-income economies, and 3.0% for High-income economies.
19) For a review of several recent studies on Industrial Policy see the subsections on 'Cooperation and National Competitiveness’ and 'Inter-Firm Cooperation’ in the Strategic Management section of Chapter II of this document.
20) Rostow (1960), Myrdal (1973), and Todaro (1989) all discuss the problems associated with using per capita GNP as an indicator of a country’s level of Development.
21) For a discussion of various Social indicators see Todaro (1989, pgs. 10S- 113).
22) For example, see Rostow (1960), Kemp (1989), Todaro (1989), and Dixon (1990).
23) See Table 7.1 of the World Development Report 1991 (1991, pg. 129).
24) I would like to thank Edward Roche, PhD for his kind assistance in developing the scale used to evaluate political systems. 178 25) A discussion of IBM’s decision to leave India appears in Encarnation and Vachani (1985).
26) For an introduction to the subject of Political Risk see Simon (1982), Ahkter and Lusch (1987), de la Torre and Neckar (1988), Friedman and Kim (1988), and Korbin (1979).
27) For a review of studies on the relationship between Political risk on FDI see Sethi and Luther (1986), and Fetehi-Sedah and Safizadeh (1989). CHAPTER IV
DATA ANALYSIS AND INTERPRETATION
Overview
This chapter presents the results of the study’s statistical analysis, and is divided into 5 sections including this brief introduction. The second section addresses the study’s first major objective; namely, to determine whether U.S.
MNEs have increased their use of external markets over the past several decades. This is accomplished by comparing 1989 data on the foreign activities of U.S. MNEs with similar data from 1966. Based on this analysis it should be possible to identify any changes that have taken place in the behavior of U.S.
MNEs during this 25 year period. The third section contains the results of the
Correlation analysis of the study’s Independent and Dependent Variables (IVs and DVs). After constructing a Correlation Matrix, the strength and direction of the relationships between the study’s 12 IVs and 4 DVs will be described and assessed. The fourth section presents the results of the Regression analysis of the study’s 4 Dependent Variables. A separate regression equation will be developed for each DV in an attempt to explain variations in the behavior of
U.S. MNEs in different countries in terms of a small group of country-specific factors. Taken together, these 2 sections address the study’s second major objective; namely, relating U.S. MNEs’ use of external markets to a specific set of environmental factors. The chapter’s final section contains a brief summary of the study's most important findings.
179 180 Comparison of the External Activities of U.S. MNEs in 1966 and 1989
Table 3 assembles information on the distribution of assets and employees in the majority-owned foreign affiliates (MOFAs) of U.S. MNEs. The table was constructed from data contained in the Commerce Department's 1966 and 1989 Benchmark Surveys of U.S. Direct Investment Abroad (USDC, 1974,
1991), and will be used to identify any major changes that have occurred in the way U.S. MNEs structure their foreign operations. Based on the arguments developed in the preceding chapters, it has been hypothesized that U.S. firms conducted a greater portion of their foreign activities though external (rather than internal) markets in 1989 than they did in 1966. Accordingly, Table 3 compares values of the study’s 4 Dependent Variables at these 2 points in time, and will be used to identify relative changes in U.S. MNEs' use of minoritv- versus maioritv-owned affiliates to enter and service foreign markets.
For all practical purposes, the Commerce Department’s Benchmark
Surveys provide information on the operations of all foreign affiliates of U.S. parent corporations in all industries and countries where activity is present. As a result, the figures in Table 3 represent universe totals of the various items being measured. In general, the information contained in the 1966 and 1989 surveys is identical, with 3 important exceptions: 1) the ownership requirements for the
1966 survey are slightly different than those for the 1989 survey, 2) total employment figures were not given in the 1966 survey, and 3) the 1966 survey includes information on the foreign activities of U.S. banks while the 1989 survey does not. Although these differences are clearly minor, taken together they may make the direct comparison of some of the figures in Table 3 somewhat misleading. A brief discussion of each of these differences is given below. 181 TABLE 3
Foreign Affiliates of U.S. MNEs (1966,1989)
1966 1989
Total No. of U.S. Parents 3,354 2,167 Total No. For. Affiliates 23,120 17,835 No. of MOFA 20,544 15,390 MOFAs a s % of Total 88.9% 86.9% Total For. Assets ($M) $128,939 $1,313,995 Total MOFA Assets ($M) $113,633 $1,067,793 %MOFA 88.1% 81.2% Total No. For. Employees (1000) 4,387 (a) . 6,621 No. Employees in MOFAs (1000) 3,874 5,111 %WORK 88.3% (a) 77.2% %AFSALES 19.1% 23.8% %INTFIRM 18.8% 18.4%
(a) These figures are estimates based on the amount of compensation paid to workers in MOFAs.
Source: U.S. Department of Commerce, Benchmark Survey of U.S. Direct Investment Abroad (1966, 1989) 182 As noted in the previous chapter, foreign affiliates meeting certain minimum sales or asset requirements and owned 10% or more by a U.S. parent were required to participate in the 1989 survey. While the same basic requirements were present in 1966, different financial and operating data was requested from U.S. parents based on their level of ownership in a given foreign affiliate. For example, as the Bureau of Economic Analysis explains in the introduction to the 1966 survey:
“There are three categories of foreign affiliates, based on the ownership interest of the U.S. reporter in the affiliate. “Associated” foreign affiliates are those in which a single U.S. reporter holds at least a 10 but less than a 25 percent ownership interest; however, the combined ownership interest of all U.S. reporters in the affiliate must be less than 25 percent. If the combined U.S. reporter ownership interest is 25 percent or more, the affiliate is classified as “allied”. “Majority-owned" foreign affiliates (a component of allied foreign affiliates) are those in which a single U.S. reporter’s ownership interest is at least 50 percent" (USCD, 1974, pg. 3)
These distinctions are important because in the 1966 survey, detailed financial and operating information was only requested from “allied” foreign affiliates. Thus, when comparing figures on the total assets and total number of local employees in all foreign affiliates of U.S. firms, it is necessary to remember that the 1966 figures do not reflect assets or employees in foreign affiliates with
10-25% U.S. ownership. Further, at least in the case of total assets, there is no way to estimate how significant these omissions may be; although, since only
11% of all foreign affiliates had 10-50% U.S. ownership in 1966, it seems reasonable to suggest that these differences are not severe. On the other hand, it is possible to estimate the total employee figure for 1966 from data on total compensation which was collected for all 3 classes of affiliates. In any case, when comparing the figures for Total Foreign Assets, Total Employees, and 183 %WORK, these problems must be taken into consideration.
The final problem involves the industries covered in the 2 survey. As
noted above, the 1966 survey includes information on the foreign activities of
U.S. banks which is not included in the 1989 survey. The likely reason for this difference is that the tremendous growth in global currency and capital markets over the past 2-3 decades has made the dollar value of certain banking activities dis-proportionately large in relation to other industries. Thus, while both surveys include Finance as a major industry heading, in the 1966 survey this sector included Banking, Holding Companies, Other Finance, and
Insurance; while in the 1989 survey, it included Finance (except Banking),
Insurance, Real Estate, and Holding Companies. Since the entire Finance sector represented only 7.8% of all U.S. foreign affiliate assets in 1966, and
12.5% of all foreign affiliate assets in 1989, it seems reasonable to conclude that the omission of U.S. banking activities in the more recent survey will not materially affect the comparison of the figures contained in Table 3.
The remainder of this section is devoted to an analysis of the changes that have occurred in the foreign activities of U.S. firms during the period from
1966-1989. Within this context, Table 3 contains a number of interesting, and unexpected findings. Perhaps the most surprising of these is the total lack of any increase in either the number of U.S. minority-owned foreign affiliates or the percentage of assets held in these arrangements. In fact, in 1966 U.S. firms actually had more minority affiliates than they did in 1989 (2576 in 1966, versus
2445 in 1989). Further, there has been a decrease in both the number of U.S. firms engaging in foreign ownership and the total number of foreign affiliates of
U.S. parents. In 1966, 3354 U.S. firms had a total of 23,120 foreign affiliates,
11.1% of which were minority owned. In 1989, only 2167 U.S. firms had foreign 184 affiliates, and a surprisingly similar 13.7% of these were minority owned. Thus, despite the avalanche of recent literature on the strategic importance of Joint
Ventures, U.S. companies seem to be making little more use of these arrangements than they did 25 years ago. This is a startling finding, and one which is not immediately explainable.
On the other hand, the percentage of total assets and total employees in majority-owned foreign affiliates (MOFAs) have both decreased slightly over the period. In 1966, roughly 88% of all assets and employees were in MOFA; but by
1989, these figures had dropped to 81% of all assets, and 77% of all employees in MOFAs. While these figures appear to indicate a moderate increase in the use of minority-owned foreign affiliates, it must be remembered that the 1966 figures do not include affiliates with 10-25% U.S. ownership. Consequently, the
1966 figures overstate the importance of MOFAs by an indeterminate amount.
This makes the comparable 1966 and 1989 figures much closer than they first appear. Thus, if any increase in the portion of U.S. foreign assets held in minority-owned affiliates has occured between 1966 and 1989, that increase has not been substantial.
But while the distribution of assets between minority- and majority- owned affiliates has remained rather constant between 1966 and 1989, the character of the “average” U.S. MOFA has changed dramatically during this period. For example, in 1966, the average MOFA of a U.S. parent had total assets of $20 million (in 1989 dollars) and employed 189 workers. In 1989, average assets had increased 4-fold to $85.4 million, and employment had almost doubled to 332. Taken together, these figures suggest that U.S. firms have “consolidated” their foreign operations during the past few decades. This view is supported by the fact that, in 1989, there were fewer foreign affiliates of 185 U.S. firms, but those operating were larger in terms of both total assets and employees.
There are several possible explanations for this phenomenon. The first is the consolidation that has taken place in many U.S. firms and industries over the past 2 decades. Numerous sources have discussed the growing
“concentration” of many U.S. industries and the increased importance of a few, large companies within the U.S. market.1 In fact, by one estimate, the largest
100 American firms now account for nearly 50% of all U.S. corporate assets.2
Given this trend, It is quite possible that a similar consolidation has occurred in the foreign activities of U.S. firms as well. In fact, such a consolidation has been encouraged by a second, more important, phenomenon; namely, the rise in global competition and the advent of homogeneous global markets.
As a result of this latter occurrence, the creation of a global, rationalized production network, and the need to achieve economies of scale and scope in production have become critical elements in a firm’s continued survival. Both of these factors clearly favor the creation of fewer, larger foreign affiliates capable of servicing a number of key markets or producing for the firm’s global network.
But as foreign affiliates begin to play a larger role in an MNE’s international production and distribution system, the need to control these affiliates increases accordingly. As a result, one would expect to see an increase in the amount of activity conducted through the MOFAs of U.S. MNEs.
This latter conclusion is confirmed by the remaining figures in Table 3.
Both %AFSALES and %INTFIRM attempt to measure the amount of activity that occurs through the “internal" markets of U.S. MNEs. Specifically, %INTFIRM measures the portion of U.S. merchandise trade (the sum of imports and 186 exports) accounted for by trade between U.S. parents and their MOFAs, while
%AFSALES measures the portion of the total sales of goods and services of
U.S. MOFAs going to other affiliated parties. Taken together, these figures suggest that while trade between U.S. parents and their MOFAs has remained rather consistent between 1966 and 1989 (accounting for roughly 18-19% of total U.S. trade in both years), a much greater portion of the output of U.S.
MOFAs now goes to other affiliated parties. In fact, In 1989, nearly a quarter of the total output of U.S. MOFAs was sold through the “internal” markets of their
U.S. parents, up from 19% in 1966. This is consistent with the tighter integration of MOFAs within the international networks of their U.S. parents, and once again contradicts the hypothesis that U.S. firms have increased their use of
“external” markets during this period.
Correlation Analysis
I. Introduction
This section presents the results of the Correlation analysis of the study's
Independent and Dependent Variables (IVs and DVs). Once again, the results are somewhat surprising, since in general, they fail to support the hypothesized relationships between variables. In fact, in 7 out of 10 cases, the correlation between a given IV and the DVs had a sign opposite that predicted. Table 4 assembles values for the 12 IVs and 4 DVs for each of the 52 countries examined in this study and represents the study’s “Data Table”. Table 5 summarizes several important descriptive statistics for each variable, including the mean, median, standard deviation, minimum and maximum values, and 1st and 3d quartile values. 187 IABLE-4
Data Table
PCGNP GDPG TOTR GOVEX INVE PQLI LIT DOCT
Argentina 2.16 -0.3 25.9 15.5 12 90 96 0.37 Australia 14.36 3.5 25.9 27.0 26 100 99 0.44 Austria 17.30 1.9 56.4 39.3 27 96 98 0.39 Belgium 16.22 1.8 127.6 50.7 20 97 99 0.33 Brazil 2.54 3.0 16.5 30.6 22 77 78 1.08 Canada 19.03 3.3(a) 46.5 23.1 23 99 99 0.51 Chile 1.77 2.7 58.2 32.5 20 91 97 1.23 China 0.35 9.7 26.7 21.8(c) 36 80 69 1.01 Columbia 1.20 3.5 27.3 14.6 20 82 88 1.24 Costa Rica 1.78 2.8 59.5 27.8 24 94 94 0.96 Denmark 20.45 2.2 61.2 41.8 19 98 99 0.40 Dorn. Rep. 0.79 2.4 47.4 20.4 26 75 77 1.76 Ecuador 1.02 1.9 40.6 14.2 20 79 82 0.82 Egypt 0.64 5.4 31.7 40.2 24 60 45 0.77 Finland 22.12 3.0 47.5 29.3 30 99 99 0.44 France 17.82 2.1 38.0 42.6 21 100 99 0.32 Germany 20.44 1.9 51.2 29.0 22 97 99 0.38 Greece 5.35 1.6 58.8 38.3 (c) 18 97 98 0.35 Guatam ala 0.91 0.4 8.9 12.0 14 64 55 2.18 Honduras 0.90 2.3 48.6 20.7 (c) 13 67 59 1.51 Hong Kong 10.35 7.1 192.0 23.3(b) 27 95 88 1.07 India 0.34 5.3 14.8 17.7 24 55 44 2.52 Indonesia 0.50 5.3 40.6 20.6 35 63 74 9.46 Ireland 8.71 1.8 128.9 57.9 21 96 99 0.68 Israel 9.79 3.2 51.8 49.1 16 96 95 0.35 Italy 15.12 2.4 33.5 47.9 24 98 97 0.23 1 TABLE 4 (cont.1
PCGNP GDPG TOTR GOVEX INVE PQLI LIT DOCT
Jamaica 1.26 1.2 71.9 41.4(c) 29 92 92 2.05 Japan 23.81 4.0 17.1 16.5 33 99 99 0.66 Korea 4.40 9.7 58.3 16.9 35 88 92 1.16 Luxembourg 20.14 3.2 182.0 37.1 (c) 20 97 100 0.33 Malaysia 2.16 4.9 126.9 30.1 30 81 73 1.93 Mexico 2.01 0.7 22.4 21.2 17 84 90 1.24 Netherlands 15.92 1.7 95.6 54.5 19 99 99 0.45 New Zealand 12.07 2.2 41.9 45.9 32 96 99 0.58 Nigeria 0.25 -0.4 43.6 28.1 13 47 43 6.44 Norway 22.29 3.6 53.3 42.7 27 99 99 0.45 Panama 1.76 0.5 27.7 31.7 3 90 88 1.00 Peru 1.01 0.4 19.4 11.6 20 71 85 1.04 Philippines 0.71 0.7 41.7 15.7 19 79 86 6.57 Portugal 4.25 2.5 70.9 43.3 30 91 85 0.41 Saudi Arabia 6.02 -1.8 59.0 13.0 (b) 21 56 24 0.74 Singapore 10.45 6.1 332.2 23.3 35 91 86 1.31 South Africa 2.47 1.5 38.0 33.0 21 66 76 1.05 (b) Spain 9.33 3.1 30.5 34.3 25 98 95 0.32 Swedan 21.57 1.8 60.3 40.6 22 99 99 0.39 Switzerland 29.88 2.1 62.6 8.7 (c) 36 99 99 0.70 Thailand 1.22 7.0 65.8 15.1 31 82 91 6.29 Trin &Tob 3.23 -5.5 66.7 36.9 19 90 96 0.95 Turkey 1.37 5.1 38.3 23.7 22 73 74 1.39 UAE 18.43 -4.5 87.0 13.0 25 74 48 1.02 UK 14.61 2.6 48.8 34.6 21 97 99 0.45 (b) Venezuela 2.45 1.0 47.4 21.8(a) 13 87 87 0.70 189 TABLE 4 fcont.)
