The Productivity Puzzle: Numbers Alone Won't Solve It

A Review Essay The productivity puzzle : numbers alone won't solve it From vantage points in management, labor, academia, and government, contributors to four recent books grapple with the productivity slowdown, with little help from economic theory

PAUL S. ADLER

Over the last two decades, there has been a major de- originality in the capacity to link the two factors, in a cline in the rate of growth in U.S. productivity . The lag synthesis of micro- and macro-economics. But so far, in the ratio of output to input has also occurred in this approach has not shed light on the most elementary many other industrial countries, including Japan. part of the productivity puzzle: Is the productivity slow- Orthodox economic theory hypothesizes a basically down basically a cause or an effect of current economic technical link between trends in output and input, problems? namely, the production function. This hypothesis has Research on productivity thus progresses somewhat been put to a severe test, for the precise extent, the ori- unevenly . Three foci of study have emerged : data analy- gins, and the significance of the productivity slowdown sis, study of management practices, and research into are yet to be analyzed with a clarity that would demon- labor relations and conditions . Productivity: Prospects for strate the usefulness of traditional economics in ana- Growth, edited by Jerome Rosow, deals with all three lyzing such problems . subjects . The interdependence of the three themes There is, in particular, a surprising contrast between makes this presentation most judicious. Three other re- the wealth of studies that attempt to quantify the decline cent books have also addressed one or other of these and calculate its causes, and the poverty of material matters. Before discussing the major issues, we will on the role played by such a lag in macroeconomic per- identify the overlapping concerns of the four volumes. formance . It should be remembered that, in general, at a Productivity: Prospects for Growth includes five contri- company and an industry level, labor productivity and butions to the task of data analysis . Solomon Fabricant profitability are not well correlated, and that in capitalist of New York University and Dale Jorgenson of Har- economies decisions are based on the latter, not the for- vard University present the growth accounting data, Je- mer. Paul Samuelson's neoclassical paradigm claims its rome Mark of the Bureau of Labor Statistics, the mea- surement consideration, and Howard Samuel and Rudy Paul S. Adler, a research economist with the Employment Research Oswald of the AFL-CIO, their views on the role of for- Center of the French Ministry of Labor, has been a guest scholar at eign trade and labor unions. the Brookings Institution and the Bureau of Labor Statistics . He is The reader seeking more detail on current data analy- currently a visiting assistant professor of economics at Barnard Col- lege, Columbia University, N.Y . sis methods can consult Aggregate and Industry-Level

15 MONTHLY LABOR REVIEW October 1982 . The Productivity Puzzle

Productivity Analysis, edited by Ali Dogramaci and management and labor aspects, as they are discussed in Nabil R. Adam, both of Rutgers University ; this is the Buehler and Shetty, Hill, and the other chapters of second volume of the Studies in Productivity Analysis Rosow. In a concluding section, we sketch some alter- series. Two papers address methodology: that by native lines of research . Ephraim Sudit of Rutgers University and Nachum Fin- ger of Ben Gurion University presents a general survey The conceptual problems of data analysis and that by Douglas Moon of Columbia University, the The literature on the productivity problem shows lit- dynamic input/output model. The formidable problems tle patience with the troublesome theoretical problems posed by time-series analysis are discussed by Lawrence of economics. These books are no exception. Fabricant's Cohen of Columbia University and Salin Neftci of Bos- overview gives theory scant attention; Mark's discussion ton College. Tom Boucher of Cornell University of measurement problems includes an extensive survey assesses technical change; J. R. Norsworthy and Mi- of the reliability of our data, but from an exclusively chael Harper of the Bureau of Labor Statistics consider pragmatic point of view. Sudit's discussion of method- capital formation, and Frank Gollop of Boston College ological issues is broad-ranging but makes no effort to and Mark Roberts of Pennsylvania State University an- draw any conclusions concerning the value of the em- alyze imported intermediate inputs. pirical work that is founded on fragile hypotheses. Most The next theme, management, is examined in Produc- of his fellow contributors to the Dogramaci/Adam col- tivity: Prospects for Growth by John Donnelly, who lection, concentrating on empirical industry-level and discusses the role of