Why Do Firms Differ, and How Does it Matter? TORIo Richard R. Nelson Strategic Management Journal, Vol 12, Special Issue: Fundamental Research Issues in Strategy and Economics (Winter, 1991), 61-74 Stable url: http://links.jstor.org/sici?sici=0143-2095%028199124%02912%3c61%03awdfdah%o3e2.0.c0%03b2-i trategic Management Journal is currently published by John Wiley Sons Your use of the jStOR archive indicates your acceptance of JSTOR,'s Terms and Conditions of Use, available at http://www.jstor.org/about/terms.htmlJstOr'sTermsandConditionsofUseprovidesinpartthatunlessyou have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://wwwjstor.org/journals/jwiley.html Each copy of any part of a STOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission jStOR is an independent not-for-profit organization dedicated to creating and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact support @ jstor. org http://www」]stor.org Wed nov211:34:43200
Strategic Management Journal, VoL. 12, 61-74(1991) WHY DO FIRMS DIFFER, AND HOW DOES IT MATTER? RICHARDR. NELSON School of International and Public Affairs, Columbia University New York, U.S.A. for students of business management and strategy, firm differences are at the heart of their inquiry. This paper explores the reasons behind this stark difference in viewpoint It argues that economists really ought to recognize firm differences explicity INTRODUCTION position certainly has been influenced by the work of scholars of firm management who have This paper is concerned with the sources and persuasively documented significant differences nificance of interfirm differences, from the among firms in an industry in behavior and viewpoint of an economist. How might an performance, and proposed that these difference economist's perspective on this differ, say, from largely reflect different choices made by firms that of a student of business management? I However, because the interests of those authors would argue that the most important difference have differed from the interests of economists, is that economists tend to see firms as players in almost no attention has been paid to the industry a multi actor economic game, and their interest or economy wide implications of such different is in the game and its outcomes, rather than in choices. Thus while the management literature the particular play or performance of individual provides a start for my argument, there is much firms. That is, economists are interested that I need to build myself, in cooperation with how the automobile industry works, and its like thinking friends erformance in ns. and not It should be recognized that, in trying to General Motors or Toyota per se, but only insofar make a case for the economic significance of as the particularities of these firms influence the discretionary firm differences, I and my co industry more broadly. This perspective is quite arguers are fighting against a strong tide in different, it seems to me, than that of a student economics, particularly in theoretical economics, behavior and performance of individual firms in such differences. The argument in economics is their own righ not that firms are all alike; economists recognize My objective in this essay is to make a strong that computer firms differ from textile firms, and case for the economic significance, in the sense in both industries, German firms almost certainly above, of discretionary firm differences. My differ from Taiwanese firms. Rather, the position is that the differences arent discretionary, but rather reflect differences in the contexts in which Key words: Firms, innovation, evolution, competition firms operate: computer design and production
62 R.R. Nelson technology and the computer market differ from modes of organizing economic activity. Finally the situation in textiles. Factor prices and a repris availabilities and product markets in Germany differ from those in Taiwan. Thus. firms are forced to be different THE DIVERGENT LITERATURES ON The tendency to ignore discretionary firm 'COMPETITIVENESS differences in part reflects that economists are not interested in behavior and performance The differences in perspective can be seen clearly t the level of firms, but rather in broader in the divergent literatures concerned with what aggregates-industry or economy wide perform- now popularly is called the competitiveness ance, It reflects, as well, some strong theoretical issue- the recent weakness of American firms, views held by most main line economists about particularly vis-a-vis Japanese ones, in industries what economic activity is all about, and about where not so long ago U.S. firms were doing the role and nature of firms in economic activity. very well. There is a sharp split between studies My argument that discretionary firm differences that focus on the differences between American within an industry exist and do matter significantly and Japanese firms, and studies by economists is part and parcel of my broader argument that that are focused on more aggregated variables neoclassical economic theory is badly limited Made in America, a publication put out in the Let me flag here, for future elaboration, what summer of 1989 by the MIT ComI do and don' t mean by the term 'discretionary. Industrial Productivity, is a good example, and I do mean to imply a certain looseness of summary, of the former line of research. While constraints, both in the short and long run, that the staff of the Commission undertook consider- gives room so that firms that differ in certain able research on its own, the multifaceted important respects can be viable in the same diagnosis it presents is quite consistent with that economic environment. I do mean that to some presented in a number of prior studies