%AGR POL RESTPRI%INT%AFS %MO %WO
Argentina 12.0 (d) 7.00 3.75 72 14.1 24.3 82.3 76.2 Australia 5.4 6.50 2.50 18 29.0 8.4 74.0 49.4 Austria 7.9 7.50 1.50 12 10.3 16.3 49.8 60.9 Belgium 2.5 6.50 0.50 12 22.2 39.4 90.1 86.6 Brazil 23.3 7.00 5.00 68 20.9 12.3 73.8 78.1 C anada 3.4 6.50 1.50 10 38.4 21.2 92.8 94.1 Chile 19.0 2.25 3.00 59 6.7 8.1 77.1 89.1 China 68.4(d) 1.75 7.00 33 0.2 11.9 37.6 34.9 Columbia 1.1 7.75 5.25 63 7.4 5.6 95.0 92.0 Costa Rica 27.4 8.00 3.25 67 6.2 56.0 92.5 93.7 Denmark 5.7 6.50 0.75 17 10.4 14.7 94.2 88.4 Dorn. Rep. 45.0 (d) 7.75 4.00 65 2.7 41.3 88.2 (f) 76.5 Ecuador 34.8 (d) 7.00 5.50 69 7.4 37.4 78.6 75.3 Egypt 35.7 5.00 5.75 62 1.9 49.3 93.1 82.7 Finland 8.7 7.50 4.00 12 4.0 4.6 98.2 96.0 France 6.7 8.50 3.00 10 19.0 20.6 79.5 82.7 Germany 4.8 6.50 0.75 7 15.5 28.6 81.9 89.0 Greece 24.6 6.75 3.50 33 5.0 7.4 95.9 97.0 Guatam ala 49.8 7.00 2.50 68 5.4 23.9 92.1 93.2 Honduras 47.7 7.25 2.75 68 15.1 32.1 98.4 99.5 Hong Kong 1.1 4.50 0.75 30 28.5 30.5 93.7 75.5 India 62.6 6.25 6.00 33 0.9 (e) 5.9 21.3 26.9 Indonesia 54.4 (d) 5.00 5.50 38 12.6 33.7 95.2 92.6 Ireland 12.7 7.25 2.00 25 41.1 60.0 99.7 95.0 Israel 4.4 6.75 3.00 44 3.7 32.1 55.0 49.0 Italy 8.1 6.75 2.75 13 10.3 12.8 85.8 86.2 190 TABLE.,.4 (CQflU
%AGR POL REST PRI %INT %AFS %MO %WO
Jam aica 24.2 6.50 4.00 64 17.5 42.2 90.0 78.3 Japan 7.4 6.25 2.75 5 5.7 18.3 41.0 33.8 Korea 19.0 5.00 5.00 21 3.6 24.4 22.4 39.1 Luxembourg 3.2 6.00 0.50 12 22.2 52.0 89.5 96.3 Malaysia 30.9 5.50 3.25 32 26.9 38.2 89.6 95.5 Mexico 22.9 7.25 14.25 56 23.6 34.4 52.2 63.7 Netherlands 4.3 6.00 1.00 11 25.8 35.6 82.3 88.9 New Zealand 10.4 (d)6.50 2.50 17 11.6 9.3 94.5(f) 93.7 Nigeria 44.6 2.50 3.25 64 19.4 84.0 90.2 87.0 Norway 6.4 6.00 1.50 13 9.9 (e) 24.6 87.0 67.5 Panam a 26.1 7.00 0.75 66 29.5 38.1 98.0 97.1 Peru 35.2 7.00 5.75 75 2.6 6.1 97.0 95.0 Philippines 41.3 6.75 4.75 62 8.7 20.7 75.7 62.8 Portugal 20.6 6.75 3.75 23 6.2 22.1 83.3 88.8 Saudi Arabia 20.0(d) 1.00 4.00 25 1.5(e) 9.7 35.6 73.8 Singapore 0.4 5.50 1.00 16 42.8 51.0 94.1 85.6 South Africa 15.5(d) 4.50 3.50 54 5.7 7.9 70.8 64.9 Spain 12.3 6.25 3.50 41 28.6 22.4 80.0 78.9 Swedan 3.5 6.25 1.50 11 14.2(e) 16.0 84.2 70.6 Switzerland 5.6 9.25 0.50 5 17.4 25.3 96.9 84.2 Thailand 64.4 5.50 4.00 28 17.2 30.3 83.5 61.3 Trin& Tob 11.9 6.00 2.75 48 2.0 (e) 60.7 59.2 80.8 Turkey 45.3 7.00 3.00 38 1.2 15.6 51.6 57.7 UAE 4.6 (d) 2.50 4.00 39 17.0 62.0 98.0 75.0 UK 2.1 6.00 1.00 9 27.6 18.9 93.4 85.8 Venezuela 12.5 7.75 4.00 53 3.3 3.9 56.8 69.7
(a) These figures are for 1988. (b) These figures are estimates based on similar country data. (c) These figues were taken from International Financial Statistics Yearbook 1990(IMF. 1991). (d) These figures were taken from World Economic Data (Schumacher. 1989). (e) These figures are based on Exports only. (f) These figures are estimates based on the number of MOFAs. TABLE 5
Descriptive Statistics for the Study’s IVs and DVs
N MEAN MEDIAN STDEV
PCGNP 52 8.56 4.32 8.42 GDPGRO 52 2.569 2.350 2.726 TOTRADE 52 61.06 48.05 53.52 GOVEXP 52 29.28 28.55 12.53 INVEST 52 23.12 22.00 6.902 PQLI 52 85.96 91.00 14.06 LIT 52 85.19 92.00 18.29 DOCTOR 52 1.354 0.795 1.818 %AGR 52 20.61 12.60 18.64 POL 52 6.144 6.500 1.656 REST 52 3.106 3.125 1.662 PRI 52 36.46 33.00 22.92 %INTFIRM 52 14.05 11.00 11.04 %AFSALES 52 7.16 24.10 17.92 %MOFA 52 79.09 86.40 20.57 %WORK 52 77.62 82.70 18.04
MIN MAX Q1 Q3
PCGNP 0.25 29.88 1.23 16.15 GDPGRO -5.50 9.700 1.525 3.50 TOTRADE 8.90 332.20 32.15 62.25 GOVEXP 8.70 57.90 18.38 39.97 INVEST 3.00 36.00 19.25 27.00 PQLI 47.00 100.00 77.50 97.00 LIT 24.00 100.00 77.25 99.00 DOCTOR 0.230 9.46 0.417 1.24 %AGR 0.40 68.40 5.45 33.83 POL 1.00 9.25 5.625 7.00 REST 0.50 7.00 1.500 4.00 PRI 5.00 75.00 13.00 62.00 %INTFIRM 0.20 42.80 5.10 21.87 %AFSALES 3.90 84.00 12.43 37.92 %MOFA 21.30 99.70 73.85 94.00 %WORK 26.90 99.50 68.05 92.45 192 Although these figures are not particularly interesting in and of themselves, they do illustrate one significant point; namely, the wide "variability” in each of the variables used in this study. This is not as trivial as it may first appear, since the application of Regression and Correlation is contingent upon precisely such variability. Finally, Table 6 presents the Correlation Matrix associated with the study’s 12 IVs and 4 DVs, while Table 7 summarizes the relationships between the Independent and Dependent variables. As noted above, relatively few strong relationships were identified between the IVs and the DVs; however, a number of strong correlations were observed between the
IVs themselves.
II. “Average” Country Profile
The information in Table 5 can be used to construct a profile of the
“typical” country hosting a substantial amount of U.S. FDI. Although this analysis is not directly related to the study’s primary objectives, it is nonetheless interesting to note the type of environmental characteristics that are associated with U.S. investment. Due to the great variety among the countries examined in this study, it must be said that no single profile accurately describes all of the country-specific factors which attract U.S. MNEs. However, based on an analysis of the mean values of the study’s 12 IVs, several observations can be made.
Perhaps the most significant of these is that U.S. MNEs appear to invest heavily in countries in the later stages of economic development. For example, the study’s “average" country had a 1989 per capita GNP of $8560, making it a
High-income economy under the World Bank’s classification system.3 Further, its GDP grew at an annual rate of 2.57% during the 1980s; while government expenditures accounted for almost 30% of GNP in 1989, and domestic 193 investment accounted for 23% of GDP. Finally, trade was an important component of this country’s economy, with the sum of merchandise imports plus exports equal to 61% of GNP. Thus, based on these economic indicators, U.S.
MNEs tend to invest in growing Developed economies; however, this image is only partially supported by the study's Social indicators.
In 1985, the “average” country had the relatively high Physical Quality of
Life rating of 86, and a Literacy rate of 85%. Roughly 21% of the population was employed in the Agricultural sector, while 1 doctor served approximately 1354 people. Unlike the Economic variables, which characterize a fully developed nation, these figures tend to describe a country in the later stages of
Development. This contention is reinforced by the study’s 3 Political variables which suggest that the “average” country has a Parliamentary form of government, and places moderate restrictions on trade and FDI. It also had the surprisingly high Political Risk rating of 36. Based on this profile, U.S. MNEs appear to invest heavily in countries in the later stages of economic and social development which have only moderate Government involvement in the economy, relative high rates of domestic investment, high Physical Qualities of
Life and Literacy rates, and a relative open Political system. On the other hand,
U.S. MNEs may be much less influenced by Host government restrictions and relative high levels of “political risk” than is commonly believed.
III. Correlations Between the Study’s Independent Variables
Table 6 presents the Correlation Matrix associated with the study’s
Independent and Dependent Variables. A cursory review of this matrix reveals a number of important correlations between the 12 IVs, but relatively few relationships between the Independent and Dependent Variables. Thus, before discussing these latter relationships, it may be useful to examine the correlation 194 TABLE 6
Correlation Matrix
PCGNP GDPGR TOTRD GOVEX INVEST PQLI LIT DOCTR
GDPGRO -.062 TOTRADE .202 .173 GOVEXP .279 -.060 .178 INVEST .259 .556* .248 -.099 PQLI .652* .070 .239 .471 + .15 LIT .478 + .096 .132 .413 + .105 .909 ** DOCTOR -.425 + .116 -.073 -.329 + .113 -.532* -.324 + %AGR -.703** .256 -.335 + -.401 + .028 -.705** -.566* .617* POL .183 -.058 -.139 .113 -.135 .418 + .530* -.222 REST -.658* .166 -.457 + -.410 + .152 -.582* -.472 + .337 + PRI -.811** -.249 -.316 + -.344 + -.485 + -.566* -.395 + .269 %INTFIRM .278 .040 .528* .226 -.021 .353 + .285 -.038 %AFSALES -.090 -.235 .413 + .137 -.098 -.149 -.177 .240 %MOFA .164 -.213 .258 .200 -.133 .185 .208 .060 %WORK .050 -.359 + .233 .243 -.282 .108 .124 -.039
%AGR POL REST PRI %INTF %AFS %MOFA
POL -.190 REST .625* -.206 PRI .491 + -.005 .536* %INTFIRM -.374 + .078 -.542* -.275 %AFSALES .046 -.144 -.184 .172 .351 + %MOFA -.200 .241 -.358 + .085 .431 + %WORK -.215 .196 -.345 + .111 .327 +
+ Indicates a moderate correlation between variables. * Indicates a substantial correlation between variables. ** Indicates a very strong correlation between variables. 195 between the IVs themselves.4 In some cases, these relationships can be
attributed to the use of “Sets” of Independent Variables, since the variables
within a given set are often highly correlated with each other. However, in other
cases these relationships are both interesting and instructive, even though they
should not be considered a valid indication of the correlation between various
Economic, Social, and Political indicators in the general population of nations.
Finally, as noted in Chapter 3, the Davis convention will be used to describe the
strength of the correlation between 2 variables. This method ascribes the
following adjectives to a relationship based on the value of its correlation
coefficient (r): “negligible", “low”, “moderate", “substantial”, and “very strong”.
Beginning with the 5 Economic Variables, PCGNP has a very strong
negative correlation with both PRI and %AGR, a substantial negative
association with REST, and a substantial positive association with PQLI. Among
other things, these relationships support the notion that higher levels of
economic development are associated with a smaller percentage of the
population employed in Agriculture, a higher Physical Quality of Life, and lower
levels of Political Risk. Although these statements are largely tautological, the
substantial negative correlation between per capita GNP and restrictions on
foreign business and FDI is of much more interest. Specifically, it would be
useful to know whether minimal restrictions on trade and FDI foster higher
levels of per capita GNP (the Neoclassical argument), or whether restrictions
are removed once a country reaches a certain level of per capita GNP.
Unfortunately, this question is well beyond the scope of the current study, but the issue is clearly important given the debate that has arisen in many
Developed Countries over the merits of Free Trade. 196 The only important relationship between GDPGRO and any of the study’s other IVs is its substantial positive correlation with INVEST. Thus, at least in the nations comprising this study, the relationship between domestic investment and domestic growth is supported. Similarly, TOTRADE has few noteworthy correlations with other IVs, the only exception is its moderate negative correlation with REST. This relationship is certainly intuitively appealing since restrictions on trade is one of the components of REST, and one would expect lower levels of trade to occur in countries with greater restrictions on trade.
On the other hand, GOVEXP is moderately correlated with several of the study’s IVs. This includes moderate positive correlations with both PQLI and LIT, and moderate negative correlations with both %AGR and REST. Once again it would be useful to establish the causal relationships between these variables, since they appear to support the currently unfashionable argument that higher levels of Government spending promote higher levels of economic and social development. Finally, there is a substantial positive association between
INVEST and GDPGRO and a moderate negative association between INVEST and PRI.
Turning to the 4 Social Variables, PQLI has numerous important relationships with other IVs. These include a very strong positive correlation with LIT (which should come as little surprise since the literacy rate is one of the
3 components of the PQLI), and a very strong negative correlation with %AGR; as well as a substantial positive correlation with PCGNP, and substantial negative correlations with DOCTOR, REST, and PRI. All of these relationships clearly support the use of PQLI as an alternative indicator of a country’s level of
Development, since higher Physical Qualities of Life are associated with higher levels of literacy, fewer people per doctor, a lower portion of the population 197 employed in Agriculture, and higher levels of per capita GNP. In addition, LIT has a very strong positive association with PQLI; and a substantial positive association with POL and a substantial negative correlation with %AGR. Here again, It would be useful to know whether higher literacy rates are the cause or result of more open political systems and lower levels of Agricultural employment.
DOCTOR has a substantial positive correlation with %AGR, and a substantial negative correlation with PQLI. In analyzing these relationships it is important to remember that DOCTOR represents the number of people per doctor (not doctors per person); and therefore, higher levels of DOCTOR indicate fewer health care providers in the general population. Accordingly, higher levels of this variable are an indication of lower levels of Development, and should be associated with higher levels of %AGR and lower levels of PQLI.
Finally, %AGR has very strong negative correlations with PCGNP and PQLI; substantial positive correlations with DOCTOR, REST and PRI; and a substantial negative correlation with LIT. Once again these relationships are consistent with a country’s level of economic and social development, suggesting that nations with a high portion of the population employed in the
Agricultural sector also tend to be in earlier stages of the Development process.
The relationships between the 3 Political Variables and other IVs are equally dramatic, with both REST and PRI manifesting numerous strong correlations. On the other hand, POL has only a substantial positive association with LIT and a moderate positive association with PQLI. The absence of any important relationships between a country’s political system and the study’s economic indicators is particularly interesting given the current fascination with the concept that “Democracy equals Prosperity”. There are several possible 198 explanation for the lack of support of this equation, not the least of which is that
POL does not provide an accurate indicator of a country’s political orientation.
On the other hand, it is equally likely that the ideological fervor generated by the collapse of the former Eastern Bloc is largely misguided, and that political systems have far less impact on economic performance than is currently purported. In fact, one does not have to look very far to find support for this contention, since during the 25 year period between 1965-1989, China experienced an average annual rate of per capita GNP growth almost twice that of the more liberal OECD nations, and an average annual rate of GDP growth almost 3 time that of the “Democracies” of Western Europe and the United
States during the 1980s.5
As previously noted, there are many important relationships between
REST and other IVs, including substantial negative relationships between
REST and PCGNP and PQLI, and substantial positive associations between
REST and %AGR and PRI. Once again, it is unfortunate that Correlation analysis cannot be used to establish the causal relationship between these variables. Finally, PRI has a very strong negative correlation with PCGNP, a substantial negative correlation with PQLI, and substantial positive correlations with %AGR and REST. Considering the nature of the political risk index used in this study, it is not surprising that countries with higher levels of economic and social development tend to have lower levels of political risk, while those with greater restrictions on foreign business activities tend to have higher levels of political risk.
