Competition, Cooperation, And Innovation

Journal

of Economic

Behavior

Competition,

and Organization

18 (1992)

cooperation,

Organizational arrangements technological progress

l-25.

North-Holland

and innovation

for. regimes

of rapid

David J. Teece* Walter A. Haas School oJ Business, University of California USA Received

September

1989, final version

received

May

at Berkeley, Berkeley, CA 94720,

1990

Discussions of the link between firm size and innovation are outmoded because the boundaries of the firm have become fuzzy in recent decades. Strategic alliances - constellations of bilateral agreements among lirms - are increasingly necessary to support innovative activities. Such alliances can facilitate complex coordination beyond what the price system can accomplish, while avoiding the dysfunctional properties sometimes associated with hierarchy. Antitrust law and competition policy need to recognize that these new organizational forms are often the functional antithesis of cartels, though they may have certain structural similarities. A more complete understanding of bilateral contracts and agreements ought to reveal when and how cooperation can support rather than impede innovation and competition.

1. Introduction Competition is essential to the innovation process and to capitalist economic development more generally. But so is cooperation. The challenge to policy analysts and to managers is to find the right balance of competition and cooperation, and the appropriate institutional structures within which competition and cooperation ought to take place. Unfortunately, the economics textbooks tell us virtually nothing about these issues. While there is usually some consideration given to the import of monopoly and competition on incentives to innovate, it is almost always implicitly assumed that the price mechanism can effect whatever coordination the economic system requires for rapid technological progress. Typically there is no discussion of how interlirm agreements, vertical and horizontal, can help the process. If interfirm agreements are discussed, it is almost *I am grateful to Thomas Jorde and two anonymous referees for their helpful insights. I am also indebted to Michael Gerlach for assistance in understanding the differences in industrial organization between the United States and Japan. Patrizia Zagnoli and Oliver Williamson provided insights into the theory and practice of strategic alliances. An earlier version of this paper was published in Japanese in Business Review, Hitotsubashi University, March 1989. 0167-2681/92/$05.00

0

1992-Elsevier

Science Publishers

B.V. All rights reserved

2

D.J. Teece, Competition,

cooperation,

and innovation

always in the context of cartel theory. It is not surprising, therefore, that the economics textbooks, at least those used in the United States, do not convey a sense that inter-firm cooperation is either desirable or a subject worthy of study. This is despite the fact that economists have always recognized the central importance of technological innovation to economic growth and welfare, and to capitalism. Adam Smith’s Weulr~ of Nations plunges immediately into discussions of ‘improvements in machinery,’ and Karl Marx’s model of the capitalist economy ascribes a central role to technological innovation in capital goods. Nor did Alfred Marshall hesitate to describe knowledge as the chief engine of progress in the economy. Paul Samuelson, in his principals text, has always acknowledged the importance of technological change, but then proceeded like all the other leading texts to largely ignore it, causing Stiglitz (1987) in the ~roo~~~gs Papers to recently lament that ‘while it is the dynamic properties of capitalism . . . that constitute the basis of our confidence in its superiority to other forms of economic organization, the theory - at least the version we teach our students - is based on a model that assumes an unchanging technology.’ When technology is taken into account, the profession at large, according to Rosenberg (1982), has treated it as events transpiring inside a black box, and has ‘adhered rather strictly to a self-imposed ordinance not to inquire too seriously into what transpires inside that box.’ Therefore, it is not surprising that the economics profession is relatively silent on matters of innovation policy and technology strategy, and on matters of economic organization and innovation more generally.’ A large but unsatisfactory literature exists in industrial organization on the relationship between market structure and innovation, and between firm size and innovation, but both the theoretical and empirical literature are almost completely silent on inter-firm and intrafirm organizational issues.’ When these issues have been addressed, it is without much of a theoretical foundation. Analysts and policy makers have tended to stress the values of pluralism and rivalry as the best organizational arrangement to promote innovation. While these are important values, they are not the only ones. Moreover, with increased global competition, such values are now adequa-

‘In a recent paper, Dasgupta (1988) set out ‘to study the sorts of social institutions which can, at least in principle, sustain an efticient level of inventive and innovative activity’ (p. 2). This led him to consider what he called the Samuelsonian contrivance, Pigovian subsidies, and Lindahl property rights. The operationality of these mechanisms is remarkably obscure. Dasgupta’s gallant efforts notwithstanding, it would appear that there is ample room in the literature for new approaches. More compelling treatments of other relevant organizational issues can be found in the economic and business history literature. See, for instance, Mowery and Rosenberg (1989). ‘For an excellent survey, see F.M. Scherer (1970).

D.J. Teece, Competition,

cooperation,

and innovation

3

tely represented in the global system, and in national economies open to international trade and investment. This paper does not question the value of rivalry and pluralism; rather, it asks whether in the United States such values ought to be balanced at the margin by a recognition of the importance of coordination, accomplished via non-market forms of cooperation, to a well-functioning national system of innovation. In what follows, the relationship between technical innovation and aspects of industrial organization other than market structure and firm in section 3, the role of cooperative size are examined. In particular, agreements (national and international) among firms are examined for their impact on innovation. Implications for business organization, for corporate strategy, and for public policy are derived. The basic conclusion of the paper is that complex forms of cooperation are usually necessary to promote competition, particularly when industries are fragmented. Very few firms can successfully ‘go it alone’ any more. Cooperation in turn frequently requires interfirm agreements and alliances. In this regard, the Japanese form of industrial organization, with complex interfirm relationships, may have distinct advantages. European and American firms are now only beginning to learn how to effectively cooperate in order to compete. Antitrust authorities in Europe appear increasingly willing to let them try;3 unfortunately, antitrust policy in the United States remains uncertain and ambiguous.4

2. Innovation and competition Ever since Schumpeter (1942, p. 106) advanced the hypothesis that ‘the large scale establishment or unit of control . . . has come to be the most powerful engine of . . . progress,’ and that ‘perfect competition looks inferior, and has no title to being set up as a model of ideal efficiency’, economists have been ambivalent as to what role market structure and competition play in the innovation process. This ambivalence is particularly pronounced with respect to theoretical discussions. Tensions in the theoretical literature are very briefly noted below. It is also contended that Schumpeter and others did not frame the debate very well for today’s circumstances, as the unit of analysis Schumpeter used - the firm - is an increasingly fuzzy concept. One reason is that cooperative interfirm agreements are so pervasive among firms 3European companies, and particularly German companies, have a long tradition of cooperation in innovation from the beginning of the development of modern science-based industries in the early years of the 20th century. Europe had no antitrust tradition, not even the common law tenets concerning combinations in restraint of trade. Antitrust came only after World War II. Accordingly, Europe has had little difficulty in creating close interfirm cooperative agreements. 4For a discussion of the antitrust treatment of cooperative agreements in Japan, Europe, and the United States, see Jorde and Teece (1989, 1990).

