strategy essay

Strategic Management Journal, Vol. 18:7, 509–533 (1997)

DYNAMIC CAPABILITIES AND STRATEGIC

MANAGEMENT

DAVID J. TEECE1*, GARY PISANO2 and AMY SHUEN3 1Haas School of Business, University of California, Berkeley, California, U.S.A. 2Graduate School of Business Administration, Harvard University, Boston, Massa- chusetts, U.S.A. 3School of Business, San Jose State University, San Jose, California, U.S.A.

The dynamic capabilities framework analyzes the sources and methods of wealth creation and capture by private enterprise firms operating in environments of rapid technological change. The competitive advantage of firms is seen as resting on distinctive processes (ways of coordinating and combining), shaped by the firm’s (specific) asset positions (such as the firm’s portfolio of difficult-to-trade knowledge assets and complementary assets), and the evolution path(s) it has adopted or inherited. The importance of path dependencies is amplified where conditions of increasing returns exist. Whether and how a firm’s competitive advantage is eroded depends on the stability of market demand, and the ease of replicability (expanding internally) and imitatability (replication by competitors). If correct, the framework suggests that private wealth creation in regimes of rapid technological change depends in large measure on honing internal technological, organizational, and managerial processes inside the firm. In short, identifying new opportunities and organizing effectively and efficiently to embrace them are generally more fundamental to private wealth creation than is strategizing, if by strategizing one means engaging in business conduct that keeps competitors off balance, raises rival’s costs, and excludes new entrants. 1997 by John Wiley & Sons, Ltd.

INTRODUCTION respect to assisting in the understanding of how and why certain firms build competitive advan- tage in regimes of rapid change. Our approach isThe fundamental question in the field of strategic

management is how firms achieve and sustain especially relevant in a Schumpeterian world of innovation-based competition, price/performancecompetitive advantage.1 We confront this question

here by developing the dynamic capabilities rivalry, increasing returns, and the ‘creative destruction’ of existing competences. Theapproach, which endeavors to analyze the sources

of wealth creation and capture by firms. The approach endeavors to explain firm-level success and failure. We are interested in both building adevelopment of this framework flows from a

recognition by the authors that strategic theory is better theory of firm performance, as well as informing managerial practice.replete with analyses of firm-level strategies for

sustaining and safeguarding extant competitive In order to position our analysis in a manner that displays similarities and differences withadvantage, but has performed less well with existing approaches, we begin by briefly reviewing accepted frameworks for strategic man-

Key words: competences; capabilities; innovation;agement. We endeavor to expose implicit assump- strategy; path dependency; knowledge assets

tions, and identify competitive circumstances *Correspondence to: David J. Teece, Institute of Management,where each paradigm might display some relative Innovation and Organization, Haas School of Business, Uni-advantage as both a useful descriptive and norma- versity of California, Berkeley, CA 94720–1930, U.S.A.

tive theory of competitive strategy. While numer-1 For a review of the fundamental questions in the field of strategy, see Rumelt, Schendel, and Teece (1994 ). ous theories have been advanced over the past

CCC 0143–2095 /97/070509–25$17.50 Received 17 April 1991  1997 by John Wiley & Sons, Ltd. Final revision received 4 March 1997

510 D. J. Teece, G. Pisano and A. Shuen

two decades about the sources of competitive recognizes but does not attempt to explain the nature of the isolating mechanisms that enableadvantage, many cluster around just a few loosely

structured frameworks or paradigms. In this paper entrepreneurial rents and competitive advantage to be sustained.we attempt to identify three existing paradigms

and describe aspects of an emerging new para- Another component of the efficiency-based approach is developed in this paper. Rudimentarydigm that we label dynamic capabilities.

The dominant paradigm in the field during efforts are made to identify the dimensions of firm-specific capabilities that can be sources ofthe 1980s was the competitive forces approach

developed by Porter (1980). This approach, advantage, and to explain how combinations of competences and resources can be developed,rooted in the structure–conduct–performance

paradigm of industrial organization (Mason, 1949; deployed, and protected. We refer to this as the ‘dynamic capabilities’ approach in order to stressBain, 1959), emphasizes the actions a firm can

take to create defensible positions against com- exploiting existing internal and external firm- specific competences to address changingpetitive forces. A second approach, referred to as

a strategic conflict approach (e.g., Shapiro, 1989), environments. Elements of the approach can be found in Schumpeter ( 1942), Penrose (1959),is closely related to the first in its focus on