Finally, it is interesting to compare the degree of multicollinearity within each “Set” of Independent Variables to the correlation between the 3 sets. In general, the Social Variables were highly correlated with each other, while the 199 Economic and Political variables manifested relatively few strong relationships
within each set. On the other hand, the relationships between Social and
Political variables were far more pronounced than those between either
Economic and Political variables and Economic, or Social variables.
Considered as a whole, the above analysis tends to support the argument that
higher levels of economic development are associated with higher levels of
social development, but uncovers little evidence that economic development is associated with a particular political system. On the other hand, there is strong
evidence that higher levels of social development are associated with fewer
restrictions on trade and FDI and lower levels of political risk. Once again, while these conclusions are evident within the group of nations hosting significant
levels of U.S. FDI, they may or may not reflect the more general relationships present in the entire population of nations.
IV. Correlations Between the Independent and Dependent Variables
While the preceding discussion is certainly of interest, it is only ancillary to the major objective of the current study; namely, to describe the relationships between the IVs and the DVs. Unfortunately, there are far fewer of these relationships than between the IVs themselves In fact, based on a cursory review of the Correlation Matrix, it must be noted that none of the Economic,
Social, and Political variables used in this study seems to have a great impact on the external activities of U.S. MNEs. This in and of itself is startling, but perhaps even more surprising is the almost total lack of support for the hypothesized relationships between the Independent and Dependent
Variables. Table 7 summarizes the empirical relationships between the study’s
IVs and DVs; and shows that in 7 out of 10 cases, the empirical relations were in the opposite direction of that expected. Further, only 2 of the 12 IVs were even ■TABLE 7 Hypothesized (HYP) and Empirical Directions of the Relationships between IVs and DVs
HYP %INTFIRM %AFSALES %MOFA %WORK
PCGNP (-) (+) (-) (+) (+) GDPGRO (+) (+) (-) (-) (•) TOTRADE (-) (+) “ (+)* (+) (+) GOVEXP (+) (+) (+) (+) (+) INVEST (-) (-) (-) (-) (-) PQLI (-) (+)* (-) (+) (+) LIT (?) (+) (-) (+) (+) DOCTOR (+) (-) (+) (+) (-) %AGR (+) (-)* (+) (■) (•) POL (-) (+) (-) (+) (+) REST (-) (-) ** (-) (-)* (-) PRI (?) (•) (+) (+) (+)
This indicates that the strength of the relationship is moderate. ** This indicates that the strength of the relationship is substantial. 201 moderately correlated with any of the DVs. Thus, the remainder of this section is
devoted to a detailed discussion of these findings, beginning with a review of
the observed relationships between each IV and the study’s 4 DVs.
One of this study’s basic hypotheses is that higher levels of external MNE
activity will occur in countries with higher levels of economic and social development. Accordingly, variables which indicate higher levels of
Development should be negatively correlated with all 4 of the study’s DVs. In
most cases this was not true. For example, PCGNP, one of the most widely used
Development indicators, was positively correlated with 3 of 4 of the DVs
(although these relationships were weak). This suggests that U.S. MNEs tend to conduct a greater portion of their activities through internal markets in countries with higher levels of per capita GNP.
This notion is confirmed by the negative correlation between GDPGRO and 3 of the study’s 4 DVs. Since higher rates of GDP growth have been associated with Developing Countries in the last decade, it was hypothesized that U.S. MNEs would use internal markets in countries that had experienced high growth rates. This was not the case, although once again the strength of the correlations were low. Further, U.S. MNEs tended to conduct a greater portion of their activities through internal markets in “high trade” nations, as evidenced by the positive correlation between TOTRADE and all 4 of the study’s
DVs (in the case of 2 of the DVs this correlation was moderate or substantial).
This is particularly surprising since it was hypothesized that high levels of trade relative to GDP were an indication of more “open” local markets, and should be associated with the use of external markets.
One of the few hypothesized relationships supported by the study was the low positive correlation between GOVEXP and all 4 of the DVs. It was 202 suggested that higher levels of government expenditures relative to GDP would
produce market inefficiencies, which should encourage U.S. MNEs to conduct a
greater portion of their activities in these countries through internal markets.
This appears to be true. Finally, the negative correlation hypothesized between
INVEST and all 4 of the study’s DVs was also supported. The argument here is
that higher levels of domestic investment relative to GDP should improve the
competitiveness of local firms, and thus encourage U.S. MNEs to make greater
use of external markets. Once again, however, the strength of this correlation
was rather low.
Turning to the Social Variables, the negative correlation hypothesized
between PQLI and the DVs was only partially supported. Two of the study’s 4
DVs had a low negative correlation with PQLI, while the other 2 had a low or
moderate positive correlation. Thus, the argument that in countries with higher
levels of (social) Development, U.S. MNEs will make greater use of external
markets cannot be accepted without reservation. On the other hand, the direction of the relationship between LIT and 3 of the study’s 4 DVs was
positive, indicating that U.S. MNEs tend to make greater use of internal markets
in countries with high literacy rates. Although no hypothesis was advanced for this variable, an argument could be made that higher levels of local literacy are associated with higher skill levels of the local workerforce, which could attract
U.S. MNEs.
The positive correlation between DOCTOR and the DVs was only partially supported, with 2 of the DVs being positively correlated and 2 negatively correlated. Finally, the positive relationship between %AGR and the study’s DVs was not supported. Instead, 3 of the 4 DVs had a low negative correlation with this variable. Once again, it appears that U.S. MNEs make 203 greater use of internal markets in countries with higher levels of Development,
contrary to the arguments developed in the preceding chapters.
Finally, the relationships between the 3 political variables and the study’s
DVs were also contrary to those hypothesized. The relationship between POL
and 3 of the 4 DVs was positive rather than negative, suggesting that U.S.
MNEs tend to make greater use of internal markets in countries with more open
political systems. But the study’s most important relationships were between
REST and the DVs. In all cases, REST was negatively correlated with the DVs,
and in 3 cases these correlations were moderate or substantial. As predicted,
restrictions on foreign business activities tended to encourage U.S. MNEs to
conduct a greater portion of their activities through external markets. Finally, the
relationship between PRI and the 3 of the 4 DVs was positive, indicating that in
countries with higher levels of risks, U.S. MNEs tend to make greater use of
internal markets.
The picture that emerges from this analysis strongly contradicts many
currently popular theories on MNE behavior. For example, rather than taking
advantage of local opportunities to use external markets, U.S. MNEs tend to
prefer internal markets in most situations. Further, the argument that MNEs will
make greater use of “New Forms of Investment” in more developed countries
found no support. Instead, U.S. MNEs seem to conduct a greater portion of their
activities through internal markets in these countries. In fact, the only variable that was consistently associated with the use of external markets was REST,
suggesting that U.S. MNEs organize the bulk of their local activities through
internal markets unless Host governments force them to do otherwise. This is a startling and rather pessimistic conclusion, but nonetheless clearly evident from the study’s findings. 204 One possible explanation of these findings is that the importance of
“global” production and distribution networks has increased in recent years, forcing MNEs to seek greater control over their local operations whenever possible. This argument has become well established in the literature, and in many ways, it represents the “conventional wisdom” of the 1960s and 1970s.
Unfortunately, this approach fails to take into consideration the potential benefits associated with more cooperative relationships between the MNE and various “stakeholder” groups.
The first 2 chapters of this dissertation examine the theoretical basis for these cooperative strategies in detail. Based on this analysis, it has been suggested that in certain economic environments (particularly Developed nations which tend to have more efficient local markets and firms) it may be advantageous for MNEs to coordinate their activities through external rather than internal markets. The fact that this argument found little support in the current study indicates either that the argument itself is flawed, or that U.S. managers have failed to recognize the benefits of this approach. These and other conclusions will be explored further in the final chapter of this work.
Since one of the primary objectives of this study is to develop regression models for each of the study's 4 Dependent Variables, it may be useful to explore the relationships between IVs and DVs from another perspective. In particular, the correlation between each DV and the study’s 12 IVs can be examined in order to identify potential predictor variables for inclusion in each of the 4 regression models. Although the results of this analysis are identical to those presented above, they have been included below as an introduction to the Regression Analysis of the study’s Dependent Variables. 205 A number of important relationships can be identified between
%INTFIRM and the study’s IVs. These include a substantial positive correlation with TOTRADE and a substantial negative correlation with REST; as well as a moderate positive correlation with PQLI, and a moderate negative correlation with %AGR. Taken together, these relationships suggest that a greater amount of trade takes place between U.S. parents and their local MOFAs in countries which have relatively fewer restrictions on trade and FDI, or higher levels of trade relative to GDP. While this should come as no surprise, the correlations between %INTFIRM and PQLI and %AGR are much more baffling since they appear to indicate that U.S. MNEs conduct a greater portion of their local activities through internal markets in countries with higher levels of (social) development. If this in fact true, it contradicts one of the study’s fundamental arguments; namely, that countries with higher levels of development tend to have more efficient external markets which should encourage MNEs to conduct a greater portion of their activities through the marketplace.
The only relationship of note between %AFSALES and any of the study’s
IVs is its moderate positive correlation with TOTRADE. This implies that U.S.
MOFAs in countries with high levels of trade relative to GDP tend to sell a greater portion of their output to other affiliated parties. One way of interpreting this finding is that MOFAs in “high-trade” nations play a more important role in the international networks of their U.S. parents than MOFAs in other countries.
On the other hand, the lack of any other strong relationships is somewhat surprising, since it suggests that neither the level of development of the Host country, its political orientation, or local restrictions on foreign business have much impact on the way U.S. MNEs structure their local operations. 206 These conclusions are supported by the absence of any strong relationships between %MOFA and the study's IVs. In fact, the only correlation present is the moderate negative association between %MOFA and REST. This is not at all surprising since one of the components of REST is restrictions on local foreign ownership; however, it is surprising that none of the study’s economic and social indicators had more than a low association with this variable. Once again, it appears that the level of development of the Host nation and its political system have little impact on U.S. MNEs decision to use majority- or minority-owned affiliates to service a particular market.
The study’s final DV, %WORK, has moderate negative associations with both GDPGRO and REST. Thus, in countries with high levels of GDP growth and those which restrict foreign business activities, U.S. MNEs tend to employ relatively more workers in minoritv-owned affiliates than in other countries. One possible explanation for this relationship is that during the 1980s, high growth countries also tended to be Developing countries. As a result, a large portion of the U.S. MOFAs operating in these nations may be involved in labor intensive, extractive industries which are often regulated or controlled by the Host government. Consequently, U.S. firms in these industries may be required to form Joint Ventures with local partners, which would increase the share of locally employed workers working in the minority affiliates of U.S. firms.
Unfortunately, there is no way to confirm or reject this suggestion from the data available in this study, but it is intuitively appealing nonetheless.
Finally, it is interesting to look at the correlation between the DVs themselves. There is a very strong positive correlation between %MOFA and
%WORK, and moderate positive correlations between %INTFIRM and both 207 %MOFA and %WORK. In general then, the study’s 4 Dependent Variables do
appear to measure slightly different aspects of external MNE activity. Thus, they
should provide a richer analysis of U.S. MNE behavior than could be obtained from a single Dependent Variable. In addition, it is interesting to note that for all
but 2 of the 12 IVs (GOVEXP and INVEST), the relationships between a given IV
and 3 or all 4 of the DVs had the same sign. This is another indication that the
DVs are providing consistent and hence reliable indicators of unique aspects of
external MNE behavior.
As for explaining the correlations between the DVs, it is quite reasonable that %MOFA and %WORK are highly correlated. Since higher values of %MOFA indicate that more of the local assets of U.S. MNEs are held in majority-owned affiliates, one would expect a greater potion of the locally employed workers of
U.S. firms to work in majority (versus minority) affiliates as well. On the other hand, the relationships between %INTFIRM and %MOFA and %WORK are a bit more complex. Remember that %INTFIRM measures the portion of U.S. merchandise trade with a particular country that is conducted between U.S. parents and their local MOFAs. One would naturally expect this figure to be higher in countries where more local U.S. assets are held in MOFAs (i.e. countries with higher values of %MOFA). Finally, given the very strong correlation between %MOFA and %WORK, any variable that is associated with
%MOFA is also likely to be associated with %WORK.
Begression Analysis I. Introduction
Chapter III contains an elaborate scheme for developing a regression model for each of the study’s 4 Dependent Variables. The plan is based on 208 identifying the IVs in each of the 3 “Sets” with the highest correlations to a given
DV and constructing a model from 3-4 of these. Further, it was stated that only
IVs with a correlation of .40 or greater would be used in a particular regression model. Unfortunately, based on the Correlation analysis presented in the preceding sections, this scheme now appears untenable. In general, very few strong relationships were identified between the study’s Independent and
Dependent Variables, and in only 3 cases was a correlation of .40 or greater present. The consistently low correlation betweeen variables has 2 important consequences: first, it makes it difficult to develop regression models which explain a large portion of the variance in the DVs; and second, it means that some other method must be used to develop these models. Fortunately, there is a statistical tool which can be used in these circumstances; namely, “Step-wise”
Regression.6
Step-wise regression provides a procedure for selecting the IVs to be included in a particular regression model, and is commonly employed when the researcher lacks a strong theoretical basis for making these decisions. The procedure is based on an algorithm which specifies that the variable capable of explaining the greatest amount of the remaining variance in the DV will be added to the model at each step of the process. As a result, step-wise regression can be used to develop extremely powerful predictive models from the smallest possible number of IVs. But while the primary objective of this procedure and the one detailed in Chapter III is identical (i.e. to develop a model which explains the variability in the DV), there are several important differences between these 2 approaches. Perhaps the most obvious is that the variable selected at each stage of the step-wise procedure is not always the one with the highest correlation (r) to the DV. In fact, only the first IV to enter the 209 model will always be the most highly correlated; subsequent IVs are selected by a slightly different criteria.
Specifically, additional variables are evaluated on their semi-partial correlations (sr) not their simple correlations (r). Thus, the variable which explains the most remaining variance in the DV (i.e. yields the highest incremental increase in R2) will be added to the model at each step. This distinction is particularly important when multicollinearity is present between the
IVs, as in the current study. Under these conditions, IVs which are highly correlated with the DV may add little explanatory power to a regression model due to their high correlation with other IVs. On the other hand, a variable that is only moderately correlated with the DV may greatly increase the power of a model when it uniquely accounts for variance that is not explained by the other variables. Thus, it is important to note that while the step-wise procedure can be used to develop effective predictive models, the theoretical significance of the variables included in a given model should be viewed with caution.
The remainder of this section is devoted to the Regression analysis of each of the study’s 4 DVs. Step-wise regression will be used to develop a model that explains the greatest amount of the variability in each DV using 3-4
IVs. Once the appropriate predictor variables have been identified, the corresponding population regression equation will be calculated, as well as the squared partial and squared semi-partial correlations (pr2 and sr2 respectively) for each predictor variable. Once again, it is important to note that the regression equations presented below are population equations, and therefore, not subject to many of the problems associated with the use of a sample.
Accordingly, the T and “p” values calculated by the computer program should 210 not be used to test the significance of the corresponding regression coefficients, since any non-zero coefficients are significant in this case. On the other hand, these statistics may give an indication of the importance of a particular variable in the regression model, and have been included for this purpose only.
Finally, a value for ‘R-sq (adj)" is also provided for each equation. This figure corrects the Coefficient of Multiple Correlation (R2) for the number of IVs included in the regression equation. Given the study’s rather low ratio of subjects to IVs, the value for R-sq (adj) may provide a somewhat more accurate indication of the amount of total variance in the DV explained by the model than the R2 value.
II. %INTFIRM
%INTFIRM attempts to measure the portion of merchandise trade between the U.S. and a particular country that consists of trade between U.S. parent corporations and their local majority-owned foreign affiliates (MOFAs).