concerned extent these differences are the result of different with why American firms have been losing out. strategies that are used to guide decision making American firms are hooked on old style mass at various levels in firms. On the other hand, I production methods, in an era where flexible do not mean that what a firm is and does is manufacturing has become a more effective mode under the tight control of high level decision of operation. Similarly, our hierarchical mode makers. And I certainly do not mean that what of organization and custom of specifying job makes a firm strong or weak at any time is well assignments narrowly, while perhaps appropriate understood. even within the firms themselves in an earlier era. now are sources of weakness although there well may be an articulated point Research and product design and development of view on this. more on these matters later stand too distant from manufacturing and pro The remainder of this essay is structured as duction engineering; thus it takes American follows. In the following section I shall flesh out companies much longer than the Japanese to my above remarks about the very significant go from conception to production, and our differences in perspective between scholars production costs and quality often are inferior trained or inclined to see discretionary firm level American firms are myopic, both in the sense of variables as important, and economists who see their failure to look at world rather than national firm differences as determined largely by more markets, and in the sense that time horizons are aggregative economic forces. Then I focus on the short. The latter partly has to do with the higl basic theoretical preconceptions of neoclassical cost of capital in the United States, but also with economic theory that lead to this position, and the way our managers think and the tools of hich make it very difficult to move any analysis they are taught in business schools distance from it. I follow with an exploration of Compared with the Japanese and Germans, our evolutionary economic theory which provides a blue collar work force comes to the work place very different view of what economic activity is poorly trained by the public education system all about and within which firm differences are This is compounded by a weakness of in-company central, and go on to consider the role of firm training and retraining programs. Together, this differences in the evolution of technology and puts American firms at a significant disadvantage
Why Do Firms Differ? 63 regarding labor skills. American firms are less diagnosis of lagging American productivity growth willing to cooperate with each other on matters rates, and the convergence of productivity and where cooperation would yield pay-off, partly living standards among the major industrial because of the attitudes of managers, but also nations. The focus of Productivity and American rtly because government looks on cooperation Leadership: The Long View(1989) is usually at with suspicion or hostility. More generally, the level of the national economy, and sometimes business and government seldom work together at the level of the sector or industry. The and often are at odd variables considered are national savings and Others might summarize the central arguments investment rates, investments in education, pro omewhat differently, but I believe the above cesses through which technology flows from oes represent fairly the kinds of propositions creators to followers, and the like. There is about firm differences made in the report. The scarcely a word about discretionary behavior at arguments are plausible and provocative, and the level of firms may provide important guidance to American It is strongly tempting, and I think right management, and for public policy headed, to propose that each of the studies has However, there are two important issues one described part of the elephant. The argument in can raise about the conclusions of the study. the MIT study, that many of the difficulties First, one can question the confidence one should American firms are having are self inflicted, is place in the causal connections asserted in studies quite persuasive. At the same time the econom like Made in America. Second, one also can ist,s proposition, that to a considerable extent question whether the variables treated there as firms are molded by the broader economic basic really are so, as contrasted with themselves conditions surrounding them, is compelling. What being determined by broader forces seems sorely needed is an analysis that sees both At this stage I want only to flag the former of these matters, in a coherent way ssue. However, there really is a big question While the authors of Made in America never about just what Japanese firms in the automobile quite got into serious analyses of environmental dustry, or the semiconductor industry, are variables, it does not seem difficult to augment doing that lies behind their evident stronger an analysis that starts at the firm level to consider performance, in various dimensions, than Amer- the environments that firms are in. Two new ican or European firms. Late this essay I books are exemplary in that they do shall focus on this uncertainty, and some of its this. Both recognize explicitly that national or implications environmental variables strongly influence firm For the present I want to focus on the latter strategy and structure, and that firms have question, because it gets sharply into view the considerable range of choice about these vari ontrast between analyses like Made in America, ables. Chandler's Scale and Scope (1990)describes and the standard views of economists about the in considerable depth how the different economi determinants