the chief executive, and by Alfred time-series analysis, struggle with the practical difficul- Neal, former president of the Committee for Economic ties associated with these problems without the benefit Development, who analyzes the role of the tax system. of a viable theoretical framework. Not surprisingly, Exploring technological change at the corporate level such studies are principally of interest to the profession- are Reginald Jones, chairman of General Electric, John al student of productivity . Diebold, chairman of the Diebold Group, Thomas Two theoretical problems in particular would seem to Donahue, secretary-treasurer of the AFL-CIo, and Rob- merit discussion. Productivity analyses inspired by the ert Ranftl of Hughes Aircraft Corp. neoclassical paradigm attempt to quantify the contribu- The problems of productivity management are also tion of each factor of production to output growth. The the theme of papers edited by Vernon M. Buehler and productivity growth that cannot thus be explained, Y. Krishna Shetty, both of Utah State University, in called the residual, has been attributed to technical Productivity Improvement. Case Studies of Proven Prac- change . Growth is thus decomposed into movements tice. Represented are 11 companies and three unions . along a production function (representing a certain tech- Contributors also include Murray Weidenbaum, former nology), and shifts of the production function (indica- chairman of the Council of Economic Advisers, and tive of a change in technology). If this sounds plausible Clement Preiwisch of the General Accounting Office . for small, marginal changes, Nelson' has already drawn The last theme, dealing with labor, is discussed in attention to the absurdity of the attempt to extrapolate Productivity: Prospects for Growth from four perspec- the procedure to major changes such as we have tives. Rosow of the Work in America Institute surveys witnessed over the postwar period. the problems associated with the various "human fac- The second question warranting additional research tors" and discusses their possible remedies. Writing on takes us further back, into the great "Capital Debate" worker participation are Stephen Fuller, vice president between Cambridge (U.S.) and Cambridge (U.K.). The of General Motors, Douglas Fraser, president of the conclusion was that the neoclassical attempt to base a United Auto Workers, and Wayne Horvitz, director of theory of distribution on the theory of production was the Federal Mediation and Conciliation Service. fatally flawed: even under competitive equilibrium con- These problems are given more theoretical treatment ditions, the remuneration of capital is not determined in Stephen Hill's Competition and Control at Work: The by its marginal productivity, because the definition of a New Industrial Sociology, the fruit of his teaching at the quantity of capital presupposes determination of the London School of Economics. distributional variable . This conclusion vitiates much of The four books incorporate some of the most ad- the growth accounting exercise, because the calculation vanced thinking in this, somewhat fragmented, area of of a stock of plant and equipment-at first sight purely research . Our review will thus attempt to assess some of physical entities-involves a nontechnical factor like the strengths and weaknesses of the state of the art. The the rate of return . Multifactor productivity studies, first two sections will deal with conceptual and analyti- however, continue to calculate a stock of capital (or a cal problems and will therefore consider the contribu- flow of capital services) by virtue, as C. E. Ferguson put tions of the Rosow volume and Dogramaci and Adam it, of an "act of faith" : "The question that confronts us collection . The following two sections will cover the is not whether the [British] Cambridge Criticism is the-

16 oretically valid. It is. Rather the question is an empiri- quality range. In an extreme case, that of the computer cal or econometric one: is there sufficient substitutabili- equipment industry, the difficulty of the task of measur- ty within the system to establish neoclassical results? . . . ing quality change has led to total capitulation, and the Until the econometricians have the answer for us, plac- price deflator is conventionally set at 1, as if there had ing reliance upon neoclassical economic theory is a been no qualitative improvement at all since the birth of matter of faith."' It is somewhat disconcerting to find the computer industry . Some, not implausible, estimates the current productivity research pursued as if the Cam- of quality changes in this industry can be shown to bridge U.K. school had never existed. boost output measures so much that the productivity Some relief from these attacks on the very legitimacy lag for manufacturing disappears entirely .4 of growth accounting models may be forthcoming from The rapid development of the service sector aggra- the sophistication