Tables 8 & 9 present the results of the regression analysis of this variable. Table
8 displays the results of the step-wise procedure, and Table 9 provides the population regression equation and various statistical information on each variable employed in the model. For each the study’s 4 DVs, step-wise regression has been used to identify the 4 IVs which explain the greatest amount of variance in the DV. In the case of %INTFIRM, these variables are:
REST, TOTRADE, DOCTOR, and PQLI. The regression model resulting from these 4 predictors explains roughly 42% of the variability in %INTFIRM, which is quite good considering the rather low correlations extent between the study’s
Independent and Dependent variables. However, in constructing the final regression model, PQLI was dropped due to its low explanatory power. 211 TABLE 8
Step-wise Regression of %INTFIRM on all 12 IVs
STEP 1 2 3 4 CONSTANT 25.236 17.414 17.548 3.820 REST -3.60 -2.52 -2.86 -2.34 t-ratio -4.56 -3.03 -3.24 -2.37 TOTRADE 0.073 0.070 0.070 t-ratio 2.84 2.72 2.72 DOCTOR 0.80 1.20 t-ratio 1.12 1.50 PQLI 0.13 t-ratio 1.14 R-SQ 29.36 39.32 40.86 42.44
BEST ALT. TOTRADE DOCTOR LIT %AGR t-ratio 4.40 1.31 0.60 -1.01 2ND ALT. %AGRGDPGRO PQLI LIT t-ratio -2.85 1.11 0.53 0.83
IABLE.9 Regression of %iNTFIRM on REST, TOTRADE, DOCTOR
The Population Regression Equation is: %INTFIRM = 17.5 - 2.86 REST + 0.0704 TOTRADE + 0.804 DOCTOR
Predictor Coef Stdev t-ratio P sr2 pr2 Constant 17.548 3.782 4.64 0.000 REST -2.861 0.8831 -3.24 0.002 .130 .180 TOTRADE 0.0704 0.0259 2.72 0.009 .092 .135 DOCTOR 0.8039 0.7196 1.12 0.269 .026 .043
R-sq = 40.9% R-sq(adj) = 37.2% 212 Thus, the final model for %INTFIRM contains 3 variables, REST,
TOTRADE, and DOCTOR, and explains about 41% of the variance in this DV
(37% when adjusted for the number of variables used in the model). The
population regression equation associated with the model is:
(4) %INTFIRM = 17.5- 2.86 REST + .0704 TOTRADE + .804 DOCTOR.
As noted above, the first variable selected by the step-wise procedure is always the one most highly correlated with the DV; in the case of %INTFIRM that variable was REST. In fact, REST alone explains almost 30% of the variance in
%INTFIRM, although the amount of unique variance accounted for in the final
model is somewhat lower than this. The second variable to enter the model
(TOTRADE) also accounts for a rather large portion of the variance in
%INTFIRM. Taken together, REST and TOTRADE explain roughly 40% of total variance, while the final variable, DOCTOR, adds only about 2% of unique explanatory power to the model.
Based on the results of this regression analysis, several observations can be made. The first is that the step-wise procedure selected one variable from each of the 3 “Sets" of IV. This is interesting because it lends a moderate degree of empirical support to the arguments that led to the segmentation of environmental factors into distinct sets, as well as to the specific divisions employed in this study. It is also interesting that, at least in this case, the 2 variables most highly correlated with %INTFIRM (REST and TOTRADE) were also the first 2 variables included in the model. As noted above, this is not always the case. In fact, the 3d variable selected by the step-wise procedure was DOCTOR, not PQLI, even though PQLI is much more highly correlated with
%INTFIRM the DOCTOR. 213 Finally, although this analysis must be viewed with caution, it is worthwhile asking whether this model sheds any light on the relationship between environmental factors and %INTFIRM. This is an important question because %INTFIRM provides one measure of the amount of activity conducted through the internal markets of U.S. corporations. Based on the preceding analysis, 2 external factors seem to explain a large portion of the differences between countries in the amount of trade that occurs between U.S. parents and their local MOFAs. These factors are: Host government restrictions on trade and
FDI (REST), and the relative importance of trade to the Host nation’s economy
(TOTRADE). Not surprisingly, Host government restrictions have a negative effect on trade between U.S. parents and their local subsidiaries, while the relative importance of trade to the Host country has a positive impact. The implications of these findings will be explored further in the final chapter of this dissertation.
III. %AFSALES
%AFSALES measures the portion of the total sales of all U.S. MOFAs in a given country that are sold to affiliated parties, and provides a broad measure of the amount of “intra-firm” trade that originates in a particular nation. Tables 10
& 11 present the results of the step-wise regression of this variable, which identified TOTRADE, PRI, GDPGRO, and DOCTOR as the 4 most powerful predictors. The model associated with these variables explains roughly 40% of the variance in %AFSALES (35% when adjusted for the number of IVs in the model). Further, each of the 4 variables contributes some additional explanatory power to the model. The corresponding population regression equation is:
(5) %AFSALES = 12.1 + .183 TOTRADE + .154 PRI - 2.05 GDPGRO+ 2.59 DOCTOR. 214 TABLE 10 Step-wise Regression of %AFSALES on all 12 IVs
STEP 1 2 3 4 CONSTANT 18.718 6.963 12.399 12.092 TOTRADE 0.138 0.174 0.183 0.183 t-ratio 3.20 4.04 4.38 4.56 PRI 0.263 0.219 0.154 t-ratio 2.62 2.21 1.54 GDPGRO -1.71 -2.05 t-ratio -2.13 -2.59 DOCTOR 2.6 t-ratio 2.19 R-SQ 17.04 27.21 33.47 39.64
BEST ALT. DOCTOR GDPGRO DOCTOR %AGR t-ratio 1.75 -2.55 1.62 1.85 2ND ALT. GDPGRODOCTOR GOVEXPGOVEXP t-ratio -1.71 2.18 1.42 1.11
TABLE 11. Regression of %AFSALES on TOTRADE, PRI, GDPGRO, DOCTOR
The Population Regression Equation is: %AFSALES » 12.1 + 0.183 TOTRADE + 0.154 PRI - 2.05 GDPGRO + 2.59 DOCTOR
Predictor Coef Stdev t-ratio P sr2 Constant 12.092 5.739 2.11 0.040 TOTRADE 0.1834 0.0402 4.56 0.000 .267 .307 PRI 0.1539 0.0999 1.54 0.130 .030 .047 GDPGRO -2.045 0.7884 -2.59 0.013 .086 .125 DOCTOR 2.590 1.182 2.19 0.034 .061 .092
R-sq = 39.6% R-sq(adj) = 34.5% 215 TOTRADE is the most powerful IV in the model, single-handedly explaining 17% of the variance in %AFSALES. It is also the variable most highly correlated with this DV. On the other hand, the 3 remaining predictors explain considerably less variance, with PRI uniquely accounting for only 3% of the variance in %AFSALES in the final model. This is interesting, since PRI was the second variable selected by the step-wise procedure, and it clearly demonstrates how the explanatory power of an IV can change dramatically as additional variables enter a regression model. This is especially true when there is a great deal of multicollinearity between IVs, as in the present study.
Further, it is important to note that while the t- and p- values associated with PRI would ordinarily indicate a regression coefficient that is not significantly different from zero, in this case they should be seen as an indication of the minimal explanatory power of this variable instead.
Once again the step-wise procedure included at least one variable from each of the 3 sets of IVs, although the Political Variable (PRI) is not particularly powerful. It is interesting to compare this model to the one developed for
%INTFIRM, since both %AFSALES and %INTFIRM provide estimates of the extent to which U.S. corporation’s use internal markets to conduct their activities in a particular country. In both cases the importance of trade to the Host country’s economy (TOTRADE) plays an key role in the model, with “high trade” nations encouraging a greater use of internal markets. It is also interesting that in countries with high levels of GDP growth, local U.S. MOFAs sell relatively less of their output to other affiliated parties. This may suggest that local affiliates in high growth countries are being used to service local markets rather than to produce for the parent’s international network. 216 IV. %MOFA
%MOFA measures the portion of total U.S. MNE assets in a particular
country that are held in local maiority-owned affiliates. Tables 12 & 13 present
the results of the regression analysis of this variable. The step-wise procedure
identified REST, PRI, PCGNP, and DOCTOR as the 4 best predictors of this DV,
even though REST is the only IV even moderately correlated with %MOFA. This
is a perfect illustration of a situation where variables which are individually only
weakly correlated with a particular DV can be used to create a reasonably
powerful predictive model. In fact, together these 4 variables account for roughly
33% of the total variance in %MOFA (27% when adjusted for the number of
variables in the model). The population regression equation associated with these variables is:
(6) %MOFA = 59.5 - 5.82 REST + .625 PRI +1.29 PCGNP + 2.89 DOCTOR.
This model has several interesting features, not the least of which is its considerably lower R-sq than either of the preceding models. Once again, the step-wise procedure included at least one predictor from each set of IVs, although the 2 Political Variables (REST and PRI) are clearly the most powerful, together explaining almost 25% of the variance in %MOFA. Accordingly, this model suggests that the “political” climate of the Host country is a critical determinant of U.S. MNEs’ decision to use a majority- or minority-owned affiliate to service a particular market. Although this argument has been widely acknowledged in the literature and has been examined in detail in other parts of this dissertation, researchers have generally considered economic and political constraints to be of equal importantance in the MNE’s decision-making process.7 On the other hand, based on these results, the political characteristics of the Host country seem to explain far more of the variance in MNE entry 217 TABLE..12 Step-wise Regression of %MOFA on all 12 IVs
STEP 1 2 3 4 CONSTANT 92.85 88.14 67.11 59.47 REST -4.4 -7.0 -5.5 -5.8 t-ratio -2.71 -3.83 -2.73 -2.95 PRI 0.35 0.57 0.62 t-ratio 2.63 3.05 3.37 PCGNP 0.95 1.29 t-ratio 1.66 2.20 DOCTOR 2.9 t-ratio 1.91 R-SQ 12.81 23.58 27.72 32.91
BEST ALT. TOTRADE: DOCTOR INVEST TOTRADE t-ratio 1.89 1.47 1.58 1.77 2ND ALT. POL POL DOCTOR INVEST t-ratio 1.76 1.31 1.26 1.64
IABLE-L3 Regression of %MOFA on REST, PRI, PCGNP, DOCTOR
The Population Regression Equation is: %MOFA = 59.5 - 5.82 REST + 0.625 PRI +1.29 PCGNP + 2.89 DOCTOR
Predictor Coef Stdev t-ratio p sr2 pr2 Constant 59.47 14.09 4.22 0.000 REST -5.824 1.972 -2.95 0.005 .124 .156 PRI 0.6246 0.1851 3.37 0.001 .162 .194 PCGNP 1.2874 0.5862 2.20 0.033 .129 .174 DOCTOR 2.890 1.515 1.91 0.063 .052 .072
R-sq = 32.9% R-sq(adj) = 27.2% 218 strategies (specifically, the choice between using a majority- versus minority- owned affiliate in a particular country) than the economic characteristics of the
Host nation. If this in fact true, it is an extremely interesting finding, and one that has not been extensively examined in the literature.
V. %WORK
The study’s final Dependent Variable measures the percent of all local workers employed by U.S. firms, that work in MOFAs. Based on the correlation analysis presented above, this variable is very strongly associated with
%MOFA; and therefore, one might expect the regression models for these 2
DVs to be quite similar. Surprisingly, the results of the step-wise procedure presented in Tables 14 & 15 do not contain a single variable in common.
Instead, GDPGRO, TOTRADE, POL, and GOVEXP were identified as the 4 best predictors of %WORK. However, as with %INTFIRM, the final variable selected by this procedure (GOVEXP) addkl very little additional power and was not included in final regression model. As a result, GDPGRO, TOTRADE, and POL have been used to develop a model which explains roughly 27% of the variance in %WORK (22% when adjusted for the number of variables). The corresponding population regression equation is:
(7) %WORK » 63.0 - 2.67 GDPGRO + .112 TOTRADE + 2.39 POL.
It is interesting that this is the only model which does not include an IV from each of the 3 sets, although 2 of the 3 sets are represented. As with
%MOFA, the explanatory power of this model is considerably lower than the models developed for %INTFIRM and %AFSALES. In fact, only about 20% of the total variance in %WORK is explained when the R2 is adjusted for the number of variables used. Further, this model is perhaps the most difficult of the 219 TABLE 14 Step-wise Regression of %WORK on all 12 IVs
STEP 1 2 3 4 CONSTANT 83.72 78.36 62.95 58.66 GDPGRO -2.37 -2.72 -2.67 -2.59 t-ratio -2.72 -3.21 -3.21 -3.11 TOTRADE 0.102 0.112 0.102 t-ratio 2.37 2.63 2.34 POL 2.4 2.2 t-ratio 1.75 1.58 GOVEXP 0.21 t-ratio 1.12 R-SQ 12.87 21.82 26.53 28.44
BEST ALT. RESTREST GOVEXPINVEST t-ratio -2.60 -2.28 1.33 -1.11 2ND ALT. INVEST GOVEXP REST PRI t-ratio -2.08 1.71 -1.28 1.02
TABLE 15 Regression of %WQRK on GDPGRO, TOTRADE, POL
The Population Regression Equation is: %WORK * 63.0 - 2.67 GDPGRO + 0.112 TOTRADE + 2.39 POL
Predictor Coef Stdev t-ratio P sr2 pr2 Constant 62.952 9.562 6.58 0.000 GDPGRO -2.671 0.8317 -3.21 0.002 .158 .177 TOTRADE 0.1123 0.0427 2.63 0.011 .105 .125 POL 2.389 1.362 1.75 0.086 .047 .060
R-sq = 26.5% R-sq(adj) = 21.9% 220 4 to justify conceptually, in large part because none of the predictor variables are in and of themselves strongly correlated with this DV.
Among other things, the strong negative impact of GDPGRO on %WORK apparent in this model is of particular interest. This implies that in countries with high rates of GDP growth, U.S. firms employ relatively more workers in minority affiliates. There is no immediately apparent rationale for this, even when one considers that high growth countries have been Developing Countries in recent years. In addition, even though REST is more highly correlated with %WORK than either TOTRADE or POL, it was not selected by the step-wise procedure for inclusion in the final model. Clearly, from a theoretical perspective it is easier to explain the impact of Host government restrictions on the local employment practices of U.S. MNEs than it is to explain the effects of trade or political systems.
On the other hand, the positive influence of TOTRADE on %WORK may stem from its role as a powerful predictor of “intra-firm” trade flows (as measured by %INTFIRM and %AFSALES). According to the regression model developed for %WORK, local U.S. MOFAs in “high trade" countries tend to employ a larger percentage of the total number of local workers working for U.S. companies.
This may be an indication that U.S. firms are using their local subsidiaries in these nations to produce goods for sale to other affiliated parties through the
MNE’s international production network. In the final analysis, however, this model may underscore the tentative nature of trying to explain the theoretical significance of variables selected by the step-wise regression process more than anything else. 221 VI. Conclusions
Several conclusions can be drawn from the preceding analysis. Perhaps the most obvious is that environmental factors seem to explain a great deal more of the variance in the “trade” practices of U.S. MNEs than in their ownership practices. This derives from the fact that the models developed for
%INTFIRM and %AFSALES explain roughly 40% of the variance in these DV, while the models for %MOFA and %WORK explain only 25-30% of the variance in these variables. One possible implication of this finding is that the choice between using a minority- or majority-owned affiliate to enter a particular country is far more complex than deciding what role a local affiliate should play in its parent’s international production network. Although the topic has not been explored in detail in this study, a great deal of work has been done on the factors which influence a firm’s decision to enter a particular country and the specific entry strategy it employs.8 Surprisingly, in many cases it appears that corporate decision-makers use non-objective, judgmental criteria to make these decisions.9 On the other hand, the decision to integrate a particular local affiliate into the parent’s global network is much more likely to be affected by critical environmental factors which can be evaluated in a rational, objective manner.
The second group of conclusions stems from the individual variables employed in the regression models. Although this analysis is subject to certain limitations arising from the nature of step-wise regression, several interesting observations can be made nonetheless. The 2 most important of these are that
TOTRADE has a strong influence on both %INTFIRM and %AFSALES, while
REST and PRI have a strong effect on %MOFA. In the first case, the implication 222 is that in countries where trade plays a relatively important role in the local
economy, the local MOFAs of U.S. firms tend to be more tightly integrated into their parents' international networks. In other words, MOFAs in “high trade” countries tend to sell more of their output through the internal markets of their
U.S. parents than MOFAs in other countries. This is an interesting finding and one that is not immediately explainable.
In the second case, it appears that the political characteristics of the Host country (as measured by REST and PRI) are a much better predictor of the local ownership preferences of U.S. firms (i.e. the choice between using minority- or majority-owned affiliates in a particular country) than the economic characteristics of the Host country. While it is not at all surprising that local restrictions on trade and FDI (REST) and the “political risk” of the Host nation
(PRI) have a strong influence on the entry strategies of U.S. MNEs, it is quite surprising that the single Economic variable included in the model for %MOFA was not more powerful. As noted above, most researchers have stressed the importance of boiii economic and political factors in explaining MNE entry strategies; however, based on this analysis, political factors seem to play a far more important role.