of 'competitiveness'. There is some conditions, institutions, and cultures of the U. S discussion in Made in America of macro or Great Britain, and Germany, molded the nature national level variables, like the exchange rate, of the modern manufacturing firms that grew up the cost of capital, or more generally the system in these different countries in the first decade of corporate finance, the effectiveness of the of the twentieth century, and influenced the public education system, government policies, industries in which the nations developed special etc. However, this is not where the focus is. It strength. However, there is nothing deterministic is firm level variables that receive the top billing, about Chandler's description of how the environ and it is presumed that these are discretionary ment shapes firms and influences their perform to a considerable degree. In contrast, the ance inclination of economists is to focus on macro, Porter's The Competitive Advantage of Nations or environmental level variables, and to play (1990)presents a similar perspective in which down or ignore the role of firm discretion environmental influences matter greatly, but the The same year that Made in America was firms have a considerable range of freedom published, three economists, Baumol, Blackman, regarding whether, or just how, they will take and Wolff, published their interpretation and advantage of the opportunities the environment
4R. R. Nelson affords. Indeed both authors see the firms as to put forth a strong general theoretical challenge some extent molding their own environment as, to the effect that innovation ought to be the for example, in calling forth significant public center of economic analysis. But it is hard to investments in education in the U.S. and Ger- overestimate the degree to which economists many continue to see the central economic problem as Chandler is an historian by training. Porter's that of meeting preferences as well as possible formal training is in economics, but his career given resources, and prevailing technologies and has been at a Business School and his research institutions. This perspective implies a rather focus has been on management. It should be limited view of what firms are about recognized that the orientation of these authors Second, partly reflecting this general orien- to 'firms' is quite different than that in most of tation, but not the only possible formulation of economics. Indeed it is apparent that for both firms'decision processes consistent authors the center of attention is the firms, and economists became wedded to a theory of firm the central questions are how are they doing behavior that posited that firms face given and and 'what makes them strong or weak. They known choice sets(constrained for example by are drawn to wider economic mechanisms and available technologies) and have no difficulty in institutions in the search for answers to these choosing the action within those sets that is the questions. Now firm performance clearly is best for them, given their objectives(generally related to broader economic performance, but I assumed to be as much profit as possible). Thus have argued above they are not the same thing. the 'economic problemis basically about getting Since neither Chandler nor Porter presents a private incentives right, not about identifying the coherent statement of the economy wide problem, best things to be doing, which is assumed to be their analyses stop considerably short of providing no problem an answer that would satisfy economists to the The perspective on the economic problem and question of why do firms differ and how does the theory of firm behavior described above do matte not invite a careful inquiry into what goes on in firms. However, the tradition in economics of treating firms as'black boxes, was not inevitable FIRMS IN NEOCLASSICAL ECONOMIC either. The fact that until recently at least, this THEORY has been the norm deserves recognition in its own right To get at that question from an economists The overall result is a view that what firms do perspective, one needs to start with a broad is determined by the conditions they face, and Inderstanding of what economic activity is all (possibly) by certain unique attributes(say a about, and what constitutes good economic choice location, or a proprietary technology )they performance or poor. Neoclassical theory, which possess. Firms facing different markets will vides the current conventional wisdom on behave and perform differently, but if the market these matters for economists, militates against conditions were reversed so would be firm paying attention to firm differences as an behaviors. Where the theory admits product portant variable affecting economic perform- differentiation, different firms will produce differ ent products but, in the theoretical literature The first is the perception of what economic any firm can choose any niche. Thus there all about. Since the formulation of are firm differ general equilibrium theory almost a century ago, autonomous quality to them the focus has largely been on how well an The theoretical orientation in economics thus and technologies. This position is far from cretionary firm differences matter. Of course universal. Empirically oriented economists have economists studying empirical or policy questions been interested in things like technical change have a proclivity to wander away from the tethers and, recently, there has been a rash of work on of theory when the facts of the matter compel economic institutions and how and why these them to do so. Thus in doing industry studies, change over time. Schumpeter some time ago economists often have been forced to recognize