4

D.J. Teece, Competition, cooperation, and innovation

in advanced industrial countries that firm boundaries are often very difficult to define in a meaningful way. Accordingly, section 3 will discuss the coordination requirements of innovation and outline the various organizational alternatives by which they can be achieved. Implications for corporate strategy and public policy are then derived. Schumpeter’s claim that large firms were necessary to promote innovation has fostered exploration of the links between innovative performance and market structure. Schumpeter linked firm size and innovation for three distinct reasons. First, he contended that only large firms could afford the cost of R&D programs. Second, large, diversified firms could absorb failures by innovating across broad broad technological fronts. Third, firms needed some element of market control to reap the rewards of innovation. The Schumpeterian legacy has spurred discussion of the link between firm size and innovation, and between market structure and innovation. Section 2.1 suggests that this discussion is inconclusive; section 2.2 indicates that it is also outmoded. Schumpeter did not frame the firm size hypothesis in a way that is particularly relevant today, as the boundaries of the firm can no longer be assessed independent of the cooperative relationships which particular innovating firms may have forged. Indeed, this paper posits that the firm’s ability to forge cooperative relationships can in many instances substitute for more comprehensive forms of integration. Put differently, in some circumstances cooperative agreements can enable smaller firms to emulate many of the functional aspects of large integrated enterprises, without suffering possible dysfunctions sometimes associated with large size. These points are now briefly explored.

2.1. Firm size, market

concentration

and innovation

As a theoretical matter, economic models of the innovation process are decidedly non-robust, showing that competition can lead to too much or too little R&D investment.5 In fragmented market structures, several kinds of market failures are commonly recognized. First, if firms are unable to exclude other firms from using technology they have developed, there is the classic ‘free rider’ problem. Even if patents prevent direct mimicking, there is likely to be a technological ‘neighborhood’ illuminated by the innovation that is not foreclosed by the patent, so the externality problem remains, though in a different form. The second problem is what is sometimes referred to as ‘overbidding’. This arises in the early development phase as competitors are stimulated to invest for the potential rewards that will go to the patent winner. Incentives to be the first to invent, or to be first to get to the patent oflice, may induce too many firms to try to invent early. In such a ‘For

a useful survey,

see Baldwin

and Scott (1987)

D.J. Teece, Competjt~on, cooperatjon, and innovarian

S

competitive race, too many resources may get applied too early. One consequence may be that firms drop out of the industry after the patent race is over but before the serious development work begins. A misallocation of resources is the unfortunate consequence. A monopolized industry avoids both the free rider and the patent race problem. Moreover, the knowledge externalities that come from successful exploration of uncharted technological areas are internalized. Economies of scale and scope may also be available. But the cost is the output restriction that the monopolist will supposedly engineer. This in turn cuts R&D investments. The results, as Nelson and Winter (1982, p. 289) explain, is that ‘it is hard to say whether there would be more or less R&D undertaken in the monopolized case than in the competitive case’. However, Nelson and Winter surmise that the balance probably tips against the monopolist, because in a centralized R&D regime which monopoly would entail, the diversity of approaches characteristic of competition would probably fall, as managers adopt simplified decision making styles, In short, the theoretical literature identifies a wide range of possible outcomes, and accordingly provides little guide to policy.

As a practical matter, the economists’ debate seems highly stylized and out of touch with global realities. Neither perfect competition nor complete monopoly is observable or realistically attainable in any industry today. Moreover, to abstract from industrial structure in such narrow terms is to miss key elements of industrial organization. First, the boundaries of the firm are extremely difficult to delineate, particularly when there are complex alliance structures in place. In Japan, for instance, Gerlach (1988) notes that alliances are ‘neither formal organizations with clearly defined, hierarchical structures, nor impersonal, decentralized markets. Business alliances operate instead in extended networks of relationships between companies, organized around identifiable groups, and bound together in durable relationships which are based on long-term reciprocity’. Much of the complexity and subtlety of these arrangements is missed by analysts who narrowly focus on legal titles and the structure of property rights, narrowly defined. Even worse, since the most familiar model that economists have of cooperative behavior is that of the cartel, many scholars have missed the efficiency inducing characteristics of alliance structures in Japan and elsewhere. Indeed, Caves and Uekusa (1976, p. 158) appear to push alliances into the straightjacket of cartel theory when they contend that ‘Japanese industries are prone to collusive arrangements but are also rich in structural incentives for the conspirators to cheat. The net outcome for the gap between actual and ideal industrial output is thus hard to predict, but

6

D.J. Teem, Competition, cooperation, and innovation

we do expect increased distortions from sporadic price discrimination, reciprocity, and the diversion of rivalry into uncontrolled nonprice forms’. Alliances are a form of contractual relationship, but they are much more than a legal contract. Alliance partners are not faceless entities operating at arms-length in Walrasian markets. Rather, exchange among alliance partners (at least in Japan) can be characterized, according to Gerlach (1988) by: (1) an emphasis on relational rather than transactional exchange, wherein the focus is toward the actors participating in rather than being the objects of exchange; (2) a state of continuous indebtedness and mutual obligation between the parties; and (3) the implicit negotiations of a socially significant order through symbolic activities and ceremonies. Alliances are defined by interests among specific subsets of firms that are familiar with each other through historical association, rather than immediate task requirements. The alliance structure in Japan appears to have enabled smaller Japanese firms to attain the degree of functional integration found within some large integrated American enterprises, but with less in the way of common ownership of the parts. Subcontractors, trading companies, and distributors take on for the firm functions often done internally in the United States. While many American firms, particularly those in high technology industries, rely on outside suppliers and distributors to a similar degree, the level of functional integration attained with these relationships in the United States is markedly lower. In Japan, the appropriate Western image is perhaps neither that of the ‘visible hand’ nor the ‘invisible hand’ but of the continuous handshake.6 Japaneses corporate strategy is therefore formulated to a considerable extent by maneuvering within alliance relationships, by industrial visions worked out among firms, and also by the broader industrial policies set by government. Where the law permits, alliance and consortia structures are increasingly coming to characterize aspects of European and American business. Though more transactional and less relational than alliances in Japan, these structures nevertheless involve high levels of cooperation, particularly for new product development and commercialization. So pervasive have these relationships become that one can no longer fruitfully explore industrial organization questions in the neoclassical tradition. Discussions of firm size which do not take cooperative agreements into account seriously mischaracterize the nature of the business firm. Discussions of product market structure that do not take the market for know-how into account are also likely to mischaracterize the nature of competition faced by a firm in a fastpaced industry. In the remainder of this paper, cooperative agreements are explored more fully. The next section discusses both the importance of

‘See Gerlach (1992, p. 2).

D.J. Teece, Competition,

coordination for the innovation available to effect it.