Finally, it is interesting that in 3 of the 4 models, the step-wise procedure selected at least one variable from each of the 3 sets of IVs. This is important not only because it lends support to the notion that the external environment can be divided into Economic, Social, and Political factors; but even more importantly, it suggests that all 3 of these classes are necessary to fuliy understand the behavior of U.S. MNEs. Far too often, researchers have focused on only one type of environmental characteristic (typically either Economic or Political), and have generally ignored the influence of the Host country’s Social environment 223 on local MNE activity. Based on the current study, however, this approach seems ill advised since it fails to include a number of potentially critical determinants of U.S. MNE behavior. If nothing else, this once again underscores the importance of adopting an “inter-disciplinary” approach to the study of International Business.
Summary of the Study's Maior Findings
This section contains a brief summary of the study’s major findings.
Readers wishing a more complete discussion of a particular finding are referred to the appropriate section in this chapter.
- Based on a comparison of data on the activities of U.S. MNEs in 1966 and 1989, there is no evidence that an increase in either the total number of minority-owned foreign affiliates or the percentage of assets held in these arrangements has occurred. Further, during this period, there has been a decrease in both the number of U.S. parents engaging in foreign ownership, and the total number of foreign affiliates of U.S. firms.
- On the other hand, there has been a slight increase in the portion of the total output of U.S. majority-owned foreign affiliates (MOFAs) going to affiliated parties through the internal markets of their parents during the 1966-1989 period.
- U.S. MNEs appear to invest heavily in countries in the later stages of economic and social development. These countries are characterized by low-moderate levels of Government involvement in the economy, relatively high rates of domestic investment, high Physical Qualities of Life and Literacy rates, and relative open political systems.
- Further, there is some evidence that Host government restrictions on trade and FDI, and relatively high levels of political risk are not as much of a deterrent to local U.S. MNE investment as is commonly assumed.
- A number of strong relationships were present between the study's 12 IVs. Specifically, there was some evidence that higher levels of economic development are associated with higher levels of social development, while there was strong evidence that higher levels of social development are associated with fewer restrictions on trade and FDI and lower levels 224 of political risk.
- Within the population of countries examined in this study (i.e. those which host a considerable amount of U.S. investment) there is little evidence that a strong relationship exists between a country’s political system and major economic indicators.
- In general, very few strong relationships were identified between the study's 12 IVs and its 4 DVs. In fact, only 2 of the IVs (REST and TOTRADE) were even moderately correlated with the DVs. Further, in 7 out of 10 cases, the empirical relationships between a given IV and the DVs were in the opposite direction of that hypothesized.
- Based on the results of the correlation analysis of the IVs and DVs, there was no support for the hypothesis that U.S. MNEs make greater use of minority-owned affiliates in countries with higher levels of development. In fact, U.S. MNEs appear to conduct a greater portion of their activities through internal markets in these countries.
- Further, based on the consistent positive correlation between REST and all 4 of the study's DVs, it appears that U.S. MNEs tend to organize their foreign activities through internal markets unless Host government restrictions force them to do otherwise.
- Based on the correlation analysis of %INTFIRM, it appears that a greater amount of trade takes place between U.S. parents and their local MOFAs in countries which have relatively fewer restrictions on trade and FDI or higher levels of trade relative to GDP.
- Based on the correlation analysis of %AFSALES, it appears that U.S. MOFAs in countries with high levels of trade relative to GDP tend to sell a greater portion of their output to other affiliated parties through the internal markets of their U.S. parents.
- Based on the correlation analysis of %MOFA, it appears that neither the level of development of the Host country nor its political system have much impact on U.S. MNE’s decision to use majority- or minority-owned affiliates in a particular country.
- Based on the correlation analysis of %WORK, it appears that U.S. MNEs tend to employ relatively more workers inminority-owned affiliates in countries with high levels of GDP growth and those that restrict foreign business activites. 225 - The step-wise regression model developed for %INTFIRM contained 3 predictors (REST, TOTRADE, and DOCTOR) and explains about 41% of the variance in this variable.
- The regression model developed for %AFSALES contained 4 predictors (TOTRADE, PRI, GDPGRO, and DOCTOR) and explains roughly 40% of the variance in this variable.
- The regression model developed for %MOFA contained 4 predictors (REST, PRI, PCGNP, and DOCTOR) and explains about 33% of the variance in this variable.
- The model regression for developed %WORK contained 3 predictors (GDPGRO, TOTRADE, and POL) and explains about 29% of the variance in this variable.
- Based on the regression analysis of the study’s DVs, it appears that environmental factors explain a great deal more of the variance in the "trade” practices of U.S. MNEs, than in their ownership practices.
- Based on the regression analysis of %INTFIRM and %AFSALES, MOFAs in "high trade” countries tend to sell more of their output through the internal markets of their U.S. parents than MOFAs in other countries.
- Based on the regression analysis of %MOFA, the political characteristics of the Host country (as measured by REST and PRI) appear to be a much better predictor of the ownership preferences of U.S. firms (i.e. the choice between using minority- or majority-owned affiliates in a particular country) than the economic characteristics of the Host country.
- Finally, in 3 out of 4 cases, the step-wise procedure included at least one Economic, Social, and Political variable in the final regression model. This not only lends support to the notion that the external environment can be divided along these lines; but even more importantly, it suggests that all 3 types of factors are necessary to fully understand the behavior of U.S. MNEs. 226 NOTES
1) An interesting examination of the recent consolidation of many U.S. industries appears in The Age of Consolidation” (Business Week, Oct., 14, 1991). Based on this review, the top 5 companies in the Appliance, Airline, Software and Tire industries controlled between 60-97% of the U.S. market in 1990, and had significantly increased their total market share since 1985.
2) According to EiU's 1990-91 Country Profile for the United States, the 100 largest U.S. firms accounted for about 48% of total U.S. corporate assets in 1985.
3) The World Bank has developed a system for classifying nations based on per capita GNP. Low-income economies are those with a per capita GNP of $580 or less, Middle-income economies have per capita GNPs between $581-5999, while High-income economies have per capita GNPs above $6000. For more details on this scheme, see World Development Report 1991. pg. 199.
4) The problems associated with correlation between the Independent Variables of a regression model (“multicollinearity") have been discussed briefly in Chapter 3. For a more thorough examination of this topic, see Hays (1988), pgs. 654-6.
5) Based on data contained in Tables 1 & 2 of the 1991 World Development Report. China had an average annual rate of per capita GNP growth of 5.7% between 1965-89, compared to only 2.5% for the OECD countries. Further, China’s average annual rate of GDP growth for the 1980-89 period was 9.7%, compared to an average annual rate of 3.0% for the OECD nations.
6) For more information on Step-wise regression, see Hays (1988) pgs. 662-8.
7) For a more complete discussion of the economic and political imperatives, see Doz (1980), Doz and Prahalad (1984), Prahalad and Doz (1987), and Bartlett and Ghoshal (1989) among others.
8) A number of models have been developed for assessing the investment potential of a particular foreign market. Typical examples include Cavusgil (1985), Douglas and Craig (1990), and Weiss (1990). These models generally attempt to evaluate a particular foreign market based on various Social, Economic, and Political factors. On the other hand, Anderson and Gatignon (1986) note that the choice between entry modes often focuses on the issues of control, risk, and return.
9) A recent article by Korbin (1979) examines the affect of Political Risk on the firm’s decision to invest in a particular market. Korbin notes that while most 227 surveys of managers indicate that Political Risk is a major determinant of the FDI decision, few firms have created formal systems for evaluating these risks, and most continue to rely on less rigorous, judgemental assessment processes. Further, Boddewyn (1983) discusses a number of ‘behavioral” factors which influence the FDI decision in his examination of the divestment process. CHAPTERV
CONCLUSIONS Introduction The final chapter of this dissertation addresses this study’s last major objective; namely, to assess the implications of the findings for the organizational structure of MNEs and the public policies of nations. However, before proceeding with this task, it may be worthwhile briefly reviewing what has been accomplished thusfar. By this point it should be clear that this project has proceeded along 2 diverse, yet equally important, lines. On one hand, an attempt has been made to develop a rigorous theoretical argument which relates changes in the international environment to corresponding changes in
MNE structure and behavior; while on the other, a rather limited aspect of that argument has been tested empirically. The objective of this approach has been to identify and test a critical link in the study’s basic thesis, and then use the results of that test to support or reject the argument itself.
Chapters I & II of this work were devoted to the theoretical development of the following thesis; that recent changes in the international environment have increased the efficiency of international markets which should encourage
MNEs to conduct a greater portion of their activities through the external marketplace. This occurs because as the efficiency of international markets increases, the benefits associated with coordinating activities through the
“internal" market of the firm decrease accordingly. This shift in the costs and benefits of internal organization should induce a corresponding shift in the way
228 229 MNEs structure their international operations. Specifically, one would expect to see a move away from equity-based control, toward more open, participative structures which allow MNEs to manage their activities through contractual arrangements with independent firms in several key markets. Using established theories and research from the fields of International Business, Strategic
Management, and Organizational Theory, the logical antecedents for this argument have been examined in detail, and a great deal of theoretical support has been presented for it.
But while developing the dissertation’s theoretical foundation was relatively straightforward, testing it empirically proved to be difficult indeed. Not only was accurate information on many types of MNE activity unavailable, but what little data could be found was often not comparable over time or between countries. As a result, it was necessary to greatly restrict the scope of the empirical study in order to increase its strength and validity. Chapters III & IV discuss these problems in detail and present the results of an empirical investigation into the impact of critical environmental factors on U.S. MNEs’ use of external markets. Specifically, the universe of U.S. foreign affiliates was examined in an attempt to determine the relationship between various country- specific factors (many of which are proxies for the efficiency of local markets), and U.S. MNEs’ choice between using a majority-owned foreign affiliate
(MOFA) or a minority-owned affiliate to enter and service a particular foreign market. But while the results of this investigation are highly valid in and of themself, their relationship to the study’s larger thesis is another matter entirely.
Thus, in discussing the results presented in the preceding chapter, 2 important questions must be answered. First, ‘What do the findings suggest about the relationships between specific environmental factors and certain 230 types of U.S. MNE behavior?’. And second, ‘What implications do they have for the larger thesis this work has tried to develop?’. This distinction is important, because each of these questions focuses on a very different aspect of the empirical study.
Further, the second question is complicated by the presence of 3 critical assumptions which relate the empirical study to the larger thesis. These are: 1)
The behavior of U.S. MNEs is representative of the behavior of all MNEs. This is not at all unreasonable considering the influence U.S. firms and management practices have had during most of the post-War era. 2) The examination of certain environmental differences between countries can be used to study changes in the international environment over time. Clearly, this assumption is somewhat more tenuous. However, it rests on the notion that various country- specific factors provide an indication of the efficiency of local markets, and that by examining MNE behavior in a broad cross-section of countries, some information can be gained on the impact recent changes in the international environment have had on firm behavior. Finally, 3) Differences in the amount of
U.S. MNE activity conducted through majority- and minority-owned affiliates can be used as proxies for the use of equity and contractual control respectively.
This means that several common forms of contractual control (such as Strategic
Alliances, R&D aggrements, Consortia, etc.) have not been examined in this study. Although this was necessary due to the lack of reliable data on these other arrangements, it is also reasonable given the prominence minority joint ventures have received in the literature.
Thus, while it is a simple matter to evaluate the relationships between 12 country-specific factors and 4 measures of the amount of activity U.S. MNEs conduct through minority-owned affiliates; it is much more difficult to assess the 231 implications these relationships have for the larger question of if and how recent changes in international markets will affect MNE structure and behavior. The remainder of this chapter is devoted to a discussion of each of these issues.
Discussion of tbfi-EmpidcflLSiudy I. Introduction
As noted above, Chapter IV contains the results of an empirical investigation into the relationship between 12 country-specific factors and 4 measures of the external activity of U.S. MNEs. Based on a broad review of relevant literature, it was hypothesized that in countries with higher levels of economic and social development, more open and participative political systems, and more efficient external markets, U.S. MNEs would conduct a greater portion of their local activities through minoritv-owned affiliates. This is consistent with the notion that the costs of using external markets in these countries is lower (relative to other countries), and therefore, firms should be encouraged to make greater use of the external marketplace. In addition, these ideas are supported by countless studies in International Business which have stressed the strategic importance of Joint Ventures, Strategic Alliances, and various other forms of contractual arrangements between firms, and have noted a veritable “explosion” in their use over the past decade. Unfortunately, neither the hypotheses nor the explosion of U.S. minority-owned foreign affiliates were supported by the empirical study.
In general, there was no evidence that a substantial increase in either the number of U.S. minority-owned foreign affiliates or the percentage of foreign assets held in these arrangements has occurred over the past 25 years. On the contrary, the results appear to indicate that U.S. firms are making greater use of 232 internal not external markets. This conclusion stems from the slight increase
observed in the percentage of total U.S. MOFA sales going to affiliated parties during the 1966-1989 period. Further, in 1989, over 80% of the total foreign assets of U.S. corporations were held in majority-owned foreign affiliates. This
represented over 22% of the total assets of U.S. parent corporations. On the other hand, minority-owned affiliates accounted for only 5% of total U.S. parent assets in 1989. Taken together, these results may suggest that U.S. MNEs have consolidated their foreign operations and are exercising even tighter control over their global subsidiaries than they did 2-3 decades ago.
This, in and of itself, is not surprising; but it must be evaluated in conjunction with the almost total lack of support for the hypothesized relationships between environmental factors and the MNE’s use of external markets. Not only did a country’s level of development, its political system, and the “openness” of its markets have little impact on the amount of local activity
U.S. firms conducted through minority-owned affiliates. In most cases the exact opposite appears to be true. In general, U.S. firms made greater use of maioritv- owned affiliates in countries with higher levels of economic and social development. Further, U.S. firms appeared to be much less affected by Host government restrictions on trade and FDI and moderate levels of political risk than is commonly believed.
One reaction to these unexpected findings is to question their validity and search for some blatant methodological error. But while the methodological problems discussed in Chapter III have undoubtedly added some margin of error to these results, that error must be considered relatively minor. In designing the empirical study, every effort was made to address the problems associated with the use of secondary data, and to select Independent and 233 Dependent Variables which maximize the explanatory power of available data sources while minimizing the technical limitations of the statistical tools being employed. Thus, the results of this investigation can be considered an accurate portrayal of the relationships between the specific environmental factors and the various measures of U.S. MNE behavior being studied, within the population of countries hosting considerable U.S. MNE investment.
In one way this is unfortunate because it means that alternative explanations for these findings must be developed. Within this context, 2 possiblities become apparent. The first explanation is that the study’s hypotheses are simply wrong and that country-specific factors associated with higher levels of development, more open political systems and higher levels of market efficiency do nat encourage MNEs to conduct a greater portion of their activities through “external” markets. The second is that the hypotheses are correct, but U.S. firms have either failed to recognize that fact or act upon it.
II. Some Possible Explanations
Assuming for a moment that the study’s hypotheses are not at fault, several interesting observations can be made about the behavior of U.S. MNEs and their managers. The most obvious is that U.S. firms seem to be making strategic decisions based on a view of the World that is no longer realistic or relevant. For example, many of the study’s findings seem to be consistent with the Strategic Management and International Business research of the 1970s and early 1980s which argued that the need to achieve economies of scale and increased efficiencies in production and distribution were the driving forces behind MNE strategy. This approach encouraged firms to rationalize their foreign operations and integrate them into tightly controlled global networks based on equity, not contractual, control. On the other hand, this dissertation 234 has argued that recent changes in the international environment have created new strategic imperatives which have far-reaching implications for MNE structure and behavior. Clearly, this “new” international environment will require firms to develop equally novel approaches to managing their international operations.
Few people would argue that the World is the same place today (or in
1989) that it was in 1966, and yet U.S. MNEs continue to structure their international activities in much the same way they did 25 years ago. The obvious question is why? Three reasons may be: 1) Managers and strategic planners of U.S. MNEs have failed to recognize the changes taking place in the international environment; and therefore, continue to use the outmoded strategies of the past. 2) Managers recognize that changes have occurred and are responding to them; but due to the overwhelming size of their existing foreign investments, these responses have as yet had little impact on the aggregate statistics which might reveal them. Finally, 3) Managers are aware of the changes taking place and recognize the need to adopt new approaches, but continue to use inappropriate strategies for some other reason.
It may be worthwhile briefly discussing each of these options. In the first case, although it is difficult to believe that managers are unaware of the events taking place in the World around them (particularly given the importance of operating and competing globally), there are several factors that may affect the ability of managers and strategic planners to analyze the international environment in an accurate, objective manner. In fact, many of these factors originate within the firm itself and act to constrain or determine the perceptions and actions of the organization’s members. This phenomenon has been the focus of a great deal of recent research in the field of Organizational Behavior 235 (see Weick, 1979) which suggests that managers accept a view of the World that is more consistent with the historical actions and perceptions of their organizations than objective reality, if this in fact true, many managers may be unable to identify the strategic imperatives facing their firms, and thus continue to act in ways that have little or no relation to the current international environment.