Why Do Firms Differ? 65 even highlight, firm differences, and differences them, or at least I do, with very little theoretical that matter. One cannot study the computer insight into why IBM is different, or Toyota, and industry sensitively without paying attention to so what le peculiarities of IBM. The recent history of There has been a certain amount of recent the automobile industry cannot be understood theoretical work by economists that looks inside without understanding Toyoto and G M. But as of firms, at their structure, and thus seems to the Baumol, Blackman, and Wolff book testifies, give promise of a theoretical window for a deeper the theoretical preconceptions shared by most look into why firms differ. The chapters by economists lead them to ignore firm differences, Holmstrom and Tirole, and by Williamson, report unless compelled to attend to them on such work. The questions explor Several recent developments in theoretical surveyed work include what determines, through economics would appear to be changing this make or buy decisions, the boundaries of a firm somewhat. Thus the same summer that Made how it is organized, the relative bargaining power in America, and Productivity and American of owners, managers, and workers, etc, But Leadership were published, the long awaited again, the ultimate reason for why firms differ also. Included in the chapters were several that some chance event, or some initial condito Handbook of Industrial Organization(1989)was is rather superficial. Implicitly they differ becaus survey theoretical work that does recognize firm made different choices profitable differences In my view, recent theoretical developments There are, first of all, the essays by Ordover in neoclassical theory have loosened two of the and Saloner, and by Gilbert, which are expressly theoretical constraints making it difficult if not concerned with theoretical that aims to impossible to see firm differences as important explain firm differences, or at least some conse- Economists are getting away from the theoretical quences of firm differences. In the models tethers of static general equilibrium theory and reported, there usually is an incumbent in the are treating technology as a variable not a given industry, or in the production of a particular And they are trying to look inside the black box oroduct, who has certain advantages over firms of the firm. However, for the most part there who might think of joining the action. The has been failure to get away from the third presence of these advantages, or threats of action tether--taking a firms choice sets as obvious to should a newcomer try to encroach, is enough it and the best choice similarly clear and obvious to make the advantages durable. Gilbert deals And because of that, the reasons for firm more generally with models where there are costs differences, in technology or organization, are to firms of changing their market positions. ultimately driven back to differences in initial However, with few exceptions the models sur- conditions, or to the luck of a draw, which may veyed in these chapters do not consider in much make choice sets different. Given the same depth or detail original sources of firm differences. conditions, all firms will do the same thing Reinganum's chapter, which surveys modern As I indicated above, I certainly do not want neoclassical models of technological innovation, to play down the role of environment in is focused on what certainly is an important constraining and molding what firms do. And I source of such differences--industrial R&D and do not want to play down the role of chance in e innovation R&D makes possible. In the causing large and durable subsequent differences models she surveys, a firms technology may among firms. But in my view the models most differ from a rival's because of the luck of an economists keep playing with do not effectively R&d draw, with the advantages made durable come to grips with what lies behind the firm by patent protection or subsequent learning curve differences highlighted in Made in America, or advantages. Given an initial difference, firms the implications of those differences may face different incentives and thus find The reason, I want to argue, is that while the different courses of action most profitable. surveyed work purports to be concerned with However, while these models may rationalize innovation,, with the introduction of something the observation that firms possess different new to the economy in the form of new technology technologies, the answers as to why certainly or a new way of organizing a firm, the models aren't very deep. And one comes away from in question completely miss what is involved
66RR. Nelson in innovation. Thus nowhere in the models moved to adapt a very different view of the Reinganum describes is the fundamental uncer- economic problem. Within this view, which I tainty, the differences of opinion, the differences will call evolutionary, firm differences play an in perceptions about the feasible paths, that tend essential role to stand out in any detailed study of technical advance, even recognized, much less analyzed in any detail. Williamsons own work on the INNOVATION AND FIRMS IN determinants of firm organization has been much EVOLUTIONARY THEORY influenced by Chandler, and he dedicates a certain space in his chapter to a transactions cost The models of technological innovation surveyed interpretation of Chandler's account of the rise by Reinganum show economists interested in the of the modern corporation, But nowhere does theory of the firm struggling to break away from he recognize explicitly the halting, trial and the orientation of general librium theory, feedback, often reactive rather than thought- which sees the economic problem as allocating