product market imperfections, entry deterrence, Nelson and Winter ( 1982), Prahalad and Hamel (1990), Teece (1976, 1986a, 1986b, 1988) andand strategic interaction. The strategic conflict

approach uses the tools of game theory and thus in Hayes, Wheelwright, and Clark (1988): Because this approach emphasizes the develop-implicitly views competitive outcomes as a func-

tion of the effectiveness with which firms keep ment of management capabilities, and difficult- to-imitate combinations of organizational, func-their rivals off balance through strategic invest-

ments, pricing strategies, signaling, and the con- tional and technological skills, it integrates and draws upon research in such areas as the manage-trol of information. Both the competitive forces

and the strategic conflict approaches appear to ment of R&D, product and process development, technology transfer, intellectual property, manu-share the view that rents flow from privileged

product market positions. facturing, human resources, and organizational learning. Because these fields are often viewedAnother distinct class of approaches empha-

sizes building competitive advantage through cap- as outside the traditional boundaries of strategy, much of this research has not been incorporatedturing entrepreneurial rents stemming from funda-

mental firm-level efficiency advantages. These into existing economic approaches to strategy issues. As a result, dynamic capabilities can beapproaches have their roots in a much older

discussion of corporate strengths and weaknesses; seen as an emerging and potentially integrative approach to understanding the newer sources ofthey have taken on new life as evidence suggests

that firms build enduring advantages only through competitive advantage. We suggest that the dynamic capabilitiesefficiency and effectiveness, and as developments

in organizational economics and the study of approach is promising both in terms of future research potential and as an aid to managementtechnological and organizational change become

applied to strategy questions. One strand of this endeavoring to gain competitive advantage in increasingly demanding environments. To illus-literature, often referred to as the ‘resource-based

perspective,’ emphasizes firm-specific capabilities trate the essential elements of the dynamic capa- bilities approach, the sections that follow compareand assets and the existence of isolating mech-

anisms as the fundamental determinants of firm and contrast this approach to other models of strategy. Each section highlights the strategicperformance (Penrose, 1959; Rumelt, 1984;

Teece, 1984; Wernerfelt, 1984).2 This perspective

how. Over time, these assets may expand beyond the point of profitable reinvestment in a firm’s traditional market.2 Of these authors, Rumelt may have been the first to self-

consciously apply a resource perspective to the field of strat- Accordingly, the firm may consider deploying its intangible assets in different product or geographical markets, where theegy. Rumelt (1984: 561) notes that the strategic firm ‘is

characterized by a bundle of linked and idiosyncratic resources expected returns are higher, if efficient transfer modes exist.’ Wernerfelt ( 1984) was early to recognize that this approachand resource conversion activities.’ Similarly, Teece (1984:

95) notes: ‘Successful firms possess one or more forms of was at odds with product market approaches and might consti- tute a distinct paradigm of strategy.intangible assets, such as technological or managerial know-

Dynamic Capabilities 511

insights provided by each approach as well as (1980). Competitive strategies are often aimed at altering the firm’s position in the industryvis-à-the different competitive circumstances in which

it might be most appropriate. Needless to say,vis competitors and suppliers. Industry structure plays a central role in determining and limitingthese approaches are in many ways complemen-

tary and a full understanding of firm-level, com- strategic action. Some industries or subsectors of industriespetitive advantage requires an appreciation of all

four approaches and more. become more ‘attractive’ because they have struc- tural impediments to competitive forces (e.g., entry barriers) that allow firms better oppor- tunities for creating sustainable competitiveMODELS OF STRATEGY

EMPHASIZING THE EXPLOITATION advantages. Rents are created largely at the indus- try or subsector level rather than at the firm level.OF MARKET POWER While there is some recognition given to firm-

Competitive forces specific assets, differences among firms relate primarily to scale. This approach to strategyThe dominant paradigm in strategy at least during

the 1980s was the competitive forces approach. reflects its incubation inside the field of industrial organization and in particular the industrial struc-Pioneered by Porter (1980), the competitive

forces approach views the essence of competitive ture school of Mason and Bain3 (Teece, 1984). strategy formulation as ‘relating a company to its environment . . . [T]he key aspect of the firm’s