A somewhat different influence on managers’ perceptions is the popular media, the business press, and the academic community. It is equally likely that these groups have inadvertently promoted a "Worldview” that, although widely accepted by managers, is largely superficial and inaccurate. A cursory review of the Wall Street Journal, Business Week, Fortune, or any number of other popular business publications reveals a painful lack of insightful analysis of many critical International Business issues. These problems are compounded by the inaccessibility of most academic journals, which have often addressed the rather narrow needs of the academic community rather than the much broader needs of international managers and other practitioners. As a result, many managers may be totally unaware of the latest International Business and
Strategic Management research; and therefore, fail to recognize the strategic options available to their companies.
Within this regard, it is interesting to note that several major academic journals are in the process of revising their format to address exactly this problem. Perhaps the most notable changes have occurred in The Columbia
Journal of World Business, but the California Management Review, Sloan
Management Review and several start-up journals appear to be directing considerably more attention to the interests and needs of international managers. Thus, it may not be unreasonable to suggest that this study’s 236 findings reflect a general failure on the part of international managers and
corporate strategic planners to recognize many of the changes taking place in
the international environment and to implement appropriate strategic responses
to these changes.
The second explanation advanced above stems from the problems
associated with using aggregate data. Although the validity of the empirical
study is strengthened by examining the entire universe of U.S. direct
investment, it is clear that the cumulative effect of prior investments may obscure
many of the activities this study has tried to uncover. This is particularly true
since the dollar amount invested in the average majority-owned foreign affiliate
is likely to far exceed the amount invested in the average minority-owned
affiliate. Thus, it could be very difficult to identify subtle shifts in the investment
patterns of U.S. MNEs through aggregate statistics on all U.S. direct investment
abroad. This argument makes a great deal of intuitive sense and may imply that the slight differences observed in various indicators between 1966 and 1989 are far more meaningful than they first appear.
Further, it might be useful to investigate changes in the investment patterns of specific industries during the 1966-1989 period. Since some industries can be expected to respond to environmental changes more rapidly than others, it may be possible to identify significant differences in the amount of
U.S. MNE activity conducted through external markets between industries, and then relate these differences to changes in the external environment. Clearly, this is an interesting area of future research.
The final suggestion proposed above is perhaps the most intriguing;
namely, that managers recognize the need to change their organizational structure and/or behavior, and yet fail to do so. There are a number of reasons 237 why this situation might occur, one of which involves the concept of “culture lag” developed by William Ogburn (Ogburn, 1957). Ogburn notes that it is common to find a “lag” between changes in social behavior and the external events or stimuli which precipitate these changes. This phenomenon is readily apparent when the much slower pace of Social change is compared to the rapid advances in such areas as technology. In terms of the present discussion, it is quite possible that managers recognize the need to respond to recent changes in the external environment, but presently lack the tools necessary to do so.
Thus, manager’s apparent inability to react to recent environmental stimuli may not be a reflection of their ability to perceive or identify these events, as much as it is an indication of their inability to accept and adapt to them.
A related problem is that managers may not know exactly how to respond to various technological and environmental changes. This problem is exacerbated by the high costs and risks (both real and perceived) associated with restructuring an MNE, or implementing a new corporate strategy.
Consequently, managers have little incentive to take actions which may result in extremely costly and uncertain outcomes, and may adopt a “wait and see” approach, and only implement a particular strategy after other firms have proven its effectiveness. Knickerbocker’s (1973) study of MNE investment behavior clearly supports this contention, as do many of the “biological” theories of organizations. In fact, this phenomenon can be easily explained in terms of the “population ecology” approach to organizations.
Based on various aspects of the theories of Environmental Adaptation and Organizational Learning, the following argument can be constructed: Within any population of firms (i.e. industry, niche, etc.), members are continually evolving and adapting to various conditions in the external environment. 238 However, when one examines this process at a specific point in time, one finds
a small number of firms implementing new approaches, the majority of firms
behaving in an historically consistent manner, and a small number of firms in
decline. Since not all firms that adopt a new strategy will succeed, firms have a
strong incentive to continue to use strategies that have been successful in the
past. Thus, most firms remain followers rather than leaders. This is an extremely
important and interesting aspect of organizational and individual behavior
which may go a long way in explaining the reluctance of most firms to respond
to novel external stimuli.
There is another related problem which stems from the nature of
organizations themselves. The power of an organization to influence the
perceptions and actions of its members has already been discussed above; but
some writers have also noted that, at least on some levels, organizations
appear to have a life of their own. This results from the structures, processes,
and behaviors that have been built into the organization over time (what Bartlett
and Ghoshal call “administrative heritage”). These forces can an organization cause it to act in ways which transcend the collective power of its individual
members. Further, they may be amplified by other factors which operate on the
individual level, such as resistance to change, the fear of losing one’s job, power, or status, and the perceived risk involved in novel behavior. Clearly, to the extent that these forces are active, it becomes an increasingly difficult and time consuming process to change the structure and behavior of the firm.
Thus, it is quite plausible that individual managers recognize the need to adopt certain organizational changes and yet are totally powerless to bring them about. Once again, the results of the current empirical study may reflect the influence of these more subtle characteristics of organizations more than 239 anything else. If this is true, it underscores the need to train managers to be more receptive to the opportunities created by changes in the external environment. This requires adopting not only novel approaches to Business education, but also more open, supportive organizational environments which encourage individual managers to engage in alternative behaviors. Clearly, this is a difficult objective to achieve, particularly considering the size and complexity of most MNEs. However, given the ever increasing pace of technological and environmental change, it may be a goal that is directly associated with the continued success of the firm.
III. The Impact of Environmental Factors
At this point it is probably worthwhile discussing the second major aspect of the empirical study; namely, the relationship between environmental factors and U.S. MNE behavior. Perhaps the most important observation to emerge in this regard is that, although a small group of environmental factors can explain a considerable amount of the variance in U.S. MNEs’ decision to use majority- or minority-owned affiliates in a particular country, there is a great deal of variance that is not explained by these factors.
This is important for 2 reasons: first, it suggests that some other group of factors plays a major role in the MNE’s decision to enter a particular foreign market, and in its choice between alternative entry strategies; and second, it raises the interesting question of whether U.S. firms are paying enough attention to environmental conditions in the Host country when they make investment decisions. In either case it is clear that the “Structuralist” approach used throughout this study is capable of explaining only one aspect of firm behavior. In order to fully understand the operations and activities of the MNE, several different approaches to organizations must be used concurrently. Within 240 this context, many of the "social” or "behavioral” theories of organizations discussed in Chapter li of this work may provide valuable insights into other critical determinants of firm behavior in general, and the foreign investment decision in particular.
On the other hand, it is also possible that managers are simply unable to incorporate a detailed analysis of the environment into their decision-making calculus. Two obvious reasons for this are: 1) many international managers and firms may find it difficult to obtain accurate, reliable information on specific environment factors, and 2) even if information is available, the perceived complexity of the external environment may discourage many firms from including it in their decision-making process. As a result, many firms may be making strategic decisions based on a set of criteria which are largely unrelated to the ultimate objectives they hope to achieve. To the extent that this is true, it may account for the declining competitiveness of many U.S. firms and industries. Thus, an investigation into the criteria and processes managers and strategic planners actually use to make foreign investment decisions may be an interesting, albeit challenging, area of future research.
IV. Public Policy Implications
Finally, a brief word on the Public Policy implications of this study are in order. Unfortunately, one is left with the rather pessimistic conclusion that (at least U.S.) MNEs prefer to conduct their activities through internal markets in most cases. In fact, the only environmental factor that had a major impact on the investment practices of U.S. MNEs was local restrictions on trade and FDI
(REST). As a result, it appears that policy-makers must continue to balance their country’s need for foreign investment with their desire to control and share the profits generated by these activities. This is a classic Public Policy dilemma, and 241 it will not be solved here; however, one novel suggestion can be made.
A great deal of the preceding discussion has revolved around the notion that the managers and strategic planners of U.S. MNEs are behaving in ways which may be irrational given the current global environment. If this is true (i.e. the thesis of this study is largely correct), then policy-makers must perform the rather difficult task of educating MNEs about the potential benefits of using external markets and engaging in local minority Joint Ventures. This provides a positive policy option which has the objective of increasing both economic activity in the Host country and the efficiency of the MNE. Clearly, this approach is preferable to imposing restrictions on trade and FDI (even though these restriction appear to modify MNE behavior) due to the negative effects such actions may have on the local activities of foreign firms in the future.
Implications for the Study's Basic Thesis
Based on the preceding discussion, it is necessary to address a second major question; namely, 'What does all of this say about the validity of the theoretical argument developed in the first 2 chapters of this dissertation?’, in its simplest form, this dissertation has argued that recent changes in the international environment should encourage (U.S.) MNEs to conduct a greater portion of their foreign activities through external markets (or more specifically, through local minority-owned foreign affiliates). Unfortunately, in most cases, the results of the empirical study do not support this argument. But while U.S. MNEs clearly prefer to manage their foreign activities through equity control and internal markets (as evidenced by their overwhelming use of majority-owned affiliates), this does not necessarily mean that the thesis of this study is untenable. 242 Nonetheless, one is placed in the rather unenviable position of having to decide whether the arguments and hypotheses advanced throughout this work are simply wrong and the benefits associated with contractual control have been grossly exaggerated, or whether the theories are correct but managers have failed to recognize the potential benefits of these alternative arrangements. This choice is complicated by 2 additional considerations: first, the thesis of this dissertation is clearly supported by established theories in the fields of Organizational Behavior, Strategic Management, and International
Busines; and second, the empirical study did identify differences in the investment behavior of U.S. MNEs between countries, but for the most part they were in the opposite direction that hypothesized.
Rather than attempt to resolve this dilemma, it is enough to note that U.S.
MNEs appear to be behaving in ways that are antithetical to the predictions of some established theories. A number of ideas have already been advanced to explain this finding; but in the present context, this quandary raises 2 extremely interesting questions. First, ‘What criteria should be used to evaluate Business research, empirical or theoretical evidence?’. Or put another way, 'Does the fact that businesses behave in a certain way necessarily invalidate a theoretical argument that they should behave in some other way?’. And second, ‘Who should determine the agenda for research in the fields of Business and
Management, the academic community or actual business practitioners?’. More specifically, ‘Should the needs of the business community dictate the research interests of the academic community?’. Clearly, these are not simple questions and their answers will have far-reaching implications for the future of both business researchers and practitioners. 243 There is little doubt that the interests of these 2 constituencies are inextricably linked. Researchers need practitioners and organizations to study, while managers need the academic community to help them identify and address various critical business issues. Yet, on another level, these 2 groups have very few values in common. Clearly, the fields of Business and
Management are “applied” sciences, and therefore, they have a responsibility to engage in research that is directly related to the activity of business. But at the same time, the academic community also bears the responsibility for evaluating business activities within a larger theoretical and social context. This is particularly true of research in the field of International Business due to its inter disciplinary nature, and the immense power MNEs have to affect the larger international community and global environment. Further, a necessary conflict of interest arises when business practitioners are solely responsible for assessing their own actions.
if one accepts this view, the growing coincidence of interests between the business and academic communities should be a source of some concern. Not only are a number of academic journals modifying their formats to become more accessible to managers, but many leading business researchers are focusing an inordinate amount of attention on the immediate needs of various non academic constituencies (most notably the Government, business and managers, and the popular media). While these actions may facilitate the dissemination of some academic research to a larger audience, they also run the risk of altering the priorities of researchers by encouraging them to focus on a very narrow range of topical issues determined by the immediate economic needs of a particular business group. This is dangerous because it leaves less support for the type of theoretical research required to address much broader 244 issues such as the Nature of Business and the role of Business in Society.
In its own way, this dissertation has attempted to address these issues by developing a logical argument about the way businesses should behave. While such normative studies are accepted by the practitioner community when they offer immediate solutions to well-defined business problems, they are much less welcome when they investigate the broader, more amorphous areas of business research. Further, perhaps as a result of their perceived irrelevance to the business community, studies into these less defined (but often important) problems are becoming the exception, not the rule in the fields of Strategic
Management, International Business, and other related areas. This may be an extremely short-sighted approach that could ultimately inhibit the kind of scholarly activity necessary to address the many complex issues associated with the activities of Multinational Enterprises.