through, process that led to the new ways of resources efficiently, given technologies. So too o put compactly, the treatment of technological Here economists seem to be basically interested organizing that Chandler describes the new literature on organizational innovation and organizational"innovation described in these in how new ways of doing things--technologies chapters simply takes the given 'choice set and and ways of organizing and governing work-are maximizing over it' presumptions of standard introduced, winnowed, and where proven useful neoclassical theory and applies them to inno- spread, as contrasted with how familiar technol vation. That is, innovation is treated as basically ogies and organizational modes are employed like any other choice, Investment costs may need Many years ago Schumpeter insisted that the to be incurred before the new product or focus of general equilibrium theory was on organizational design is ready to be employed, questions that, over the long run, were of minor but in neoclassical theory this is true of other importance compared with the question of how apital goods like a bridge or a machine. There Capitalist economies develop, screen, and selec- may be high risks involved in doing something tively adopt new and better ways of doing things new, in a formal sense of that term, but this is Many of the writers surveyed by Reinganum call treated as statistical uncertainty with the correct themselves 'neo Schumpeterian probability distribution known to all as is standard However, the dynamic processes Schumpeter in micro economic theory. TI InnovatIo n may described are not captured by the new neoclassical yield a new latent or manifest public good, and models. As he put it ' in dealing with Capitalism this raises theoretical problems of 'market failure, you are dealing with an evolutionary process but this is no different than investment in, say, He clearly had in mind a context in which people and organizations, had quite different views about ut what if effective treatment of innovation what kinds of innovations would be possible, and (and perhaps other activities) requires breaking desirable, and would lay their bets differently away from the assumptions of clear and obvious There are winners and losers in Schumpeter's choice sets and correct understanding of the process of creative destruction, and these are consequences of making various choices? Does not determined mainly in ex-ante calculation, but it really make sense to work with a model that largely in ex-post actual contest presumes that the transistor, or the M form of In his 1911 Theory of Economic Development organization, were always possible choices out Schumpeter saw the key innovative actors as there and known to all relevant parties, and that 'entrepreneurs. His 'firms' were basically the hey simply were chosen and thus came into vessels used by entrepreneurs, and other decision existence and use when conditions made profit- makers forced to adapt to the changes wrougl able the relevant investments? Does the assump- by entrepreneurial innovators or to go under. By tion that 'actors maximize, help one to analyze the time(1942) he wrote Capitalism, Socialism, situations where some actors are not even aware and Democracy, Schumpeter's view of the sources of a possibility being considered by others? of innovation had changed, or rather it might be If one reflects on these issues, one may be better to say that there had been a transformation
Why Do Firms Differ? 67 of the principal sources of innovation from an The concept of strategy in this theory of the earlier era, and Schumpeter's views reflected this firm is basically what business historians and transformation. Modern firms, equipped with scholars of management mean, as contrasted with research and development laboratories, became game theorists. It connotes a set of broad the central innovative actors in Schumpeter's commitments made by a firm that define and heory. The chapter by Cohen and Levin in the rationalize its objectives and how it intends Handbook admirably surveys the wide range of pursue them. Some of this may be written down empirical research that has been inspired by ome may not be but is in the management humpeter, particularly the research concerned culture of the firm. Many economists would be ith the relationships among innovation, firm size wont to propose that the strategy represents and other characteristics, and market structure. firms solution of its profit maximization problem In our book, An Evolutionary Theory of but this seems misconceived to me. In the first Economic Change(1982), Winter and I spent place, the commitments contained in a strategy quite a bit of space presenting a"theory of the often are as much a matter of faith of top firm' which is consistent with, and motivates, a management, and company tradition, as they are Schumpeterian or evolutionary theoretic view of of calculation. Second, firm strategies seldom economic process and economic change. Our determine the details of firm actions, but usually formulation drew significantly on Simon(1957), at most the broad contours. Third, and of vital on Cyert and March (1963), and on Penrose importance, there is no reason to argue a priori (1959), as well as on Schumpeter. With the vision that these commitments are in fact optimal or of hindsight, it is clear that our writing then was even not self destructive. If it is proposed handicapped by insufficient study of the writings that competition and selection force surviving of Chandler, particularly his Scale and Scope strategies to be relatively profitable, this should (1966) be a theorem not an assumpti Since the time we wrote, there have been a The concept of firm structure in this literature number of