Strategic conflict environment is the industry or industries in which it competes.’ Industry structure strongly influ- The publication of Carl Shapiro’s 1989 article,

confidently titled ‘ The Theory of Businessences the competitive rules of the game as well as the strategies potentially available to firms. Strategy,’ announced the emergence of a new

approach to business strategy, if not strategicIn the competitive forces model, five industry- level forces—entry barriers, threat of substitution, management. This approach utilizes the tools of

game theory to analyze the nature of competitivebargaining power of buyers, bargaining power of suppliers, and rivalry among industryinteraction between rival firms. The main thrust

of work in this tradition is to reveal how a firmincumbents—determine the inherent profit poten- tial of an industry or subsegment of an industry. can influence the behavior and actions of rival

firms and thus the market environment.4The approach can be used to help the firm find a position in an industry from which it can Examples of such moves are investment in

capacity (Dixit, 1980), R&D (Gilbert and New-best defend itself against competitive forces or influence them in its favor (Porter, 1980: 4). berry, 1982), and advertising (Schmalensee,

1983). To be effective, these strategic movesThis ‘five-forces’ framework provides a sys- tematic way of thinking about how competitive require irreversible commitments.5 The moves in

question will have no effect if they can beforces work at the industry level and how these forces determine the profitability of different costlessly undone. A key idea is that by manipu-

lating the market environment, a firm may beindustries and industry segments. The competitive forces framework also contains a number of able to increase its profits. underlying assumptions about the sources of com- petition and the nature of the strategy process.3 In competitive environments characterized by sustainable and

stable mobility and structural barriers, these forces mayTo facilitate comparisons with other approaches, become the determinants of industry-level profitability. How-we highlight several distinctive characteristics of ever, competitive advantage is more complex to ascertain in

the framework. environments of rapid technological change where specific assets owned by heterogeneous firms can be expected to playEconomic rents in the competitive forces a larger role in explaining rents.framework are monopoly rents (Teece, 1984). 4 The market environment is all factors that influence market

Firms in an industry earn rents when they areoutcomes (prices, quantities, profits) including the beliefs of customers and of rivals, the number of potential technologiessomehow able to impede the competitive forces employed, and the costs or speed with which a rival can( in either factor markets or product markets) enter the industry.

which tend to drive economic returns to zero.5 For an excellent discussion of committed competition in multiple contexts, see Ghemawat (1991 ).Available strategies are described in Porter

512 D. J. Teece, G. Pisano and A. Shuen

This literature, together with the contestability which the strategic conflict literature is relevant to strategic management. Firms that have aliterature (Baumol, Panzar, and Willig, 1982), has

led to a greater appreciation of the role of sunk tremendous cost or other competitive advantage vis-à-vis their rivals ought not be transfixed bycosts, as opposed to fixed costs, in determining

competitive outcomes. Strategic moves can also the moves and countermoves of their rivals. Their competitive fortunes will swing more on totalbe designed to influence rivals’ behavior through

signaling. Strategic signaling has been examined demand conditions, not on how competitors deploy and redeploy their competitive assets. Putin a number of contexts, including predatory

pricing (Kreps and Wilson, 1982a, 1982b ) and differently, when there are gross asymmetries in competitive advantage between firms, the resultslimit pricing (Milgrom and Roberts, 1982a,

1982b ). More recent treatments have emphasized of game-theoretic analysis are likely to be obvious and uninteresting. The stronger competitor willthe role of commitment and reputation (e.g.,

Ghemawat, 1991) and the benefits of firms simul- generally advance, even if disadvantaged by cer- tain information asymmetries. To be sure, incum-taneously pursuing competition and cooperation6

( Brandenburger and Nalebuff, 1995, 1996). bent firms can be undone by new entrants with a dramatic cost advantage, but no ‘gaming’ willIn many instances, game theory formalizes

long-standing intuitive arguments about various overturn that outcome. On the other hand, if firms’ competitive positions are more delicatelytypes of business behavior (e.g., predatory pric-

ing, patent races ), though in some instances it has balanced, as with Coke and Pepsi, and United Airlines and American Airlines, then strategicinduced a substantial change in the conventional

wisdom. But by rationalizing observed behavior conflict is of interest to competitive outcomes. Needless to say, there are many such circum-by reference to suitably designed games, in

explaining everything these models also explain stances, but they are rare in industries where there is rapid technological change and fast-shift-nothing, as they do not generate testable predic-

tions (Sutton, 1992). Many specific game- ing market circumstances. In short, where competitors do not have deep-theoretic models admit multiple equilibrium, and

a wide range of choice exists as to the design of seated competitive advantages, the moves and countermoves of competitors can often be use-the appropriate game form to be used. Unfortu-

nately, the results often depend on the precise fully formulated in game-theoretic terms. How- ever, we doubt that game theory can comprehen-specification chosen. The equilibrium in models

of strategic behavior crucially depends on what sively illuminate how Chrysler should compete against Toyota and Honda, or how United Air-one rival believes another rival will do in a

particular situation. Thus the qualitative features lines can best respond to Southwest Airlines since Southwest’s advantage is built on organizationalof the results may depend on the way price

competition is modeled (e.g., Bertrand or attributes which United cannot readily replicate.8