3. Innovation,

coordination

cooperation,

process

7

and innovation

and the organizational

mechanisms

and cooperation

The contention advanced in this paper is that the intensely competitive environment in which high-tech firms find themselves, coupled with the global dispersion of productive-technical competence, often requires complex forms of cooperation among competing firms. This paper stresses the need for both ‘operational’ and ‘strategic’ coordination to develop and profitably commercialize new technology. 3.1. The global dispersion of competence Global industrial competence became widely dispersed within a century of the industrial revolution. The predominant method of organizing these competences was first the factory. As Chandler (1977) so ably describes, the transportation and telecommunications revolution facilitated the rise of the modern (vertically) integrated corporation. The vertically integrated enterprise, beginning around 1900, internalized R&D activities [Mowery and Rosenberg (1989)]. U.S. industrial R&D emerged a private-sector affair, with R&D closely tied to the large enterprise through in-house labs. By 1945, the United States had an industrial structure dominated by large industrial enterprises, with a research system appended to it. American firms became dominant across a broad front of technologies. The decades of the 1970s and 1980s saw the manifestation of three trends, two of which have been in progress since 1945: (1) a catching up of Europe and Japan with the United States, and the greater dispersion of technological and organizational resources which has accompanied it; (2) enhanced global competition through a gradual liberalization of trading and investment regimes; and (3) the growth in venture capital, from the late 1960s in the United States, and in Europe and Japan in the 1980s. The growth of venture capital has provided, at least in the United States, a second avenue for developing new technologies - namely, the ‘start-up’ firm. A consequence of these trends is that industrial competence is now more widely dispersed, both organizationally and geographically. Couple this with the generic nature of many new technologies - like microelectronics and biotechnology - and one recognizes that the competitive challenges associated with developing and commercializing innovation are very considerable. That is not to say that in earlier periods - such as the second half of the 19th century - technology capacity was not geographically dispersed. However, global communications were such that the distinction between local and global was economically much sharper than it is today; the ability

8

D.J. Teece, Competition,

cooperation, and innovation

of firms to quickly assemble and efficiently operate complex forms of interfirm agreements was severely limited. The capacity to organize and operate complex and geographically dispersed organizational forms is now widely available; with enhanced competition, the need to select efficient structures is even more pressing.

3.2.

The need for operational coordination

Innovation is a special kind of economic activity, with very special kinds of informational and coordination requirements. This subsection examines these requirements, without addressing the question of whether they are organized most appropriately by markets, hierarchies, alliances or other organizational forms. I. Accessing complementary assets. Innovative new products and processes will not yield value unless they are commercialized. It is during commercialization that the greatest organizational challenges arise and where the bulk of the resource commitment occurs. The profitable commercialization of technology requires timely access to complementary assets on competitive terms. Thus an innovating firm or consortium that has developed the core technology needed for a new product or process with good commercialization prospects has taken only the first step. It must then secure access to complementary technologies and complementary assets on favorable terms in order to successfully commercialize the product or process.7 Assets such as marketing, competitive manufacturing, reputation, and after-sales support are almost always needed. The range of complementary assets that are necessary is indicated in fig. 1. These assets need not bear a relationship to the innovating unit that is either strictly vertical or horizontal.* Moreover, these assets will often need to be specialized to the innovation. For example, the commercialization of a new drug is likely to require the dissemination of information over a specialized information channel. In some cases, the complementary assets may be the other parts of a system. For instance, computer hardware may require the development of specialized software both for the operating system and for applications. The interdependence between the innovation and the relevant complementary assets can, of course, vary tremendously. At one extreme, the complementary assets may be virtually generic, have many potential suppliers, and be relatively unimportant when compared with the technological breakthrough represented by the innovation. At the other end, successful commercialization of the innovation may depend crtitically on a bottleneck ‘See Teece (1986). ‘Vertical relations are upstream or downstream. competing for the same customers in the same market.

Horizontal

relationships

involve

firms

D.J.

Fig. 1. Representative

Teece, Competition, cooperation, and innovation

complementary

assets needed to commercialize

technological

9

know-how.

asset that has only one possible supplier. Between these two extremes there is and the assets the possibility of ‘cospecialization’ - where the innovation depend on each other. An example of this would be containerized shipping, which requires specialized trucks and terminals that can work only in conjunction with each other. A salient aspect of innovation 2. Coupling developer to users and suppliers. is that it requires a close coupling of the developer of the new technology to the user. Commercially successful innovations require linking scientific, engineering, entrepreneurial and management skills with an intimate understanding of user needs. Indeed, in some fields, such as scientific instruments, it is often the user that stimulates innovation and comes up with a new product concept or product prototype which is passed back upstream for further development work. Hence, innovation requires considerable vertical interaction and communication flows. Moreover, these flows must occur expeditiously. With uncertainty, learning, and short product life cycles, there must be organizational systems in place to facilitate timely feedback, midcourse correction, redesign, and rapid commercialization. The necessary linkages and feedbacks are summarized in fig. 2. Kline and Rosenberg (1986) have described the process as ‘chain linked’. Their characterization recognizes aspects of a linear model - such as a flow of activity through design to development production and marketing - but also recognizes constant feedback between and among ‘stages’. Moreover, ‘the linkage from science to innovation is not solely or even preponderantly

10

D.J. Teece, Competition,

POTENTIAL

cooperation,

AND 0R 1 PRODUCE

XDETAILED t DESIGN

1 ANALYTIC

1AND

Fig. 2. Model showing

flow paths

and innovation

of information

and cooperation.

Symbols used on arrows: C = central-chain-of-innovation f =feedback loops F = particularly important feedback K-R: Links through knowledge to research and return paths. If problem solved at node K, link 3 to R not activated. Return from research (link 4) is problematic - therefore dashed line. D: Direct link to and from research from problems in invention and design. I: Support of scientific research by instruments, machines, tools, and procedures of technology. s: Support of research in sciences underlying product area to gain information directly and by monitoring outside work. The information obtained may apply anywhere along the chain. Source: Kline and Rosenberg (1986).

at the beginning of typical innovations, but rather extends all through the processes, to process . . . science can visualized as lying alongside development be used when needed’. The correct identification of needs is critical to the profitable expenditure of R&D dollars. R&D personnel must thus be closely connected to the market and to marketing personnel. R&D managers must have one foot in the lab and one in the marketplace. Knowing what to develop and design, rather than just how to do it, is absolutely essential for commercial success. Developing this understanding involves a complex interplay between science and engineering, manufacturing, and marketing in order to specify product functions and features. It is not just a matter of identifying user needs and assessing engineering feasibility. One must also separate those user needs which are being met by competition and those which are not, This may not