theoretical papers on firm capabilities also is in the spirit of Chandler, as is the and behavior that draw both on Chandler presumption that strategy tends to define a and on our early formulation, and which add desired firm structure in a general way, but not significantly to the picture. Papers by Teece the details. Structure involves how a firm is ( 1980, 1982), Rumelt(1984), Cohen and Levin- organized and governed, and how decisions thal(1989), Dosi, Teece and Winter (1989), actually are made and carried out, and thus Prahalad and Hamel (1990), Pavitt(1987, 1990), largely determines what it actually does, given Cantwell (1989, 1990), Kogut(1987), Henderson the broad strategy. A firm whose strategy calls (1990), Burgelman and Rosenbloom (1989), for being a technological leader that does not Langlois (1991), and Lazonick(1990), all present have a sizeable R&D operation, or whose R&d a similar or at least a conformable theoretical director has little input into firm decision making. view, although with differences in stress. The clearly has a structure out of tune with its recent paper by Teece, Pisano, and Shuen(1990) strategy. However, the high level strategy may provides an overview of many of these works, be mute about links between its R&d lab and and I believe correctly states that the common universities, whether to have a special biotech element is a focus on firm specific dynamic group, etc capabilities Change in strategy may require a change in e This emerging theory of dynamic firm capabili- management as well as a change in articulation ies can be presented in different ways. Here it indeed for the latter to be serious may require is convenient to focus on three different if the former. However, within this theory of the strongly related features of a firm that must be firm structure is far more difficult to change recognized if one is to describe it adequately: its effectively than is strategy. While changing formal strategy, its structure, and its core capabilities. organization, or at least the organization chart While each has a certain malleability, major is easy, and selloffs and buyups are possible changes in at least the latter two involve significantly changing the way a firm actually onsiderable cost. Thus they define a relatively goes about making operating level decisions and stable firm character carries them out is time consuming and costly to
68 R.R. Nelson do. Or rather, while it may not be too difficult of the decision making processes, the links to destroy an old structure or its effectiveness, between R&D and production and marketing it is a major task to get a new structure in shape etc. This means that at any time there will be and operating smoothly. Thus to the extent that certain kinds of r&d projects that a firm can a major change in strategy calls for a major carry out with some confidence and success and change in structure, effecting the needed changes a wide range of other projects that, while other ay take a long tim firms might to able to do them, this particular The reason for changing structure, of course, firm can not, with any real confidence to change, possibly to augment, the things a R&d capabilities may be the lead ones in firm is capable of doing well. Which brings the defining the dynamic capabilities of a firm discussion to the concept of core capabilities. However, in a well tuned firm, its production Strategy and structure call forth and mold procurement, marketing and legal organizations organizational capabilities, but what an organiz. must have built into them the capabilities to ation can do well has something of a life of its support and complement the new product and process technologies emanating from R&D. In Winter and I have proposed that well working Teece's terms, the firms capabilities must include firms can be understood in terms of a hierarchy control over or access to the complementary of practiced organizational routines, which define assets and activities needed to enable it to profit lower order organizational skills, and how these from innovation. And in an environment of are coordinated, and higher order decision Schumpeterian competition, this means the capa procedures for choosing what is to be done at bility to innovate, and to make that innovation lower levels. The notion of a hierarchy of profitable, again and again organizational routines is the key building block The concept of organizational capabilities, and under our concept of core organizational capabili- the theory that winter and I proposed as to what ties. At any time the practiced routines that are determines and limits them, does not directly built into an organization define a set of things imply any coherency to the set of things a firm the organization is capable of doing confidently. can do. However, Dosi et al.(1989)argue that If the lower order routines are not there for in effective firms, there is a certain coherency doing various tasks, or if they are but there is There would appear to be several reasons. The no practiced higher order routine for invoking ones stressed by Dosi et al. basically are associated them in the particular combination needed to with localized learning in a dynamic context, and accomplish a particular job, then the capability follow on the arguments that Winter and I made to do that job lies outside the organization,s some time ago that, to be under control,a extant core capabilities routine needs to be practiced. Firms need to The developing theory of dynamic firm capabili- learn to get good at certain kinds of innovation ties I am discussing here starts from the premise and at the things needed to take advantage of that, in the industries of interest to the authors, these, and this requires concentration or at least firms