of more recent econometric tech- vates this problem . It is remarkable that as we narrow niques. Much of the capital debate concerned the circu- our focus from GNP, to private business sector output, larity of reasoning in the neoclassical theory, attacking and further to manufacturing output, the productivity its explanatory power, but perhaps not its descriptive slowdown appears progressively less dramatic . This power. To our knowledge, however, none of the parti- seems perhaps normal, when one contrasts automation sans of the growth accounting techniques has made this trends in manufacturing with those of service industries case. Most of the technical debates to date-for exam- like shoe-shining . But the image of a technically back- ple, those surrounding the replacement of Laspeyres ward service sector is belied by the example of comput- and Paasche indexes by Divisia indexes-are by com- erization in telecommunications, banking, and insur- parison of limited import . ance . The basic problem posed by such theoretical interro- Two hypotheses thus compete in explaining the dif- gations concerns the usefulness of the neoclassical ference between the roles of manufacturing and services paradigm for dynamic analysis in conditions of real- in the productivity slowdown . The first is that we world complexity. As Joan Robinson has written,' there mismeasure and underestimate service output ; pushed is something inherently wrongheaded in trying "to find far enough, this hypothesis could lead to the argument out from the record of what actually happened, what that there has been no serious productivity lag. Against growth of output would have been if the value of capital such skeptics, it can, however, be shown that in the had grown as much as it did without any technical manufacturing sector, too, and in particular in many in- progress having taken place." The value for long-term dustries where measurement problems are least impor- analysis of the distinction between shifts of and along a tant, there seems to have been a significant productivity production function seems at best extremely limited. slowdown . The second hypothesis reverses the perspec- The concrete problems of a choice of productivity in- tive, to emphasize the collapse of the service sector's ap- dicators are thus posed against a backdrop of vast theo- parent productivity . Could this reflect a real breakdown retical disputes; and the latter permeate the former. The usefulness of multifactor indexes, on the one hand, in attempting to define quantities of the different inputs, is limited by the need to assume that factors are remuner- Books reviewed ated at their marginal product. If this assumption is of dubious legitimacy for capital, the case of labor is not Jerome Rosow, ed., Productivity : Prospects for Growth . simple either. Obviously, different qualities of labor New York, Van Nostrand Reinhold, Work in Ameri- have different productive potentials ; but it is much less ca Series, 1981, 340 pp . $19.00. clear that relative pay reflects these differences. Ali Dogramaci and Nabil R. Adam, eds ., Aggregate and The use of simple labor productivity indexes, on the Industry-Level Productivity Analysis. (Volume 2 of other hand, is theoretically uncontroversial. But their Studies in Productivity Analysis.) Hingham, Mass ., use does little to reduce the productivity puzzle to its Martinus Nijhoff, 1981, 195 pp . $25 .00 purely quantitative dimension. The substitution of capi- Vernon M. Buehler and Y. Krishna Sherry, eds., Pro- tal for labor must be somehow incorporated into the ductivity Improvement: Case Studies of Proven Prac- analysis . Relying on labor productivity, therefore, sup- tice. New York, AMACOM, American Management poses the development of a model of accumulation, Associations, Inc., 1991, 273 pp. S19.95. which the neoclassicists thought they had provided . Beyond these properly economic disputes, there is Stephen Hill, Competition and Control at Work : The New Industrial Sociology. Cambridge, Mass ., Massa- also confusion over broader issues . chusetts Institute of Technology Press, 1981, 280 pp. Measures of output, including those of the Bureau of $25 .00 cloth, $9.95 paper. Labor Statistics, are often approximate, especially in the many industries with no clearly defined products or