Fortunately, some writers continue to pursue these more esoteric lines of research, as evidenced by numerous works reviewed in this dissertation. But the kind of visionary investigations conducted by Adam Smith, David Ricardo,
Karl Marx, Max Weber, John Maynard Keynes, and many others, seem conspicuously absent from the current fare of “quick-fix”, “how-to” strategy manuals and business handbooks. To the extent that this perception is true, it poses a serious threat not only to the future of our various fields, but also to the larger community of nations and peoples that are directly affected by the operations of Multinational Enterprises and the activity of Business in general. LIST OF REFERENCES
Abell, Derek and Hammond, John; Strategic Market Planning: Problems and Analytical Approaches: Englewood Cliffs, N. J.; Prentice-Hal I; 1979
Abernathy, William and Chakravarfhy, Balaji; “Government Intervention and Innovation in Industry: A Policy Framework”; Sloan Management Reveiw; Spring, 1979; pgs. 3-18
Adelman, Irma and Morris, Cynthia; Society. Politics, and Economic Development: Baltimore; Johns Hopkins University Press; 1967
Ajami, Riad; “Global Strategic Alliances: The New Transnationals”; Working Paper; 1989
Ajami, Riad and Khambata, Dara; “Global Strategic Alliances: The New Transnationals”; Journal of Global Marketing; 1991, Vol. 5, #1/2
Akhter, Humayun and Lusch, Robert; “Political Risk: A Structural Analysis”; Advances in International Marketing; 1987, Vol.2; pgs. 81-101
Alchian, Armen and Demsetz, Harold; “Production, Information Costs and Economic Organization”; American Economic Association, 1972, Vol. 62, #5; pgs. 777-795
Allen, Thomas; Utterback, James; et. al.; “Government Influence on the Process of Innovation in Europe and Japan"; Research Policy; 1978, Vol. 7; pgs. 124- 149
Aliber, R.; “A Theory of Direct Investment"; in Kindleberger, C. (ed.); The International Corporation: Cambridge, MA.; MIT Press; 1970
Amine, Lyn; “Toward a Conceptualization of Export Trading Companies in World Markets”; Advances in International Marketing; Vol. 2,1987; pgs. 200-238
Anderson, Erin and Gatignon, Hubert; “Modes of Entry: A Transaction Cost Analysis and Propositions”; Journal of International Business Studies; Spring, 1986
245 246 Argyris, Chris and Shon, Donald; Organizational Learning: A Theory of Action Perspective: Reading, MA.; Addison-Wesley; 1978
Armour, Henry and Teece, David; “Organization Structure and Economic Performance: A Test of the Multidivisional Hypothesis”; Bell Journal of Economics; 1978, Vol. 9, #1; pgs. 106-122
Auster, Ellen; “International Corporate Linkages: Dynamic Forms in Changing Environments”; Columbia Journal of World Business; Summer, 1987; pgs. 3-6
Bain, Joe; Industrial Organization: New York; John Wiley & Sons; 1956
Banks, Arthur (ed.); Political Handbook of the World: Binghampton, N.Y.; CSA Publications; 1990
Bartlett, Christopher; “Multinational Structural Change: Evolution versus Reorganization”; in Otterbeck, Lars (ed.); The Management of the Headauarters-Subsidiarv Relationships in Multinational Corporations; Hampshire, England; Gower Publishing Company Ltd.; 1981
Bartlett, Christopher and Ghoshal, Sumantra; “Tap Your Subsidiaries for Global Reach"; Harvard Business Review; Nov.-Dee., 1986; pgs. 87-94
Bartlett, Christopher and Ghoshal, Sumantra; “Organizing for Worldwide Effectiveness"; California Management Review; Fall, 1988; pgs. 54-74
Bartlett, Christopher and Ghoshal, Sumantra; Managing Across Borders: The Transnational Solution: Boston; Harvard Business School Press;1989
Beamish, Paul; “The Characteristics of Joint Ventures in Developing Countries"; Columbia Journal of World Business; Fall, 1985; pgs. 13-19
Beamish, Paul; “Joint Ventures in LDCs: Partner Selection and Performance”; Management International Review; 1987 Vol. 27, #1; pgs. 23-37
Behrman, Jack; “Transnational Corporations in the New International Economic Order”; Journal of International Business Studies; Spring/Summer, 1981; pgs. 29-42
Behrman, Jack; “The Future of International Business and the Distribution of Benefits”; Columbia Journal of World Business; 1985, 20th Anniversary Issue; pgs. 15-22
Blair, Frank (ed.) Countries of the World and Their Leaders Yearbook: Detroit; Gale Research Inc.; 1991 247 Blinder, Alan; “A Japanese Buddy System That Could Benefit U.S. Business”; Business Week; Oct. 14,1991; pg. 32
Boddewyn, Jean; “Foreign and Domestic Divestment and Investment Decisions: Like or Unlike?”; Journal of International Business Studies; Winter, 1983; pgs. 23-35
Boddewyn, Jean; “Political Aspects of MNE Theory”; Journal of International Business Studies; Fall, 1988; pgs. 341-363
Boddewyn, J. , Soehl, Robin and Picard, Jacques; “Standardization in International Marketing: Is Ted Levitt in Fact Right?”; Business Horizons; Nov.- Dee., 1986; pgs. 69-75
Boyacigiller, Nakiye and Adler, Nancy; “The Parochial Dinosaur: Organizational Science in a Global Context”; Academy of Management Review; 1991, Vol.16, #2; pgs. 262-290
Brooks, Harvey; “Technology as a Factor in U.S. Competitiveness"; in Scott, Bruce and Lodge, George (ed.); U.S. Competitiveness in the World Economy: Boston; Harvard Business School Press; 1984; pgs. 328-356
Buckley, Peter and Casson, Mark; The Future of the Multinational Enterprise: London; Macmillian; 1976
Buckley, Peter; “A Critical Review of of Theories of the Multinational Enterprise"; Aussenwritschaft, Vol. 36,1981; pgs. 70-87
Buckley, Peter; “The Limits of Explanation: Testing the Internalization Theory of the Multinational Enterprise”; Journal of International Business Studies; Summer, 1988; pgs. 181-193
Burns, Tom and G.M. Stalker: The Management of Innovation: London; Tavistock; 1961
Burrell, G. and Mogan, Garth; Sociological Paradigms and Organizational Analysis: London; Heinemann Educational Books; 1979
Business International; Investing. Licensing and Trade Conditions Abroad (various Countries); New York; Business International Corp.; 1990-91
Business Week; “The Age of Consolidation"; Oct, 14,1991; pgs. 86-94 248 Calvet, A. L.; “A Synthesis of Foreign Direct Investment Theories and Theories of the Multinational Enterprise”; Journal of International Business Studies; Spring/Summer, 1981; pgs. 43-59
Campbell, Donald and Stanley, Julian; Experimental and QuasirExperimental Designs for Research: Boston; Houghton Mifflin Company; 1963
Carlos, Ann and Nicholas, Stephen; “Giants of Earlier Capitalism: The Chartered Trading Companies as Modern Multinationals”; Business History Review; Autumn, 1986; pgs. 398-419
Casson, Mark; “Transaction Costs and the Theory of the Multinational Enterprise”; in Rugman, Alan (ed.); New Theories of the Multinational Enterprise: Bechenham, England; Croom Helm Ltd.; 1982; pgs. 24-43
Casson, Mark (ed.); The Growth of International Bussiness: London; Allen and Unwin; 1983
Casson, Mark, et. al.; Multinationals and World Trade: London; Allen and Unwin; 1986
Caves, Richard; “International Corporations: The Industrial Economics of Foreign Investment"; Economica; Feb. 1971; pgs. 1-27
Cavusgil, S.; “Guidelines for Export Market Research”; Business Horizons; Nov.- Dee., 1985; pgs. 27-33
Chandler, Alfred; Strategy and Structure: Chapters in the History of the Industrial Enterprise: Cambridge, MA; MIT Press; 1962
Chandler, Alfred, Bruchey, Stuart and Galambos, Louis; The Changing Economic Order: Readings in American Business and Economic History: New York; Harcourt, Brace & World Inc.; 1968
Chandler, Alfred; “The Multi-Unit Enterprise: A Historical and International Comparative Analysis and Summary”; in Harold Williamson (ed.); Evolution of International Management Structures: Philidelphia; University of Delaware Press; 1975
Chandler, Alfred; The Visible Hand: The Managerial Revolution in American Business: Cambridge, MA; Belknap Press of Harvard University Press; 1977
Chandler, Alfred; “The Evolution of Modern Global Competition”; in Porter, Michael (ed); Competition In Global Industries: Boston; Harvard Business School Press; 1986; pgs.405-448 249 Clegg, Jeremy; Multinational Enterprise and World Competition; A Comparative Study of the USA. Japan. UK. Sweden and West Germany; London; Macmillan Press; 1987
Clegg, Jeremy; “The Determinants of Aggregate international Licensing Behavior: Evidence from Five Countries”; Management International Review; 1990, Vol. 30, #3; pgs. 231-251
Coase, R. H.; “The Nature of the Firm"; Economica, 1937, #4; pgs. 386-405
Cohen, Jacob and Cohen, Patricia; Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences: Hillsdale, N. J.; Lawrence Erlbaum Associates, Publishers; 1983
Commons, J. R.; Institutional Economics: Madison; University of Wisconsin Press; 1934
Contractor, Farok; “The Role of Licensing in International Strategy”; Columbia Journal of World Business; Winter, 1981; pgs. 73-81
Contractor, Farok; “Contractual and Cooperative Forms of International Business: Towards a Unified Theory of Modal Choice”; Management International Review; 1990, Vol.30, #1; pgs. 31-54
Crazier, Michel: The Bureaucratic Phenomenon: Chicago; University of Chicago Press; 1964
Cyert, Richard and March, James; A Behavioral Theory of the Firm: Englewood Cliffs, N.J.; Prentice-Hall; 1963
Daniels, John; Pitts, Robert and Tretter, Marietta; “Strategy and Structure of U.S. Multinational: An Exploratory Study"; Academy of Management Journal; 1984, Vol.27, #2; pgs. 292-307
Datta, Deepak; “International Joint Ventures: A Framework for Analysis”; Journal of General Management; Winter, 1988; pgs. 78-91
Davidson, William; “Administrative Orientation and International Performance”; Journal of International Business Studies; Fall, 1984; pgs. 11-23
Davis, James; Elementary Survey Analysis: Englewood Cliffs, N. J.; Prentice- Hall; 1971
Davis, Stanley and Lawrence, Paul; Matrix: Reading, MA; Addison-Wesley Publishing Co.; 1977 250 de la Torre, Jose and Neckar, David; “Forecasting Political Risks for International Operations”; International Journal of Forecasting; 1988, Vol. 4; pgs. 221-241
Deiury, George (ed.); World Encyclopedia pf..Political-Sy.slems and Parties; New York; Facts on File Publications; 1987
Dixon, Chris: Rural Development in the Third World: London: Routledge; 1990
Douglas, Susan and Craig, C.; “Information for International Marketing Decisions”; in Tracey, Murray and Ingo, Walter (eds); International Business Handbook; 1988-9
Douglas, Susan and Wind, Yoram; “The Myth of Globalization”; Columbia Journal of World Business; Winter, 1987; pgs. 19-29
Doz, Yves; “Strategic Management in Multinational Companies”; Sloan Management Review; Winter, 1980; pgs 27-46
Doz, Yves and Prahalad, C. K.; “Patterns of Strategic Control Within Multinational Corporations”; Journal of International Business Studies; Fall, 1984; pgs. 55-71
Doz, Yves and Prahalad, C. K.; The Multinational Mission: Balancing Local Demands and Global Vision: New York; Free Press; 1987
Doz, Yves and Prahalad, C. K.; “Managing DMNCs: A Search for a New Paradigm”; Strategic Management Journal; Special Issue on Global Strategy; Summer 1991
Drucker, Peter; “The Changed World Economy”; The McKinsey Quarterly; Autumn, 1986; pgs. 2-26
Drucker, Peter; “The Coming of the New Organization”; Harvard Business Review; Jan.-Feb., 1988; pgs. 45-53
Dunning, John; American Investment in British Manufacturing Industry: London, England; Allen and Unwin; 1958
Dunning, John; “Trade, Location of Economic Activity and the MNE: A Search for an Eclectic Approach"; in Ohlin, B.; Hesselborn, P., and Wijkman. P. (eds.); The International Allocation of Economic Activity: London, England; Macmillian; 1977 251 Dunning, John; “Explaining Changing Patterns of International Production: In Defense of the Eclectic Theory"; Oxford Bulletin of Economics and Statistics; Nov., 1979; pgs. 269-295
Dunning, John; “Towards an Eclectic Theory of International Production”; Journal of International Business Studies; Spring/Summer, 1980; pgs. 9-31
Dunning, John; “The Eclectic Paradigm of International Prodiution: A Restatement and Some Possible Extensions”; Journal of International Business Studies; Spring, 1988(a); pgs. 1-31
Dunning, John; Explaining International Production: London; Unwin Hyman; 1988(b)
Dunning, John; Multinationals. Technology and Competitiveness: London; Unwin Hyman; 1989
Dunning, John and McQueen, Mathew; “The Eclectic Theory of the Multinational Enterprise and the International Hotel Industry”; in Rugman, Alan (ed.); New Theories of the Multinational Enterprise: Beckenham, England; Croom Helm Ltd.; 1982; pgs. 79-106
Dunning, John and Rugman, Alan; “The Influence of Hymer’s Dissertation on the Theory of Foreign Direct Investment”; American Economic Review; Vol. 75, #2,1985; pgs. 228-232
Dymsza, William; “Trends in Multinational Business and Global Environments: A Perspective"; Journal of International Business Studies: Winter, 1984; pgs. 25- 46
Economist, The; “Alphabet Soup”; Sept. 21,1991; pg. 33
Economist Intelligence Unit (EIU1: Country Profiles (various Countries); London; EIU; various years (1989-91)
Egelhoff, William; “Strategy and Structure in Multinational Corporations: An Information Processing Approach”; Administrative Science Quarterly; Sept. 1982; pgs. 435-458
Egelhoff, William; “Patterns of Control in U.S., UK, and European Multinational Corporations”; Journal of International Business Studies; Fall, 1984; pgs. 73-83
Egelhoff, William; “Strategy and Structure in Multinational Corporations: A Revision of the Stopford and Wells Model”; Strategic Management Journal; 1988, Vol. 9; pgs. 1-14 252 Encarnation, Dennis and Vachani, Sushil; “Foreign Ownership When Host Rules Change"; Harvard Business Review; Sept.-OctM 1985; pgs. 152-160
Ergas, Henry; "Does Technology Policy Matter?”; in Guile, Bruce and Brooks, Harvey (ed.); Technology and Global Industry: National Academy Press; 1987; pgs. 191-245
Euromoney; "A Slightly Riskier Place”; Sept., 1989; pgs. 206-7
Farge, Nathan and Wells, Louis; "Bargaining Power of Multinationals and Host Governments”; Journal of International Business Studies; Fall, 1982; pgs. 9-23
Farma, E. F. and Jensen, M. C.: "Separation of Ownership and Control”; Journal of Law and Economics; 1983; Vol. 26, #2; pgs. 301-325
Fatehi-Sedah, K. and Safizadeh, M.; "The Association Between Political instability and Flow of Foreign Direct Investment”; MIR; 1989, Vol. 29, #4; pgs. 4- 13
Fayol, Henri; General and Industrial Management: New York; Pitman; 1947 (original 1916)
Franko, Lawrence; “New Forms of Investment in Developing Countries by US Companies; A Five Industry Comparison”; Columbia Journal of World Business; Summer, 1987; pgs. 39-56
Franko, Lawrence; “Global Competition: Who’s Winning, Who’s Losing, and the R&D Factor a s One Reason Why”; Strategic M anagem ent Journal; Vol. 10, 1989; pgs. 449-474
Friedman, Roberto and Kim, Jonghoon; “Political Risk and International Marketing; Columbia Journal of World Business; Winter, 1988; pgs. 63-74
Gates, Stephen and Egelhoff, William; "Centralization in Headquarters- Subsidiary Relationships”; Journal of International Business Studies; Summer, 1986; pgs. 71-92
Geringer, Michael and Hebert, Louis; “Control and Performance of International Joint Ventures"; Journal of International Business Studies; Summer, 1989; pgs. 235-254
Ghoshal, Sumantra; “Global Strategy: An Organizing Framework"; Strategic Management Journal; 1987, Vol. 8; pgs. 425-440 253 Globerman, Steven; “Government Policies Toward Foreign Direct Investment: Has a New Era Dawned?”; Columbia Journal of World Business; Fall, 1988; pgs. 41-49
Gomes-Casseres, Benjamin; “Joint Venture Instability: Is It A Problem?”; Columbia Journal of World Business; Summer 1987; pgs. 97-101
Grant, Robert; “Multinationality and Performance Among British Manufacturing Companies”; Journal of International Business Studues; Fall, 1987; pgs.79-89
Hamel, Gary; “Competition for Competence and Inter-Partner Learning within International Strategic Alliances”; Strategic Management Journal; Special Issue on Global Strategy; Summer 1991
Hannan, Michael and Freeman, John; “The Population Ecology of Organizations”; American Journal of Sociology; Vol. 82, 1977; pgs. 929-964
Hannan, Michael and Freeman, John; Organizational Ecology: Boston; Harvard University Press; 1988
Harrigan, Kathryn; Strategies for Joint Ventures: Lexington, MA.; Lexington Books; 1985
Harrigan, Kathryn; “Joint Ventures and Competitive Strategies"; Strategic Management Journal; 1988, Vol. 9; pgs. 141-158
Haspeslagh, Philippe; “Portfolio Planning: Uses and Limits”; Harvard Business Review; Jan.-Feb., 1982; pgs. 58-73
Hays, William; Statistics: Fort Worth; Holt, Rinehart and Winston; 1988
Helleiner, Gerald; Intra-firm Trade and the Developing Countries: New York; St. Martin’s Press; 1981
Henderson, Bruce; Henderson on Corporate Strategy: Cambridge, MA.; Abt Books; 1979
Hennart, Jean-Francois; “A Transaction Costs Theory of Equity Joint Ventures”; Strategic Management Journal; 1988, Vol. 9; pgs. 361-374
Hennart, Jean-Francois; “Can the “New Forms of Investment” Substitute for the “Old Forms?” A Transaction Cost Perspective"; Journal of International Business Studies; Summer, 1989; pgs. 211-233 254 Hill, Charles; “Cooperation, Opportunism, and the Invisible Hand: Implications for Transaction Cost Theory”; Academy of Management Review; 1990, Vol. 15, #3; pgs. 500-513