are in a Schumpeterean or evolutionary coherency, rather than random spreading of context. Simply producing a given set of products efforts. Further, in many technologies one innd with a given set of processes well will not enable vation points more or less directly to a set of a firm to survive for long. To be successful for following ones, and the learning and complemen any length of time a firm must innovate. The tary strengths developed in the former effort capabilities on which this group of scholars focus provide a base for the next round abilities for innovation and to take but i think it also is the case that to be economic advantage of innovation effective a firm needs a reasonably coherent In industries where technological innovation is strategy, that defines and legitimatize, at least important, a firm needs a set of core capabilities loosely, the way the firm in R&D. These capabilities will be defined governed, enables it to see organizational gaps and constrained by the skills, experience, and or anomalies given the strategy, and sets the knowledge of the personnel in the r&d depart- ground for bargaining about the resource needs firm procedures for forming new ones, the character its next step forward. Absent a reasonably
Why Do Firms Differ? 69 oherent and accepted strategy, decision making THE EVOLUTION OF TECHNOLOGY about rival claims on resources has no legitimate basis. Decisions from above have no supportive In real capitalist economies, in contrast with the rationale, and there is no way to hold back log neoclassical models, technical advance proceeds olling bargaining among claimants other than through an evolutionary process, with arbitrary high level decisions. There is no real products and processes competing with guidance regarding the capabilities a firm needs other and with prevailing technology in real time, to protect, enhance, or add in order to be rather than solely in ex-ante calculation. Some fective the next round of innovative of the innovations will be winners, other losers With the vision of hindsight the whole process But I think I simply am restating what looks messy and wasteful, and a more coherent Chandler, Lazonick, Williamson, and other schol- planning approach to technological advance ars of the modern corporation, have been saying appears attractive requires that firms innovate and change, a firm misguided efforts to plan and control significand for some time. To be successful in a world that However, it is striking how inefficient and must have a coherent strategy that enables it to technical advance have been. Where, for one decide what new ventures to go into and what reason or another, society has been denied the to stay out of. And it needs a structure, in the advantages of multiple independent approaches sense of mode of organization and governance, to advance technology, which flows naturally that guides and supports the building and from a basis of independent rivalrous firms sustaining of the core capabilities needed to carry almost always the approach chosen has turned out that strategy effectivel out. after the fact, to have major limitations If one thinks within the frame of evolutionary And since alternatives had not been developed theory, it is nonsense to pre that a fir calculate an actual'best'strategy. A basic premise comparison, there has been lock in. A number of evolutionary theory is that the world is too of U.S. military R&D efforts since 1960 are complicated for a firm to comprehend, in the striking examples. Nuclear power programs are irm understands its world in another. The fact is that in virtually every field neoclassical theory. There are certain character- where we have had rapid technical advance that istics of a firms strategy, and of its associated has met a market test or its equivalent, we have structure, that management can have confidence had multiple rivalrous sources of new technolog will enhance the chances that it will develop the While Winter and I formally modelled company apabilities it needs to succeed. There are other r&d programs as generating results through a characteristics that seem a prescription for failure. random draw, in fact in the industries that I However, there is a lot of room in between know well there has tended to be a certain where a firm(or its management)simply has to consistency in the r&D efforts of particular lay its bets knowing that it does not know how compa This consistency reflects a basically they will turn out stable company 'strategy, and the core R&D Thus diversity of firms is just what one would and other dynamic capabilities it has expect under evolutionary theory. It is virtually to carry it out. Where company strategies and inevitable that firms will choose somewhat associated capabilities differ significantly, their different strategies. These, in turn, will lead to patterns of innovation are likely to differ signif firms having different structures and different core cantly as well pabilities, including their r&D capabilities. This has an important consequence often Inevitably firms will pursue somewhat different overlooked in the literature on technological paths. Some will prove profitable, given what imitation. When one firm comes up with a other firms are doing and the way markets successful innovation, its competitors may differ evolve, others not. Firms that systematically lose significantly among themselves in their ability money will have to change their strategy and effectively to imitate or develop something structure and develop new core capabilities, or comparable. Contrary to many economic models operate the ones they have more effectively, or effective technological imitation very often drop out of the contest requires the imitating firm to go through many