17 MONTHLY LABOR REVIEW October 1982 e The Productivity Puzzle in the efficiency with which this sector performs its me- explaining the slowdown . The data are far from show- diating and informational functions? Unfortunately, lit- ing this; but, above all, one would want to ask: why tle research has been conducted on the industrial have the major economies proved themselves to be so dynamics of these functions.' incapable of surmounting such a handicap? The vigor of Deeper conceptual problems are not absent here ei- the upturns in GNP growth since 1973 has slowed recog- ther: how should we treat nonmarket goods? Pollution nition in this country that the long-term growth path control expenses are commonly included in the cost side has been shifted downward. of production, but are difficult to include in the output Under the title "Free the Fortune 500," Weidenbaum side as, for example, clean air. Do market prices bear presents the now-classic case for assigning the role of sufficient relation to utility to justify our reliance on chief culprit to government regulation . No statistics, them for evaluating economic performance? There is a and certainly not Denison's, have been advanced to venerable tradition of rejecting output (and therefore substantiate his claim. The text is a candidly ideological productivity) statistics as irrelevant to real welfare. The manifesto that gives the reader a glimpse into the mind rub, of course, is that even if the data reflect the specifi- set of the recently resigned head of the Council of Eco- cally market forms of welfare calculation, it is such cal- nomic Advisers. culations which orient real-world decisions. As limited The most serious candidate for blame is capital for- as these measures are, they therefore have a key role to mation-the object of a study by Norsworthy and play in analysis . Harper in Aggregate and Industry-Level Productivity The Rosow and Dogramaci/Adam volumes give Analysis. The proportion of GNP going to investment these problems but scant attention. has been remarkably stable over the last decade, but as GNP growth has slowed, so has capital formation. Other Looking for scapegoats data in this contribution indicate that the price of capi- Beyond the conceptual and measurement difficulties, tal services sharply accelerated from 1973, almost there has nevertheless probably been a fall in labor pro- reaching the rate of increase in hourly labor compensa- ductivity growth rates. This deceleration is sufficiently tion. The combination of higher interest rates, massive important in a large enough range of indicators, both increases in the labor force owing to the arrival of the aggregate and industry level, to overcome most skepti- baby-boom generation and to the "mobilization" of cism. Do we have an explanation for it? women, as well as more direct pressure on real wage In the aggregate data, the slowdown is particularly levels, may have thus led to such a cheapening of labor dramatic since 1973. In the total factor productivity relative to capital as to slow the substitution of the lat- framework, this shows up as a precipitous decline in the ter for the former.' The principal difficulty with these main factor contributing to growth, the residual . This explanations of the productivity slowdown is that the fact alone should be sufficient to show that Edward F. reduction in investment flows only marginally affects the Denison's interpretation of the residual as primarily re- "productivity" of the stock of capital. A further flecting advances in knowledge cannot be sustained .e hypothesis is explored by Alfred Neal in Productivity: Whatever slowdown one may imagine taking place in Prospects for Growth; he blames "excessive" taxation for research and development, the accumulation of knowl- insufficient investment. The argument is weakened by edge can hardly be imagined to have braked so the ubiquity of the slowdown in countries with widely suddenly . different taxation trends. Energy costs have also been A first hypothesis might be that companies today incriminated, their rise rendering redundant a certain treat labor as a quasi-fixed factor, and that therefore the fraction of the capital stock because of energy/equip- adjustment of employment to production is slower than ment complementarities. it used to be. This has been verified statistically, and Any or all of these factors may have played a role, many of the contributors to the Buehler/Shetty volume but a key lesson from John Maynard Keynes seems to claim that increased labor flexibility is the key to in- have been forgotten: the "animal spirits" of the investor creasing corporate profitability. While this may explain a will surmount many such obstacles if the weather fore- certain (downside) volatility of productivity ratios over casts for the business climate are good.' In particular, the shorter period, the question remains as to why the that somewhat tired old culprit, deficiency in savings, slowdown persists. cannot constitute a real brake in a modern economy in Indeed, the U.S. debate has been characterized by a which investment is financed on a credit-based, for- great resistance to the idea that the recent recessionary ward-contract system . If business prospects are good, trends could be other than purely cyclical or exogenous- low levels of retained corporate earnings will be supple- ly generated. Jorgenson,' in Productivity: Prospects for mented by extra external finance, and a lack of deposits Growth and elsewhere, develops the exogeneity thesis, in the lending institutions will be overcome by money- arguing for the importance of energy prices in creating credit .