Hill, Charles and Kim, W. Chan; “Searching for a Dynamic Theory of the Multinational Enterprise: A Transaction Cost Model”; Strategic Management Journal; 1988, Vol. 9; pgs. 93-104
Hill, Charles; Hwang, Peter and Kim, W.; “An Eclectic Theory of the Choice of International Entry Mode”; Strategic Management Review; 1990, Vol. 11; pgs. 117-128
Hirsch, S.; “An International Trade and Investment Theory of the Firm”; Oxford Economic Papers; Vol. 28, #2, July, 1976
Hirshleifer, J.; “Economics from a Biological Viewpoint”; Journal of Law and Economics; 1977, Vol. 20, #1; pgs. 1-52
Horaguchi, Haruo and Toyne, Brian; “Setting the Record Straight: Hymer, Internalization Theory and Transaction Cost Economics”; Journal of International Business Studies; Third Qt., 1990; pgs. 487-494
Hout, Thomas; Porter, Michael and Rudden, Eileen; “How Global Companies Win Out"; Harvard Business Review; Sept.-Oct. 1982; pgs. 98-108
Hymer, Stephen; The International Operations of National Firms: Cambridge, MA.; MIT Press; 1976 (original 1960)
Hymer, Stephen; “The Multinational Corporation and the Law of Uneven Development”; in Bhagwati, J. (ed.) Economics and the New World Order: New York; World Law Fund; 1971
International Monetary Fund; Direction of Trade Statistics Yearbook 1990: Washington, D.C.; IMF; 1991
International Monetary Fund; International Financial Statistics Yearbook 1990: Washington, D.C.; IMF; 1991
Itaki, Masahiko; “A Critical Assessment of the Eclectic Theory of the Multinational Enterprise”; Journal of International Business Studies; 1991, Vol. 22, #3; pgs. 445-460
Jain, Subhash; “Perspectives on International Strategic Alliances”; Advances in International Marketing; 1987 Vol. 2; pgs. 103-120 255 Jensen, M. C. and Meckling, W. H.; Theory of the Firm: Managerial Behavior, Agency Costs and Capital Structure”; Journal of Financial Economics; 1976, Vol. 3; pgs. 305-360
Johnson, H. G.; The Efficiency and Welfare Implication of the International Corporation”; in Kindleberger, C. (ed.); The international Corporation: Cambridge, MA.; MIT press; 1970
Johnston, Russell and Lawrence, Paul; “Beyond Vertical Integration-the Rise of the Value-Adding Partnership”; Harvard Business Review; July-August, 1988; pgs. 94-101
Jorde, Thomas and Teece, David; “Competition and Cooperation: Striking the Right Balance”; California Management Review; Spring 1989; pgs. 25-37
Kemp, Tom; Industrialization in the Non-Western World: New York; Longman; 1989
Keppel, Goeffrey; Design and Analysis: Englewood Cliffs, N.J.; Prentice Hall; 1982
Kester, W. Carl: Japanese Takeovers: The Global Contest for Corporate Control; Boston; Harvard Business School Press; 1991
Kidder, Louise: Research Methods in Social Relations: New York; Holt, Rinehart and Winston; published for the Society for Psychological Study of Social Issues (SPSSI); 1981
Killing, Peter; “How to Make a Global Joint Venture Work”; Harvard Business Review; May-June, 1982; pgs. 120-127
Killing, Peter; Strategies for Joint Venture Success: New York; Praeger; 1983
Kindleberger, Charles: American Business Abroad: New Haven; Yale University Press; 1969
Knickerbocker, F.; Oligopolistic Reaction and Multinational Enterprise: Cambridge, MA.; Harvard University Press; 1973
Kobayashi, Noritake; “Strategic Alliances with Japanese Firms”; Long Range Planning; 1988, Vol. 21, #2; pgs. 29-34
Kogut, Bruce; “Normative Observations on the International Value-Added Chain and Strategic Groups”; Journal of International Business Studies; Fall, 1984; pgs. 151-167 256 Kogut, Bruce; “Joint Ventures: Theoretical and Empirical Perspectives”; Strategic Management Journal; 1988, Vol. 9; pgs. 319-332
Kogut, Bruce and Singh, Harbir; “The Effect of National Culture on the Choice of Entry Mode”; Journal of International Business Studies; Fall, 1988; pgs. 411- 432
Kojima, K.; Direct Foreign Investment: A Japanese Model of Multinational Business Operations: London; Croom Helm; 1978
Korbin, Stephen; “Political Risk: A Review and Reconsideration”; Journal of International Business Studies; Spring/Summer, 1979; pgs. 67-80
Korbin, Stephen; “An Empirical Analysis of the Determinants of Global Integration”; Strategic Management Journal; Special Issue on Global Strategy, Summer 1991; pgs. 17-31
Kotler, Philip: “Global Standardization-Courting Danger”; Journal of Consumer Marketing; Spring 1986, Vol. 3, #2; pgs. 13-15
Krugman, Paul: Rethinking International Trade: Cambridge, MA.; MIT Press; 1990
Lawrence, Paul and Lorsch, Jay; Organizations and Environments: Cambridge, MA.; Harvard Business School; 1967
Lawrence, Paul; “The Harvard Organization and Environment Research Program"; in Van De Ven, Andrew and Joyce, William (ed.); Perspectives on Organization Design and Behavior: New York; John Wiley & Sons; 1981
Lemak, David and Bracker, Jeffrey; “A Strategic Contingency Model of Multinational Corporate Structure”; Strategic Management Journal; 1988, Vol. 9; pgs. 521-526
Levitt, Theodore; “The Globalization of Markets"; Harvard Business Review; May-June, 1983; pgs. 92-102
Lincoln, James and Kalleberg, Arne; Culture. Control, and Commitment: A Sludv of Work Organization in the United States and Japan: Cambridge, England; Cambridge University Press; 1990
Lodge, George and Walton, Richard; “The American Corporation and its New Relationships”; California Management Review; Spring, 1989; pgs.9-24 257 Lorange, Peter; “A Framework for Strategic Planning in Multinational Corporations”; Long Range Planning; June, 1976; pgs. 30-36
Lundstedt, Sven; "A Note on Asking Questions”; Journal of General Psychology; 1968, Vol. 79; pgs. 229-239
Magee, S. P.; Technology and the Appropriability Theory of the Multinational Corporation”; in Bhagwati, Jagdish (ed.); The New International Economic Order: Cambridge, MA.; MIT Press; 1976
Magee, S. P.; "Multinational Corporations, the Industry Technology Cycle, and Development”; Journal of World Trade Law; July-Aug., 1977; pgs. 297-321
Makler, Harry, Martinelli, Alberto and Smelser, Neil; The New International Economy: Beverly Hills, CA.; SAGE Publications Inc.; 1982
March, James and Simon, Herbert; Organizations: New York; John Wiley and Sons; 1958
McClain, D.; "Foreign Direct Investment in the United States: Old Currents, “New Waves", and the Theory of Direct Investment"; in Kindleberger, C. and Audretsch, D. (eds.); The Multinational Corporation in the 1980s: Cambridge, MA.; MIT Press; 1983
McManus, J.; The Theory of the International Firm"; in Paguet, G. (ed); Ibfi Internaional Firm and the Nation State: Toronto; Collier-Macmillian; 1972
Meltzer, Allan; T he Growth of Government Revisited”; in Tinbergen, Jan (ed.); Perspectives on an Economic Future: Forms. Reforms, and Evaluations: New York; Greenwood Press; 1991
Meyer, J. and Rowen, B.; "Institutionalized Organizations: Formal Structures as Myth and Ceremony"; American Journal of Sociology; 1977, Vol. 83; pgs. 340- 363
Miles, Raymond and Snow, Charles; Organizational Strategy. Structure and Process: New York; McGraw-Hill, 1978
Morris, Deigan and Hergert, Michael; Trends in International Collaborative Agreements"; Columbia Journal of World Business; Summer, 1987; pgs. 15-21
Morris, Morris; Measuring the Conditions of the World’s Poor: The Physical Quality of Life Index: London; Frank Cass; 1979 258 Myrdal, Gunnar; Against the Stream: Critical Essavs in Economics: New York; Pantheon Books; 1973
National Academy of Engineering; “The Competitive Status of U.S. industry”; Washington, D.C.; National Academy Press; 1985
National Academy of Engineering; “Perspectives on Technology and Industrial Competitiveness”; Transcript from an NAE Roundtable held Jan. 7,1986
Nelson, Richard; High Technoloov Policies: A Five Nation Comparison: Washington, D. C.; American Enterprise Institute for Public Policy Research
Netter, John and Kunter, Michael; Applied Linear Statistical Models: Regression. Analysis of Variance, and Experimental Designs: Homewood, III.; R.D. Irwin; 1985
Newbold, Paul; Statistics for Business and Economics: Englewood Cliffs, N.J.; Prentice Hall;1991
Ogburn, William; “Culture Lag as Theory”; Sociology and Social Research; Jan., 1957, Vol. XLI; pgs. 167-173
Ohmae, Kenichi; Triad Power: The Coming Shape of Global Competition: New York; Free Press; 1985
Ohmae, Kenichi; “Becoming a Triad Power: The New Global Corporation”; International Marketing Review; Autumn, 1986; pgs. 7-20
Ohmae, Kenichi; “The Global Logic of Strategic Alliances”; Harvard Business Review; March-April, 1989; pgs. 143-154
Ohmae, Kenichi; The Borderless World: Power and Strategy in the Interlinked Economy: New York; Harper Business, a Division of Harper Collins Publishers; 1990
Oman, Charles; “New Forms of Investment in Developing Countries”; Working Paper; Paris; OECD Development Center; 1984
Perlmutter, Howard; “The Tortuous Evolution of the Multinational Enterprise”; Columbia Journal of World Business; Jan.-Feb., 1969; pgs. 9-18
Perlmutter, Howard and Heenan, David; “Cooperate to Compete Globally” Harvard Business Review; March-April, 1986; pgs. 136-152 259 Pfeffer, Jeffery and Salancik, Gerald; The External Control of Organizations: A Resource Dependence Perspective: New York; Harper and Row; 1978
Porter, Michael; Competitive Strategy: Techniques for Analyzing Industries and Competitors: New York; Free Press; 1980
Porter, Michael; Competitive Advantage: Creating and Sustaining Superior Performance: New York; Free Press; 1985
Porter, Michael (ed.); Competition in Global Industries: Boston; Harvard Business School Press; 1986;
Porter, Michael; The Competitive Advantages of Nations: New York; Free Press; 1990
Prahalad, C.K. and Doz, Yves; “Strategic Control-The Dilemma in Headquarters-Subsidiary Relationship"; in Otterbeck, Lars (ed); The Management of the Headquarters-Subsidiary Relationships in Multinational Corporations: Hampshire, England; Gower Publishing Company Ltd.; 1981
Prahalad, C. and Doz, Yves; The Multinational Mission: Balancing Local Demands and Global Vision: New York; The Free Press; 1987
Price Waterhouse; Price Waterhouse Information Guide: Doing BuainessJn (Country XI: Price Waterhouse; various years (1985-1991)
Pugh, Dereck and Hickson, David; Organizational Structure in its Context: West me ad, England; Saxon House; 1976
Read, Robert; “The Growth and Structure of Multinationals in the Banana Export Trade”; in Casson, Mark: The Growth of International Business: London; George Allen & Unwin (Publishers) Ltd.; 1983
Robinson, Richard; “Background Concepts and Philosophy of International Business from World War II to the Present”; Journal of International Business Studies; Spring/ Summer, 1981; pgs. 13-21
Robinson, Richard; Performance Requirements for Foreign Business: U.S. Management Response: New York; Praeger Publishers; 1983
Root, Franklin; “Some Trends in the World Economy and Their Implications for International Business”; Journal of International Business Studies; Winter, 1984; pgs. 19-32 260 Rostow, Walt; The Stages of Economic Growth: A Non-Communist Manifesto: Cambridge, England; Cambridge University Press; 1991 (original 1960)
Rugman, Alan; “International Diversification by Financial and Direct Investment”; Weltwirschaftliches Archiv; Vol. 113, Sept. 1977
Rugman, Alan; “Internalization as a General Theory of Foreign Direct Investment; A Reappraisal of the Literature”; Weltwirschaftliches Archiv; Vol. 116, June, 1980; pgs. 365-379
Rugman, Alan (ed.); New Theories of the Multinational Enterprise: London; Croom Helm Ltd.; 1982
Rugman, Alan; “New Theories of the Multinational Enterprise: An Assessment of Internalization Theory”; Bulletin of Economic Research; 1986, Vol. 38, #2; pgs. 101-118
Rumelt, Richard; Strategy. Structure, and Economic Performance: Boston; Harvard Business School Press; 1974
Salvatore, Dominick; International Economics: New York; Macmillian Publishing Co.; 1987
Sanger, David; “Behind the Thai Boom: The Japanese”; New York Times', May, 10,1990
Sarathy, Ravi; “The Interplay of Industrial Policy and International Strategy: Japan’s Machine Tool Industry”; California Management Review; Spring, 1989; pgs. 132-160
Schumacher, Rose (ed.); World Economic Data: Santa Barbara; ABC-CLIO; 1989
Schumpeter, Joseph; Capitalism. Socialism and Democracy: New York; Harper & Brothers; 1942
Sethi, S. Prakash and Luther, K.A.; “Political Risk Analysis and Direct Foreign Investment: Some Problems of Definition and Measurement”; California Management Review; Winter, 1986; pgs. 57-68
Simon, Jeffrey; “Political Risk Assessment: Past Trends and Future Prospects”; Columbia Journal of World Business; Fall, 1982; pgs. 62-71
Southard, F. A.: American Industry in Europe: Boston; Houghton Mifflin; 1931 261 Spekman, Robert; “Strategic Supplier Selection: Understanding Long-Term Buyer Relationships”; Business Horizons; July-Aug., 1988; pgs.75-81
Stoever, William; “Endowments, Priorities, and Policies: An Analytical Scheme for the Formation of Developing Country Policy Toward Foreign Investment”; Columbia Journal of World Business; Fall, 1982; pgs. 3-15
Stopford, John and Wells, Louis; Managing the Multinational Enterprise: Organization of the Firm and Ownership of the Subsidiaries: New York; Basic Books, Inc.; I972
Taylor, Frederick: Scientific Management: New York; Harper & Row; 1947 [This volume contains both Shop Management (1903) and Principals of Scientific M anagem ent M 91111
Teece, David; “The Multinational Enterprise: Market Failure and Market Power Considerations”; Sloan Management Review; Spring, 1981; pgs. 3-17
Teece, David; “Transaction Cost Economic and the Multinational Enterprise: An Assessment”; Journal of Economic Behavior and Organization; 1986, Vol. 7; pgs. 21-45
Teichova, Alice, Levy-Leboyer, Maurice, and Nussbaum, Helga; The Multinational Enterprise in Historical Perspective: Cambridge, England; Cambridge Universitty Press; 1986
Thompson, Jam es D.; Organizations in Action: New York; McGraw-Hill; 1967
Thurow, Lester; “Revitalizing American Industry: Managing in a Competitive World Economy"; California Management Review’ Fall, 1984; pgs. 9-41
Thurow, Lester; “The Zero-Sum Solution"; Public Management; Dec. 1987; pgs. 12-19
United Nations Research Institute on Social Development (UNRISD); Contents and Measurement of Socioeconomic Development: Geneva; UNRISD; 1970
U.S. Department of Commerce; U.S. Direst Investment Abroad. 1966 Final Data: Washington, D.C.; U.S. Government Printing Office; 1974
U.S. Department of Commerce; U.S. Direst Investment Abroad. 1989 Benchmark Survey. Preliminary Results: Washington, D.C.; U.S. Government Printing Office; 1991 262 Van de Van, Andrew and Joyce, William; Perspectives on Organization Design and Behavior: New York; Wiley; 1981
Vernon, Raymond; “International Investment and International Trade in the Product Cycle”; Quarterly Journal of Economics; Vol. 80,1966; pgs. 190-207
Vernon, Raymond; “The Product Cycle Hypothesis in a new International Environment"; Oxford Bulletin of Economics and Statistics; Vol. 41, #4, Nov., 1979; pgs. 255-267
Vernon, Raymond; ‘Can U.S. Manufacturing Come Back?”; Harvard Business Review; July-Aug., 1986; pgs. 98-106
Vernon, Raymond; “Coping With Technological Change: U.S. Problems and Prospects”; in Guile, Bruce and Brooks, Harvey (ed.); Technology and Global Industry: National Academy Press; 1987
Vernon, Raymond; “Organizational and Institutional Responses to International Risk"; in Wortzel and Wortzel (ed.); Strategic Management of Multinational OQiporaiiQnaLlhe Essentials; 1990
Weber, Max: The Theory of Social and Economic Organization: New York; Free Press; 1947
Weick, K. E.; The Social Psvcholoov of Organizing: Reading, MA.; Addison Wesley; 1979
Weigal, Van: A Unified Theory of Global Development; New York; Praeger; 1989
Weisberg, Sanford; Applied Linear Regression: New York; Wiley; 1985
Weiss, Kenneth; “Management of the International Marketing Function”; in Tracey, Murray and Ingo, Walter (eds.); International Business Handbook: 1988/89
Wilkins, Mira; “The History of European Multinationals: A New Look”; Journal of European Economic History; Winter, 1986; pgs. 483-6
Williamson, Oliver; Markets and Hierarchies: Analysis and Antitrust Implications: New York; The Free Press; 1975
Wind, Yoram and Douglas, Susan; “International Portfolio Analysis and Strategy: The Challenge of the 1980s”; Journal of International Business Studies; Fall,1981; pgs. 69-73 263 Woodard, Joan; “Management and Technology"; Problems and Progress in Industry. 3; London; H.M.S.O.
Woodward, Joan; Industrial Organization: Theory and Practice: London; Oxford University Press; 1965
World Bank; World Development Report 1991: Oxford; Oxford University Press; 1991
Yamamura, Kozo; “A Compromise with Culture; The Historical Evolution of the Managerial Structure of Large Japanese Firms"; in Harold Williamson (ed.); Evolution of International Management Structures: Philidelphia; University of Delaware Press; 1975
Yip, George; “Global Strategy...ln a World of Nations?"; Sloan Management Review; Fall, 1989; pgs. 29-41
Yoffie, David and Miller, Helen; “An Alternative to Free Trade or Protectionism: Why Corporations Seek Strategic Trade Policy”; California Management Review; Summer, 1989; pgs. 111-131
Yoshino, M. and Lifson, Thomas; The Invisible Link: Japan's Sogo Shosha and the Organization of Trade: Cambridge, MA.; MIT Press; 1986
Zengage, Thomas and Ratcliffe, Tait; The Japanese Centurv: Challenge and Response: Hong Kong; Longman Group (Far East) Ltd.; 1988