D.J. Teece, Competition,

cooperation,

and innovation

11

become clear until the product is introduced, in which case the ability to redesign quickly and efficiently may be of the utmost importance. In short, commercialization is an extremely important ingredient to successful innovation. This model recognizes the existence and exercise of tight linkages and feedback mechanisms which must operate quickly and efficiently. These linkages must exist within the firm, among firms, and between firms and other organizations, such as universities. Of course, the positioning of the firm’s boundaries, for example, its level of vertical integration, determines in part whether the required interactions are intratirm or interfirm. Thus no matter how innovation proceeds, it is likely to require access to capabilities which lie beyond the initiating or driving entity. These capabilities may lie in universities, other parts of the enterprise, or in other unaffiliated enterprises. The role of some of these key organizational units is now explored. We will examine both the development and the commercialization of new technology. In a series of important pieces, von Hippel (1977, 1988) has presented evidence that, in some industries in the United States, industrial products judged by users to offer them a significant performance improvement are usually conceived and prototyped by users, not by the manufacturers. The manufacturers’ role in the innovation process in these industries is to become aware of the user innovation and its value, and then to manufacture a commercial version of the device for sale to user firms. This pattern of innovation involving vertical cooperation is contrary to the usual assumption that product manufacturers are responsible for the innovation process from finding to filling the need. Successful management of the process requires that product engineering skills (rather than R&D skills) be resident in the manufacturer, and that manufacturers search to identify user solutions rather than user needs.’ A further implication is that there may be a symbiotic vertical relationship between users and equipment manufacturers that depends upon social and geographical proximity. Balancing the role that users play in stimulating innovation upstream is the role that suppliers play in stimulating downstream innovation. A good deal of the innovation in the automobile industry, including fuel injection, alternators, and power steering, has its origins in upstream component suppliers. lo The challenge then becomes how to ‘design in’ the new components, and possibly how to avoid sole source dependency on the part of the automotive companies. As discussed below, deep and enduring relationships need to be established between component developer-

‘Note that user innovation requires two kinds of technology transfer: first from user manufacturer, and then from manufacturer to the developer-user and other users. “‘For example, Bendix and Bosch developed fuel injection and Motorola the alternator.

to

12

D.J. Teece, Competition, cooperation, and innovation

manufacturer and supplier to ensure adoption and diffusion of the technology. A related set of vertical relationships involving innovation has been remarked upon by Rosenberg (1972) in his treatise on technology and American economic growth. The machine tool industry in the 19th century played a unique role both in the initial solution of technical problems in user industries, such as railroad locomotive manufacture. Rosenberg’s description seems to suggest that the users played some role in the development of new equipment. He notes that before 1820 in the United States, one could not identify a distinct set of firms who were specialists in the design and manufacturer of machinery. Machines were either produced by users or by firms engaged in the production of metal or wooden products.” Machineryproducing firms were thus first observed as adjuncts to textile factories. However, once established, these firms played an important role as the transmission center in the diffusion of new technology.” 3. Coupling to competitors. Successful new product and new process development innovation often requires horizontal as well as vertical cooperation. Horizontal linkages can assist in the definition of technical standards for systemic innovation. Horizontal linkages can also assist firms to overcome the appropriability (spillover) problems because the set of firms receiving the benefits is likely to include a greater portion of firms which have incurred R&D costs. The effect of greater appropriability is to encourage greater investment in new technology.’ 3 In addition, collaborative research reduces needless duplication of effort. r4 Independent research activities often proceed down identical or near-identical technological paths. This is often wasteful and can be minimized if research plans are coordinated.” Furthermore, innovation and commercialization of new products and processes are often high cost activities. The scale and scope of assets needed will often lie beyond the capabilities of a single firm. Thus, cooperation both horizontal and vertical - may be the only viable means for moving forward. In addition, cooperation will also reduce wasteful duplicate expenditures on research and development. Innovation also entails significant risk. While successful innovation and its financial rewards are often highly visible, behind the scenes there are usually many failed efforts and unproductive paths. ‘Dry holes’ and ‘blind alleys’ are commonplace. Risk can be diversified and spread through cooperation. Indeed, when risk is particularly high “See Rosenberg (1972, p. 98-99). “Ibid., p. 102. 13This has been shown empirically by Levin and Reiss (1984). “‘See David (1985, p. 42). ISNeedless to say, uncertainty often requires that multiple (but not identical) paths be pursued simultaneously. See, for example, Nelson (1984, chapter 2).

technological

D.J. Teece, Competition, cooperation, and innovation

13

because the technology being pursued is both expensive and undeveloped, cooperation may be the only way that firms will undertake the needed effort. Until very recently, it has been fashionable in the United States to argue that diversity is the leitmotif of successful innovation. Unquestionably, a system of innovation that converges on just one view of possibilities is likely to close off productive avenues of inquiry. However, in a private enterprise economy without some horizontal coordination and communication, there is no guarantee that the level of diversity obtained is ideal. If firms are able to coordinate their research programs to some degree, uneconomic duplication in some instances can be minimized without the industry converging on a single technological approach. Another dimension of coordination is 4. Connections among technologies.“j that which must take place among various technologies. Particular technological advances seldom stand alone. They usually are connected both to (a) prior developments in the same technology, and (b) to complementary or facilitating advances in related technologies. In addition, a generic technology may be capable of (c) a wide variety of end-product applications. Each is discussed in turn, along with related reasons for cooperation.” Many technologies evolve in an evolua. Connections to prior technologies. tionary fashion, with today’s round of R&D activities building on yesterday’s, which in turn builds on the day before’s. Thus, with respect to memory devices, advance is cumulative along a particular technological trajectory, from 1 K to 4 K to 64K to 256 K to 1 megabyte and so on, up to the theoretical limits of a particular technology. Similarly, in the aircraft industry, the DC-3 improved upon the DC-2, and subsequent aircraft built on what was learned with the DC-3. Hence, whether an enterprise is able to participate in one technology often depends on whether it participated in the earlier generation. If it did not, and the technologies in questions have path dependencies, then in order to enter at a subsequent stage the enterprise in question will have to link up, in some fashion, with enterprises familiar with the prior art. Relatedly, experience and competence in a particular technological regime may count for little, and may in fact be a handicap, when a significant shift in technological regime occurs. A regime shift signals opportunities for new entrants; but to engage these opportunities, new entrants may need to link up with incumbent firms who may not have the relevant new technology, but do in fact have complementary capacities, such as marketing and distribution. 16This “It is approach implicitly see Teece

treatment is based in part on Nelson (1984, pp. 8-10). obvious that the treatment here breaks with the traditional ‘book of blueprints’ to technology often embedded in economic models of the innovation process. We postulate market failure in the market for know-how. For a more extensive treatment, (1980).