18 The problem would thus appear to be systemic rather duction process. The Quality Circle view, in contrast to than localized. Any particular difficulty can be sur- the "quality of worklife" philosophy to which Rosow mounted, and, often, transformed into a stimulus. The and others allude, excludes any commitment to real co- search for the origin of, and the cure for, the productiv- operation in which the gains of labor would not be pre- ity "problem" has therefore recently turned to manage- mised on the prior increase of company profits. ment and labor, the major actors in a socioeconomic Some cracks do, nevertheless, appear in the manage- system, the dynamism of which may be faltering. ment orthodoxy . Nucor Corp . insists on the importance of job security and has implemented group bonus The role of management schemes that include foremen and maintenance crew . The link between productivity and management is Crompton Co., Inc., has instituted a 36-hour, 3-day difficult to establish because product change and mar- workweek paid 40 hours. Hughes Aircraft declares its keting flexibility are often more direct determinants of commitment to designing "meaningful" jobs by enlarg- commercial survival and success than the technical effi- ing the range of tasks. ciency with which a firm produces a hypothetically sta- The union contributions by Cass Alvin of the AFL- ble product. Accordingly, management itself tracks CIO echo somewhat alone in this landscape. The conser- profitability rather than the more abstract notion of vatism of his interlocutors would indeed seem to consti- productivity . tute a major handicap in putting the United States back The second part of Productivity: Prospects for Growth onto the map of innovative entrepreneurship. Aber- discusses a number of management problems related to nathy, Clarke, Hayes, and Kantrow'° have recently the productivity issue. The principal area of analysis is launched a major critique of this conservatism. They at- the dynamism of technological change in the firm. tribute the decline in the relative strength of U.S. com- Diebold sketches the (well known and still) fascinating panies to the short-term, bottom-line myopia of account of the Office of Tomorrow, with a refreshingly corporate decisionmaking . Overemphasis on quarterly pragmatic touch as to the limits both of the current and annual results, according to the Harvard authors, technology and above all of its impact on office-work cripples American corporations' capacity for long-term productivity . This contribution is valuable in reminding technological programming. Symptomatic of the disease us that the availability of new technologies does not is the U.S. managers' tendency, perfectly explicit in the guarantee their rapid implementation-the delays are case studies, and above all in the "Free the Fortune often measured in decades. Furthermore, implementa- 500" contribution, to interpret every constraint on their tion does not guarantee improvement of the standard prerogatives as an intolerable shackle on individual cre- productivity indexes, for new technologies create new ativity. Whence the paradox: in the United States, tasks. where Government intervention and unions are smaller Other contributions include a disappointingly low-key and weaker than in most other developed countries, the union assessment of technological trends by Donahue, blame attributed to Government and unions in causing somewhat in contrast with the more thought-provoking the current crisis is greatest. piece by Oswald, AFL-CIO research director, on the gen- The difficulty, of course, with this critique of manage- eral productivity question. The contribution of John ment, is that in less expert hands it can easily slide into Donnelly, the chairman of Donnelly Mirrors, Inc., is the same "blame the victim" mode that constitutes one useful in outlining one manager's perception of the im- of management's own shortcomings . Can one sustain portance of practical labor-management cooperation in the argument that the current economic woes of the the framework of a Scanlon Plan. United States are principally due to a particularly in- This latter approach to labor, seeking to transform competent group of managers? Is not their myopia the the presence of unions from a handicap into an advan- most rational programming strategy in a period of great tage for corporate competitiveness, is in sad contrast to uncertainty? Is it not the flip side of the flexibility of the approaches outlined in the case study volume operations that European managers so envy? Is not published by AMACOM (a division of American Manage- long-term technological programming vastly easier for ment Associations). The reader cannot but be impressed those in second place who are imitating the by the presence of such important companies as Kaiser frontrunner?" Aluminum and Chemicals, Hughes Aircraft, and Burger Alternatively, one could perhaps hypothesize that the King, even if the papers themselves are disappointingly cyclical characteristics of capitalist growth can be short and lacking in detail. The message is basically dampened in the shorter term but not eliminated . The that productivity demands more Taylorism, more con- problem is thus rephrased : in the current worldwide re- trol, more incentive pay schemes, and a small dose of cessionary climate the only way to limit the cost of the Japanese-style Quality Circles. The last are designed to market system's congenital myopia is by aligning short- capitalize on workers' intimate knowledge of the pro- and long-term prospects . Such a reconnection implies a