Indeed, the entrepreneurial side of strategy —howCournot) or on the presence or absence of stra- tegic asymmetries such as first-mover advantages. significant new rent streams are created and

protected—is largely ignored by the game-The analysis of strategic moves using game theory can be thought of as ‘dynamic’ in the theoretic approach.9 Accordingly, we find that

the approach, while important, is most relevantsense that multiperiod analyses can be pursued both intuitively and formally. However, we use the term ‘dynamic’ in this paper in a different sense, referring to situations where there is rapid

8 Thus even in the air transport industry game-theoretic formu-change in technology and market forces, and lations by no means capture all the relevant dimensions of

‘feedback’ effects on firms.7 competitive rivalry. United Airlines’ and United Express’s difficulties in competing with Southwest Airlines because ofWe have a particular view of the contexts in United’s inability to fully replicate Southwest’s operation capabilities is documented in Gittel (1995 ). 9 Important exceptions can be found in Brandenburger and Nalebuff ( 1996) such as their emphasis on the role of com-6 Competition and cooperation have also been analyzed ouside

of this tradition. See, for example, Teece (1992 ) and Link, plements. However, these insights do not flow uniquely from game theory and can be found in the organizational economicsTeece and Finan (1996 ).

7 Accordingly, both approaches are dynamic, but in very literature (e.g., Teece, 1986a, 1986b; de Figueiredo and Teece, 1996 ).different senses.

Dynamic Capabilities 513

when competitors are closely matched10 and the that may deter entry and raise prices above long- run costs, but because they have markedly lowerpopulation of relevant competitors and the iden-

tity of their strategic alternatives can be readily costs, or offer markedly higher quality or product performance. This approach focuses on the rentsascertained. Nevertheless, coupled with other

approaches it can sometimes yield powerful accruing to the owners of scarce firm-specific resources rather than the economic profits frominsights.

However, this research has an orientation that product market positioning.12 Competitive advan- tage lies ‘upstream’ of product markets and restswe are concerned about in terms of the implicit

framing of strategic issues. Rents, from a game- on the firm’s idiosyncratic and difficult-to- imitate resources.13theoretic perspective, are ultimately a result of

managers’ intellectual ability to ‘play the game.’ One can find the resources approach suggested by the earlier preanalytic strategy literature. AThe adage of the strategist steeped in this

approach is ‘do unto others before they do unto leading text of the 1960s (Learnedet al., 1969) noted that ‘the capability of an organization is itsyou.’ We worry that fascination with strategic

moves and Machiavellian tricks will distract man- demonstrated and potential ability to accomplish against the opposition of circumstance or compe-agers from seeking to build more enduring

sources of competitive advantage. The approach tition, whatever it sets out to do. Every organiza- tion has actual and potential strengths and weak-unfortunately ignores competition as a process

involving the development, accumulation, combi- nesses; it is important to try to determine what they are and to distinguish one from the other.’nation, and protection of unique skills and capa-

bilities. Since strategic interactions are what Thus what a firm can do is not just a function of the opportunities it confronts; it also dependsreceive focal attention, the impression one might

receive from this literature is that success in the on what resources the organization can muster. Learnedet al. proposed that the real key to amarketplace is the result of sophisticated plays

and counterplays, when this is generally not the company’s success or even to its future develop- ment lies in its ability to find or create ‘a com-case at all.11

In what follows, we suggest that building a petence that is truly distinctive.’14 This literature also recognized the constraints on firm behaviordynamic view of the business enterprise—

something missing from the two approaches we and, in particular, noted that one should not assume that management ‘can rise to anyhave so far identified—enhances the probability

of establishing an acceptable descriptive theory occasion.’ These insights do appear to keenly anticipate the resource-based approach that hasof strategy that can assist practitioners in the

building of long-run advantage and competitive since emerged, but they did not provide a theory or systematic framework for analyzing businessflexibility. Below, we discuss first the resource-

based perspective and then an extension we call strategies. Indeed, Andrews ( 1987: 46) noted that ‘much of what is intuitive in this process isthe dynamic capabilities approach. yet to be identified.’ Unfortunately, the academic literature on capabilities stalled for a couple of decades.MODELS OF STRATEGY