D.J. Teece, Competition, cooperation, and innovation

14

b. Connections to complementary technologies. Technological advances are often linked together because of systems interdependencies. Thus, the electric lighting system - dynamos, distribution system, and incandescent bulbs - was severely limited by nondurable bulb filaments that existed until Edison and his associates invented a high resistance filament. Artificial intelligence technologies depend both on the availability of computer software and hardware. Containerization of ocean shipping could not move forward until new facilities to handle containers had been put in place in ports, in distribution centers, and in railroads and trucking. The innovation, of course, also required the construction of ships with the requisite capacities. In a system technology, an advance in one part of the system may not only permit but require changes in other parts. The tightness of interdependence and the requirements for organizational connectedness are discussed later. Suffice to say here that tight interdependencies require interaction and information and materials flows among organizations. Whether the appropriate governance mechanism is a contract, a joint venture, equity crossholdings, or merger is a matter we consider in section 4. may simultac. Connections to enabling technologies. New technologies neously affect several different activities, typically because they provide a common core technology to several businesses (e.g., microelectronics, composites, biotechnology). Thus advances in semiconductors increase the technical capabilities of industries that use chips, from computers to telecommunications. Advances in semiconductors also facilitate more powerful computeraided engineering and design and computer-integrated manufacturing.18 Organizations must be designed in order to facilitate the complex cooperation which utilization of the enabling technology requires. Without the appropriate organizational mechanism, the full potential of the enabling technology will not be realized.

3.3. Strategic coordination In section 3.2, the operational dimensions of coordination were sketched. These activities must take place if any system is to be innovative. Strategic coordination refers to coordination of a different type. It refers to activities which affect the distribution of returns to innovation through impacts on prices and competitive entry. Because of fundamental weaknesses in the system of intellectual property Moreover, because of the law,19 leakage and free riding are commonplace. use of strategic trade and industrial policies by some nation states, strategic “See “See

National Advisory Committee on Semiconductors Report Levin, Klevorick, Nelson and Winter (1987, p. 3).

(1989)

D.J. Teece, Competition, cooperation, and innovation

15

coordination either by firms or governments is often necessary if innovating firms are to capture value from technology. Strategic coordination can thus enhance the public as well as private interests. Strategic coordination is often desirable to ensure that capacity levels at various stages in an industry are matched to the level of residual demand, to ensure that beneficial standards are adopted, and to ensure that strategic moves of competitors are blunted if not deterred, and to ensure that economies and advantages associated with increasing returns are attained. Unfortunately, in the United States, strategic behavior is often uncritically viewed as anticompetitive, with negative effects on economic welfare assumed. While such behavior may be ‘anticompetitive’ in the sense that it limits the entry of competitors, strategic behavior can sometimes enhance national if not global economic welfare.20 This is because it can help innovators capture enough of the return from innovation to keep them in the business of innovation, thereby invigorating competition. There is no opportunity in this paper to document all the various forms of strategic coordination. Suffice to point out that the timely assembly of complementary assets may well have both a strategic and an operational dimension, as is discussed more fully elsewhere [Teece (1986)].

4. Governance structures to facilitate 4.1. The organizational

innovation

menu

There are of course a wide variety of organizational mechanisms that can be used to effectuate the coordination which successful innovation requires. Economists at some level have been aware of this for decades. Thus, ‘Marshall introduces organization as a fourth factor of production; J.B. Clark gives the coordinating function to the entrepreneur; Knight introduces managers who coordinate’ [Coase (1937)]. But it is only recently that the emerging literature significant progress has been made.21 Unfortunately, on the economics of organization has yet to deal with matters of innovation. The state of the art is such that many economists are still in the uncomfortable position of claiming that the price mechanism achieves the necessary coordination while at other times entrepreneurs and managers do

“‘Jacquemin and Slade (1989) single out two areas where governments can usefully encourage cooperative behavior: R&D-intensive industries, and declining industries. The work of Brian Arthur (1988) also points out the critical role of increasing returns and lock-in by small chance events. Government policy obviously can be of great importance, positively and negatively, in such environments. 2’Ohver Williamson (1975, 1985) has pioneered the economic analysis of organizations building on fundamental insights presented by Coase (1937).

16

D.J. Teece, Competition, cooperation, and innovation

the coordination.22 While Williamson (1975, 1985) has elaborated the distinctive merits of markets and certain hierarchical forms, alternative organizational arrangements until recently have received relatively short shrift in the literature. But interlirm agreements and alliances must clearly be among the alternatives considered. The following subsections will explore each of the various options in more detail.

4.2.

The price system

The invisible hand is one of the oldest notions in economic science. Prices are considered sufficient statistics; sufficient that is to guide resource allocations toward Pareto optimal outcomes. The invisible hand theorems require, it would seem, that economic agents know not only today’s supply and demand but supply and demand for all future periods. Otherwise, we cannot be confident that the market outcomes generated will be optimal. In particular, investment levels will be incorrect unless we know about future market demand and the future investment plans of competitors and suppliers. Any single investment will, in general, only be profitable provided first that the volume of competitive investment does not exceed a limit set by demand, and second, that the volume of complementary investment reaches the correct leve1.23 However, there is no special machinery in a private enterprise, market economy to ensure that investment programs are made known to all concerned at the time of their inception. Price movements, by themselves, do not generally form an adequate system of signalling. Indeed, Koopmans (1957) has been rather critical of what he calls the ‘overextended belief’ of certain economists in the efficiency of competitive markets as a means of allocating resources in a world characterized by ubiquitous uncertainty. The main source of this uncertainty, according to Koopmans, is the ignorance which firms have with respect to their competitors’ future actions, preferences, and states of technological information. In the absence of a complete set of forward markets in which anticipation and intentions could be tested and adjusted, there is no reason to believe that with uncertainty competitive

*‘As Coase (1937) notes, ‘it is surely important to enquire why coordination is the work of the price mechanism in one case and of the entrepreneur in another.’ Coase went on to develop an approach to economic organization that saw the firm as suppressing the price mechanism in circumstances where transaction costs were high. Hayek (1945) saw the situation somewhat similarly, there being three fundamental choices: central planning, competition (decentralized planning), and monopoly - the latter he refers to as the ‘half-way house’ about which many people talk but which few like. Despite recent progress, the economic literature has not gone much beyond Hayek and Coase. ‘%ee Richardson (1960), p. 31.