19 MONTHLY LABOR REVIEW October 1982 * The Productivity Puzzle stabilization of macroeconomic conditions . Because pain of distressing unemployment . (Milton Friedman's markets are in themselves unable to provide such stabil- identification of Capitalism and Freedom rests on ob- ity, capitalist growth seems to necessitate its imposition scuring the general constraint in order to vaunt the free- by non-market forces, via the further institutionalization dom of its particularity.) of social consensus and conflict-resolution mechanisms. This leads to an interesting if somewhat fragmented discussion of Taylorism that contrasts favorably with The role of labor what one often finds in the U.S. literature . Hill follows The frequency with which incentive pay schemes are much of the recent research which characterizes mentioned by the contributors to the case studies is per- Taylorism as an expression of this asymmetry in the la- haps not to be simply attributed to the blame-the-victim bor process: management control over the immediate la- syndrome. Assuring the cooperation of labor is a major bor process is gained at the expense of craft-type permanent task; poor labor relations can be very costly worker autonomy. But he tempers this account by a in terms of excess supervisory personnel, of under-per- discussion of the limits of Taylorism: its partial adop- formance of workers, of underutilization of plant, and tion in management circles, the resistance of workers to of lack of product quality and timeliness. If these costs its effects, and, most importantly, the fact that the pro- are less important than those associated with a deficit of duction process always necessitates some degree of co- technical and organizational adaptation, they are by no operation-even within the framework of conflict. means negligible . The conflictuality of labor-management relations is, Stephen Hill's book presents a valuable framework in this perspective, somewhat independent of the degree for the analysis of these problems . Written from an En- of institutionalization taken by the forms of its resolu- glish perspective, but with a solid grasp of U.S. devel- tion . By contrast, U.S. discussion of quality of worklife opments, its dual reference to Max Weber and to a programs seems hampered by the assumption that coop- context where class conflict is manifest could prove a erative and adversarial relations can and should be two tonic for a U.S. audience. Especially in the current peri- totally distinct modes of labor-management interaction. od when labor leaders have rediscovered the pertinence It is as if an overly consensual (and individualist) ideol- of a "class war" rhetoric . ogy blocked recognition by management and by unions U.S. industrial sociology has been dominated by a that plant-level conflict was healthy and that coopera- Durkheimian perspective which privileges the reproduc- tive moments within this conflictual relation were per- tion of a community of values. The absence of consen- fectly normal. Whence a fruitless polarization between sus thus constitutes the horizon of much social the cynics and the naive. thinking: conflict is ever present but always on the hori- The import of such research for the productivity zon, beyond theoretical grasp. This approach contrasts puzzle is considerable, for many discussants locate the with that of Weber, for whom the conflict of interests is root of productivity decline in shop-floor tensions. The the starting point of social analysis . value of Hill's work is to remind such "radicals"-who The fundamental hypothesis of Hill's work is that an- appear at all points of the political spectrum-that tagonistic interests compete within the firm. This con- growth in capitalist economies is not a zero-sum game. flict is not just over income distribution, but also over Workers' gains are not simply capitalists' losses, be- power, and in particular allocative power on the shop cause in the longer run such gains are one of the most floor (work rules, staffing patterns and levels, work in- potent stimuli to technical change and hence to produc- tensity, and so on). The fact that U.S. unions are seen tivity growth. Whether worker resistance plays this role as having de-emphasized allocative struggles in ex- depends on the dynamism of the system . change for concessions in income distribution should not, in Hill's view, be interpreted as implying that shop- The dynamism of socioeconomic systems floor conflicts can be relegated to the status of a prob- The productivity puzzle is a valuable indicator of the lem of maintaining consensus within the unions. The current state of economics, reflecting this discipline's basic separation of ownership or control and productive difficulties -heoretical, quantitative, historical, and so- activity-as opposed to their unity in a cooperative sys- ciological . Richard Nelson has drawn the uncomplimen- tem-makes competition and conflict primary, if not tary parallel with the drunk looking for his lost watch permanent, features of the capitalist firm. under the lamp post "because that's where the light is." Hill's Weberianism is not the diluted version to But why is the economics profession tipsy? Part of the which U.S. audiences are accustomed . Power within the reason may be its excessive focus on formulating policy capitalist firm is inexorably asymmetrical . The wage re- recommendations, an objective not always conducive to lation is a power relation, not just a "contract," because major theoretical research . the worker, while free not to enter this or that particu- The role played by this policy focus might, however, lar employment contract, must enter some contract on shift from debilitating to revivifying. The urgent need