D.J. Teece, Competition, cooperation, and innovation

17

markets of the kind described in the textbooks produce efftcient outcomes.24 The information-circulating function which economic theory attributes to competitive markets is quite simply not discharged by any existing arrangements with the detail and forward extension necessary to support efficient mechanisms are needed. outcomes.25 Ancillary org anizational Arrow (1959) suggests that the firm performs the dual role of entrepreneur and auctioneer when the system is out of equilibrium. Richardson (1960) argues that some form of market imperfection is essential to the process of economic adjustment. Telser (1985) has articulated the case for efficient cartels where firms discuss new investment decisions and share capacity in order to overcome the shortcomings of the price system. There is no arena in which uncertainty is higher and the need to coordinate greater than in the development and commercialization of new technology. Moreover, there are no theoretical grounds for believing that an economy that is all tooth and claw will out perform one that involves a judicious mix of tooth, claw, and intramarket and intermarket cooperation. Adam Smith’s2‘j warning - that ‘people of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public, or in some contrivance to raise prices’ - needs to be tempered with the recognition that in global markets the difficulty of assembling all the relevant parties to effectuate an international conspiracy is an insurmountable challenge in industries’ experiencing rapid technological change. It also needs to be tempered with a recognition that the exchange of information on common values such as future demand is often socially desirable. This is because prices clearly are not sufficient signals in all cases. The price mechanism does not always find prices efficiently, and it may fail to elicit those responses from economic agents that will maintain an optimal allocation of resources. The presence of scale economics or collective goods are just two considerations which upset the invisible hand theorems. Yet in the traditional textbook, the price system magically allows all of the necessary coordination to occur smoothly and efficiently. In the Walrasian system, which is implicit in textbook thinking about equilibrium price determination, profits cause output expansion, losses contraction.27 In 24Koopmans (1957, p. 146) points out that because of this deficiency economic theorists are not able to speak with anything approaching scientific authority on matters relating to individual versus collective enterprise. *‘Ibid., p. 163. 26This did not appear to be an important part of Smith’s (1976) overall thesis. “As Aoki (1984) explains, the prices of goods and services are set by the auctioneer, who adjusts them according to the law of supply and demand. Given a system of prices, the excess of sales price of entrepreneurial output over the cost of production may be either positive, null, or negative. This excess is termed benefice de Penterprise by Walras (1954). A positive or negative benefice is a sign of disequilibrium, and entrepreneurs respond to this signal according to the law of cost price; that is, they increase their scale of production when the benefice is positive and reduce it when the benejce is negative. The presumption that firms strive for higher incomes and

18

D.J. Teece, Competition, cooperation, and innovation

equilibrium, firms make neither profits nor losses. In this system, entrepreneurs, together with the auctioneer, act as coordinators to bring harmony to the competitive pursuit of self-interest. In this simple view, the only information firms need to develop and commercialize innovation is provided by the auctioneer in terms of price. A pure unassisted Walrasian-type market system is clearly inadequate to effectuate most types of innovation. Yet many economists cling to the belief that it is adequate, or very nearly so. This would appear to stem from what Koopmans calls an overextended belief in the efftciency of competitive markets, coupled with a lack of understanding of the nature of the innovation process. 4.3. Internal

organization

Coase (1937) saw the costs of using the price system, and in particular the costs of discovering the relevant prices, as the fundamental reason why it is profitable to establish firms. Williamson (1975) took several cues from this thesis and has developed a more sophisticated contractual variant of the Coase thesis. He shows that it is not so much the ditliculty of establishing prices as the difficulty of regulating economic activity with incomplete contracts that provides the reason to abandon markets in favor of internal organization. Unless safeguards can be provided against opportunistic recontracting, transactions must be brought inside the firm where a managerial hierarchy and appropriate incentives can ameliorate the consequences of opportunism. But the full integration solution has liabilities, not least of which is that it can impair the autonomy so necessary for aspects of the innovation process to proceed. Still, it has obvious advantages for effectuating strategic coordination. For example, Michael Borrus (1988, p. 231) recognizes that new institutional forms are needed in U.S. microelectronics if the industry is to compete effectively in the future and at the same time avoid antitrust problems.” However, his solution - the establishment of a number of lower losses through entry and exit is implicit. However, entrepreneurs in their purely functional roles are only catalytic agents who accelerate combinations of atomistic factors of production only when the benefice is positive. Thus, in a state of equilibrium, entrepreneurs make neither profit nor loss. ‘Profit in the sense of benefice de I’enterprise depends upon exceptional and not upon normal circumstances.’ *sBorrus (1988) comments: Almost a decade ago it was clear that Japanese producers would emerge as enduring players in the semiconductor industry, and would, as a consequence, radically alter the industry’s terms of competition. Yet it has taken almost that long for U.S. firms to cooperate sufficiently to begin to devise appropriate responses. It is almost a truism that had the industry been able to coordinate its actions strategically a decade ago, an adequate response would have been far less costly and far more likely to succeed. To accomplish such strategic coordination, the US. chip industry needs an ongoing analytic capacity, embedded in an electronics industry-wide institution, with the ability to carry on competitive analysis

D.J. Teece, Competition,

cooperation,

and innovation

19

holding companies for Silicon Valley firms - implicitly endorses internal organization (accomplished via merger) as the solution to the problem.2g While it may be better than the current fragmented structure, there are many reasons why one should be skeptical of its viability. Many of these relate to the change in incentives and compensation structures which the arrangements would imply. 3o As explained below, when high technology activities are at issue, contractual agreements, alliances, and joint ventures are likely to be superior to full-scale internal organization.

4.4. Strategic

alliances and interfirm

agreements

There are a wide range of contractual agreements between and among firms to facilitate the development and commercialization of new technologies. Strategic alliances are constellations of bilateral and possibly multilateral contracts and understandings among firms, typically to develop and commercialize new technology. These may well constitute a new organizational form. Agreements are bilateral when A agrees to secure Y from B as a condition for making X available, and both parties understand that agreement will be of these continued only if reciprocity is observed. 31 When constellations agreements exist between or among firms, they can be considered to constitute alliances. While the term is often used rather loosely, a strategic alliance is defined here as a constellation of agreements characterized by the commitment of two or more partner firms to reach a common goal, entailing the pooling of their resources and activities. A strategic alliance might include the following: (i) an exclusive purchase agreement; (ii) exclusionary market or manufacturing rights; (iii) technology swaps; (iv) joint R&D or codevelopment agreements; (v) co-marketing arrangements. A strategic alliance denotes some degree of strategic as well as operational coordination. of foreign market and technology strategies, and with sufftcient prestige to offer strategic direction on which planning can occur. There is, of course, a substantial problem associated with industry-strategic planning. To assure its health, the industry needs strategic coordination short of market sharing. All U.S. industries facing international competition ought to be permitted to develop industry-wide competitive assessment capability, at least whenever the industry can demonstrate substantial involvement of foreign governments in assisting foreign competitors, 29Borrus (1988, p. 233) remarked: [I]t is possible to envision chip-firm holding companies built around common manufacturing facilities .._ R&D resources shared among the holding company’s chip Iirms would eliminate the problem of duplication of R&D among smaller companies. The high capital costs of staying in the technology race could be shared among firms in the form of shared flexible fabrication facilities _. Shared facilities would permit high usage of capacity In essence, the holding company structure would gain the advantages associated with consolidation without the disadvantages associated with integration. “See, in particular, Williamson (1985). “‘Ibid., p. 191.

20

D.J. Teece, Competition, cooperation, and innovation

Alliances can be differentiated from exchange transactions, such as a simple licensing agreement with specified royalties, because with an exchange transaction the object of the transaction is supplied by the selling firm to the buying firm in exchange for cash. An alliance by definition can never have one side receiving only cash; moreover, it consists of a constellation of bilateral agreements. Alliances need not involve equity swaps or equity investments, though they often do. Equity alliances can take many forms, including minority equity holdings, consortia, and joint ventures. Strategic alliances are capable of achieving the various forms of operational and strategic coordination described in sections 3.2 and 3.3. Complementary assets, user know-how, complementary technologies and enabling technologies can all be accessed in this way, though the terms are likely to be different from what could be obtained through a intralirm agreement were it feasible since the costs and risks of innovation can be lowered through such arrangements. Strategic alliances have increased in frequency in recent years, and are particularly characteristic of high technology industries. Joint R&D, knowhow, manufacturing, and marketing agreements go well beyond exchange agreements because they can be used to access complementary technologies and complementary assets. The object of the transactions, such as the development and launch of a new product, usually does not exist at the time the contracts are inked. Whether equity or nonequity forms of alliances are the most desirable governance structures depends on a variety of circumstances.32 Equity stakes provide a mechanism for distributing residuals when ex ante contractual agreements cannot be written to specify or enforce a division of returns. If equity membership also provides board membership, higher level strategic coordination is also possible in some cases. These governance structures are summarized in fig. 3. Strategic alliances appear to be an attractive organizational form for an environment characterized by rapid innovation and geographical and organizational dispersion in the sources of know-how. As compared to the price system, they enable investment plans for complementary assets to be coordinated more concisely than the price system would allow; as compared to hierarchy, incentives are not dulled through bureaucratic decision making. In short, the strategic alliance appears to be a hybrid structure well suited to today’s global realities in industries experiencing rapid technological change. Such industries frequently require operational and strategic coordination. Alliances facilitate reciprocal specialization among firms, such as when one firm does development and its partner manuufacturing. But as suggested elsewhere [Teece (1986)], market failures are such that the distribution of the j2For

a fuller treatment,

see Pisano

and Teece (1989).

21

D.J. Teece, Competition, cooperation, and innovation Nonequity

Equity

Exchange

short-medium term cash-based contracts

passive stock holdings for portfolio diversification

Alliance

mid-long-term bilateral contracts (non-cash based), non-operating joint ventures and consortia

operating joint ventures and consortia, minority equity holdings, and cross-holdings

Cartel

price fixing and/or output restricting agreements

price fixing and/or output restricting agreements

Fig. 3. Taxonomy

profits that partners.

may

follow

need

not

of interfirm

arrangements.

be equally

5. Strategic alliances and the logic of managerial

distributed

among

alliance

capitalism

The above analysis might appear, at first glance, to be at odds with Chandler’s recent claim (1990) that firms that deny whay he calls ‘the logic of managerial enterprise’ will not survive. Chandler’s historical analysis suggests that innovating firms that fail to invest in facilities of sufficient scale will lose market share and technological leadership to competitors. Thus, according to Chandler the early lead which the British had in dyestuffs was lost to the Germans because British firms failed to make the necessary investments to capture economies of scale and scope. This is the principal reason why German firms outcompeted the British; vertical integration is thus essential to managerial capitalism according to the Chandler argument. The thesis in this paper is that the logic of managerial capitalism may well be eroding, although not as fast as the growth of strategic alliances involving U.S.-based firms might suggest. The ‘logic’ of managerial capitalism is being tempered in some industries because of the heightened importance of generic enabling technologies and time to market. Both favor small autonomous organization units that can avoid the complex decision-making complexities usually, though not always, associated with large integrated hierarchies. Large amounts of information critical to innovation must move horizontally; if forced to move vertically before it can move horizontally, its timeliness as well as its value will erode dramatically. A valued price system has no advantage in this regard; small

22

D.J. Teece, Competition, cooperation, and innovation

autonomous organizational units communicating and interacting directly with each other are likely to develop and implement an action plan faster and more efficiently than if they were responding just to price signals. When product life cycles are short, the ability to quickly and effectively coordinate business units into a coherent operation often assumes great significance. One additional reason for the frequency of strategic alliances involving U.S. high-tech firms, however, is the decline in U.S. manufacturing skills, particularly in relation to Japan and Germany. U.S. firms still seem to possess a strong advantage in early-stage technological activity, but their advantage at commercialization has eroded markedly in recent decades. It is therefore advantageous for innovative firms which are unable to build and operate world-class manufacturing facilities to team up with those that can. Unfortunately for U.S. firms, the market for know-how is riddled with imperfections, and the ability of innovators to capture a significant share of the rents through alliances is highly attenuated - except in unusual circumstances where the innovator has a strong worldwide patent position. When intellectual property protection is weak, strategic alliances - implying a constellation of interlirm agreements and the ability to temper opportunistic behavior - are likely to be more effective than a pure licensing arrangement. But an alliance is likely to be inferior to vertical integration if the innovator is in the position to integrate at low cost. In the United States one can perhaps agree with Chandler that there aren’t intrinsic theoretical disabilities associated with vertical integration; but there may be disabilities associated with an integration strategy built on decaying U.S. manufacturing skills. Hence, strategic alliances may often be the best alternative available to U.S. firms, though that alternative may well be inferior to an integration strategy based on world-class manufacturing. Hence while strategic alliances may be the best that many U.S. firms can do, it may not be enough, as Chandler suggests, to survive in today’s highly-competitive global environment.

6.

Conclusion

Successful technological innovation requires complex forms of business organization. To be successful, innovating organizations must form linkages, upstream and downstream, lateral and horizontal. Advanced technological systems do not and cannot get created in splendid isolation. The communication and coordination requirements are often quite stupendous, and in practice the price system alone does not suffice to achieve the necessary coordination. A variety of organizational arrangements exist to bring about the necessary coordination. The price system described in the textbooks provides only a useful backdrop in market-oriented economies. By itself it is not up to the

D.J. Teece, Co~~tiii5n.

cooperation, and in~5~ation

23

task because investments required to effectuate innovation cannot be appropriately coordinated by price signals alone. Fully integrated companies, on the other hand, must be careful not to suffocate creativity and to dampen incentives. Strategic alliances constitute viable alternatives in many instances. Alliance structures can facilitate innovation, and are increasingly necessary as the sources of innovation and the capacities necessary to effectuate commercialization become increasingly dispersed. A variety of implications follow for management and public policy. Managers must become adept at managing not just their own organization, but also their relationships and alliances with other firms. Very often difference skills are required for each, which makes management tasks more complex and challenging. Strategic alliances also must be designed to be selfreinforcing. There is the danger that changing circumstances can upset delicate balances, thereby causing relationships and agreements to unravel. Equity can be judiciously used to anchor alliances, but it may not suffice as a safeguard if inadequate in amount, and if not bolstered by other mechanisms. At present there is considerable organizational learning occurring with respect to such matters, both nationally and internationally, and new arrangements which are balanced and durable will undoubtedly be crafted. The greatest challenge probably goes to public policy, particularly in the United States, where there has been a failure to recognize the importance of cooperation. This has manifested itself in the absence of inter-agency coordination in the federal government, with science and technology policy appearing to be weak and uncoordinated, and a reluctance in some quarters to permit and encourage the private sector to forge the interfirm agreements, alliances and consortia necessary to develop and commercialize new technologies like high-density television or superconductors. A key reason for this is the shadow that neoclassical thinking casts over antitrust policy. It renders antitrust policy hostile to many forms of beneficial collaboration, because of fear that such arrangements are a subterfuge for cartelization and other forms of anticompetitive behavior. Until a greater understanding emerges as to the organizational requirements of the innovation process and antitrust uncertainties are removed, antitrust policy in the United States is likely to remain a barrier to innovation because it has the capacity to stifle beneficial forms of interfirm cooperation. Antitrust policy, which has always been rather hostile to horizontal agreements, must accommodate this feature of capitalist economies if it is to promote enterprise performance and economic welfare.33 The framework also has strong implications for business history. It suggests the viability in some instances of new hybrid organizational arrangements, such as strategic alliances, linking firms with complementary 33For fuller elaboration,

see Jorde and Teece (1989, 1990).