20 for vigorous policy remedies to current economic prob- hierarchies. The coherence of these social forms with lems will not, we believe, be satisfied by a reliance on the dominant macroeconomic relations of productivity the automaticity of market adjustments . The demand and income growth-"deepening" or "widening" for serious policy may thus, indirectly, become a stimu- modes of accumulation-assures a harmonious balance lus for the revival of those theoretical trends that have in the expansion of output and demand; their incoher- for too long been relegated to the margins of economic ence generates a protracted, Kondratieff like period of theory : the heterodoxies of institutionalist and "funda- instability. mentalist" Keynesian theories. Periods of coherence naturally exhaust their dyna- The most fruitful areas of research may be at the inter- mism . Tensions accumulate. The diffusion of finite sets section of Joseph Schumpeter and Nelson, in its of organizational and technological innovations reaches proximity with that developed by certain French re- higher plateaus . Virtuous circles become vicious. No searchers along the lines suggested by Michel Aglietta .II meta-auctioneer guarantees the timely replacement of It would associate the analysis of macroeconomics to failing system-stabilizers . that of social institutions, going beyond the neoclassi- In such a perspective, the downward shift in growth cal, market-centered model by breaking with its implicit paths, of which the productivity deceleration is but a assumption that real developments, such as a productiv- symptom, is attributable neither to a single cause nor ity slowdown, can be accounted for by the juxtaposition the accidental conjunction of several causes. Longer of purely exogenous shocks and the spontaneous equili- downswings are part of our economic history, as the brating market mechanism. system exhausts and then recreates the social-structural Market mechanisms need to be integrated into a his- conditions of accumulation . torical model that explains their (always limited) Economic history, the real history of cycles, short pertinence to any given epoch. Periods of economic his- and long, of accumulation and crashes, is made in the tory are thus distinguished according to their money- interstices of "economics" as Academia currently imag- creation regimes, wage-setting institutions, price de- ines it. At least, such might be the lesson of the produc- termination mechanisms, and international trade tivity puzzle. 0

FOOTNOTES

Activity, 1, 1981) discusses this ' Richard R. Nelson, "Research on productivity growth and Labor" (Brookings Papers on Economic between market valuation productivity differences: dead ends and new departures," Journal of hypothesis in relation to Tobin's q ratio practical import of the Cam- Economic Literature, September 1981, pp . 1029-64. and replacement costs of capital. The bridge U.K . position may lie in the modeling of the tensions generat- C. E. Ferguson, The Neoclassical Theory of Production and Distri- ed by the real-world gap between the competing financial estimates of bution (London and New York, Cambridge U.P ., 1969). capital stocks approximated in q's numerator and denominator. 'Joan Robinson, Contributions to Modern Economics (New York, '° William J. Abernathy, Kim B. Clark, and Alan M . Kantrow, Academic Press, 1978). "The New Industrial Competition," Harvard Business Review, Septem- the Council of Eco- 'Unpublished paper by Michael J. McKee of ber-October 1981, pp . 68-81; and William J. Abernathy and Robert nomic Advisers staff. H. Hayes, "Managing Our Way to Economic Decline," Harvard Busi- ` But see Robert S. Cohen, The Internationalization of Capital and ness Review, July-August 1980, pp . 67-77. Research, U.S. Cities (Ph.D . dissertation, New School for Social "Robert Z. Lawrence, Phase 1 Report: International Trade, to the 1979), and Thomas Stanback Jr . and Thierry Noyelle, Services/the National Science Foundation. Essay number two: Trade performance New Economy (Montclair, Allanheld, Osmun, 1981). patterns, unpublished paper, (Washington, The Brookings Institution, 'Edward F. Denison, Accounting for Slower Growth (Washington, April 1982). The Brookings Institution, 1979). " See, in particular, Robert Boyer and Pascal Petit, "Employment 'Dale W. Jorgenson, interview, Challenge, November-December and Productivity in the EEC," Cambridge Journal of Economics, Vol. 1980, pp . 16-25. Note, however, that the rate of capital/labor substi- 5 No. 1, March 1981, pp . 47-58. Also see Robert Boyer, "Wage For- tution does not seem to have slowed in manufacturing. mation in Historical Perspective: the French Experience," Cambridge 1979, pp . 99-118 . 'Gregory Schmid, "Productivity and Reindustrialization : A Dis- Journal of Economics, Vol . 3, No . 2, June senting View," Challenge, January-February 1981, pp . 24-29. "Michel Aglietta, A Theory of Capitalist Regulation: the U.S. Expe- ' Martin Baily, "Productivity and the Services of Capital and rience (London, N.L.B ., 1979).