24

D.J. Teece, Competition,

cooperation,

and innovation

capabilities and capacities, over both the integrated alternatives and standalone firms coordinating their activities exclusively via the price mechanism.34 These organizational forms may well represent a new and dramatic organizational innovation in American business history. In retrospect, the emergence and proliferation of alliances, dating from about 1970, may turn out to be as significant an organizational innovation as the moving assembly line and the multidivisional structure. This trend does not deny what Chandler calls the logic of managerial enterprise; it may simply suggests its declining attainability for U.S. companies. 34These instances

are more clearly delineated

in Dosi, Teese and Winter

(1990).

References Aoki, M., 1984, The cooperative game theory of the firm (Clarendon Press, London). Arrow, Kenneth J., 1959, Toward a theory of price adjustment, in: Moses Abramovitz et al., eds., The allocation of resources (Stanford University Press, Stanford, CA) 41-51. Arthur, W. Brian, 1988, Competing technologies: An overview, in: Giovanni Dosi et al., eds., Technical change and economic theory (Pinter Publishers, London) 59@607. Baldwin, W. and J.T. Scott, 1987, Market structure and technological change (Harwood Publishers, Chur). Borrus, Michael, 1988, Competing for control: America’s stake in micro electronics (Ballinger, Cambridge, MA). Caves, R. and M. Uekusa, 1976, Industrial organization in Japan (The Brookings Institute, Washington, DC). Chandler, Alfred D., 1977, The visible hand (Harvard University Press, Cambridge, MA). Chandler, Alfred D., 1990, The enduring logic of industrial success, Harvard Business Review, March-April, 434444. Coase, Richard, 1937, The nature of the firm, Economica 4, 386405. David, D., 1985, R&D consortia, High Technology, Oct. Dasgupta, Partha D., 1988, The welfare economics of knowledge production, Oxford Review of Economic Policy 4, no. 4, Winter. Dosi, Giovanni, David J. Teece and Sidney Winter, 1990, Toward a theory of corporate coherence: Preliminary remarks. Unpublished manuscript (University of California at Berkeley, Berkeley, CA). Gerlach, Michael, 1992, Alliance capitalism (University of California Press, Berkeley, CA) forthcoming. Hayek, F., 1945, The use of knowledge in society, American Economic Review 35, 519-530. Jacquemin, A. and M.E. Slade, 1989, Cartels, collusion and horizontal merger, in: R. Schmalensee and R.D. Willig, Handbook of Industrial Organization 1 (Elsevier, Amsterdam). Jorde. Thomas M. and David J. Teece, 1989, Innovation, cooperation and antitrust, High Technology Law Journal 4, no. 1, Spring, l-l 13. Jorde. Thomas M. and David J. Teece, 1990, Innovation and cooperation: Implications for competition and antitrust, Journal of Economic Perspectives 4, no.3, 75-96. Kline, S.J. and Nathan Rosenberg, 1986, An overview of innovation, in: R. Landau and N. Rosenberg, eds., The positive sum strategy (National Academy Press, Washington, DC) 275-305. Koopmans, Tjalling, 1957, Three essays in the state of economic science (McGraw-Hill, New York). Levin, R. and P. Reiss, 1984, Tests of a Schumpeterian model of R&D and market structure, in: Z. Griliches, eds., R&D, patents and productivity (University of Chicago Press, Chicago, IL). Levin, R., R. Klevorick and Sidney Winter, 1987, Appropriating the returns from individual research and development, Brookings Papers on Economic Activity.

D.J. Teece, Competition,

cooperation,

and innovation

25

Mowery, David C. and Nathan Rosenberg, 1989, Technology and the pursuit of economic growth (Cambridge University Press, New York). National Advisory Committee on Semiconductors, 1989, A strategic industry at risk, Report to the President and the Congress (U.S. Government Printing Office, Washington, DC). Nelson, Richard, 1984, High technology policies (American Enterprise Institute, Washington, DC). Nelson, Richard and Sidney Winter, 1982, An evolutionary theory of economic change (Harvard University Press, Cambridge, MA). Psiano, Gary and David J. Teece, 1989, Collaborative arrangements and global technology strategy: Some evidence from the telecommunication equipment industry, in: R.A. Burgelman and R.S. Rosenbloom, eds., Research on Technological Innovation, Management and Policy 4 (JAI Press, Greenwich, CT) 227-256. Richardson, George, 1960, Information and investment: A study in the working of a competitive economy (Oxford University Press, London). Rosenberg, Nathan, 1972, Technology and American economic growth (Harper & Row, New York). Rosenberg, Nathan, 1982, Inside the black box: Technology and economics, vol. 7 (Cambridge University Press, New York). Scherer, F.M., 1970, Industrial market structure and economic performance (Rand-McNally, Chicago, IL). Schumpeter, J.A., 1942, Capitalism, socialism and democracy (Harper, New York). Smith, Adam, 1976, Wealth of nations, book 1, chapter X, Part II (University of Chicago Press edition, Chicago, IL). Stiglitz, Joseph, 1987, Technical change, sunk costs and competition, Brookings papers on economic activity, vol. 3 (Special Issue on Microeconomics) (The Brookings Institution, Washington, DC) 883-937. Teece, David J., 1980, Economics of scope and the scope of enterprise, Journal of Economic Behavior and Organization 1, no. 3, 2233247. Teece, David, J., 1986, Profiting from technological innovation, Research Policy 15, no. 6, 286305. Telser, L., 1985, Cooperation, competition and efficiency, Journal of Law and Economics 28, 271. von Hippel, Eric., 1977, The dominant role of the user in semiconductor and electronic subassembly process innovation, IEEE Transactions on Engineering Management EM-24, no. 2, May. von Hippel, Eric., 1988, The sources of innovation (Oxford University Press, New York). Walras, Leon, 1969, Elements of pure economics (translated by William Jaffe from the original French version, 1874) (Kelley, New York). Williamson, Oliver E., 1975, Markets and hierarchies (Free Press, New York). Williamson, Oliver E., 1985, The economic institutions of capitalism, chapter 6 (Free Press, New York).