EMPHASIZING EFFICIENCY New impetus has been given to the resource- based approach by recent theoretical develop-

Resource-based perspective ments in organizational economics and in the theory of strategy, as well as by a growingThe resource-based approach sees firms with

superior systems and structures being profitable not because they engage in strategic investments

12 In the language of economics, rents flow from unique firm- specific assets that cannot readily be replicated, rather than from tactics which deter entry and keep competitors off10 When closely matched in an aggregate sense, they may

nevertheless display asymmetries which game theorists can balance. In short, rents are Ricardian. 13 Teece ( 1982: 46) saw the firm as having ‘a variety of endanalyze.

11 The strategic conflict literature also tends to focus prac- products which it can produce with its organizational tech- nology.’titioners on product market positioning rather than on

developing the unique assets which make possible superior14 Elsewhere Andrews ( 1987: 47) defined a distinctive com- petence as what an organization can do particularly well.product market positions ( Dierickx and Cool, 1989 ).

514 D. J. Teece, G. Pisano and A. Shuen

body of anecdotal and empirical literature15 that process.18 Quite simply, firms lack the organiza- tional capacity to develop new competenceshighlights the importance of firm-specific factors

in explaining firm performance. Cool and Schen- quickly (Dierickx and Cool, 1989). Secondly, some assets are simply not readily tradeable, fordel (1988) have shown that there are systematic

and significant performance differences among example, tacit know-how (Teece, 1976, 1980) and reputation ( Dierickx and Cool, 1989). Thus,firms which belong to the same strategic group

within the U.S. pharmaceutical industry. Rumelt resource endowments cannot equilibrate through factor input markets. Finally, even when an asset( 1991) has shown that intraindustry differences

in profits are greater than interindustry differences can be purchased, firms may stand to gain little by doing so. As Barney (1986) points out, unless ain profits, strongly suggesting the importance of

firm-specific factors and the relative unimportance firm is lucky, possesses superior information, or both, the price it pays in a competitive factorof industry effects.16 Jacobsen (1988) and Hansen

and Wernerfelt (1989) made similar findings. market will fully capitalize the rents from the asset. Given that in the resources perspective firmsA comparison of the resource-based approach

and the competitive forces approach (discussed possess heterogeneous and sticky resource bundles, the entry decision process suggested byearlier in the paper) in terms of their implications

for the strategy process is revealing. From the this approach is as follows: (1) identify your firm’s unique resources; (2) decide in which mar-first perspective, an entry decision looks roughly

as follows: (1) pick an industry (based on its kets those resources can earn the highest rents; and (3) decide whether the rents from those assets‘structural attractiveness’); (2) choose an entry

strategy based on conjectures about competitors’ are most effectively utilized by ( a) integrating into related market( s), (b) selling the relevantrational strategies; (3) if not already possessed,

acquire or otherwise obtain the requisite assets to intermediate output to related firms, or ( c) selling the assets themselves to a firm in related busi-compete in the market. From this perspective, the

process of identifying and developing the requi- nesses (Teece, 1980, 1982). The resource-based perspective puts both verti-site assets is not particularly problematic. The

process involves nothing more than choosing cal integration and diversification into a new stra- tegic light. Both can be viewed as ways of captur-rationally among a well-defined set of investment

alternatives. If assets are not already owned, they ing rents on scarce, firm-specific assets whose services are difficult to sell in intermediate mar-can be bought. The resource-based perspective is

strongly at odds with this conceptualization. kets ( Penrose, 1959; Williamson, 1975; Teece, 1980, 1982, 1986a, 1986b; Wernerfelt, 1984).From the resource-based perspective, firms are

heterogeneous with respect to their resources/ Empirical work on the relationship between per- formance and diversification by Wernerfelt andcapabilities/endowments. Further, resource endow-

ments are ‘sticky:’ at least in the short run, firms Montgomery ( 1988) provides evidence for this proposition. It is evident that the resource-basedare to some degree stuck with what they have

and may have to live with what they lack.17 This perspective focuses on strategies for exploiting existing firm-specific assets.stickiness arises for three reasons. First, business

development is viewed as an extremely complex However, the resource-based perspective also invites consideration of managerial strategies for developing new capabilities (Wernerfelt, 1984).

15 Studies of the automobile and other industries displayed Indeed, if control over scarce resources is thedifferences in organization which often underlay differences

amongst firms. See, for example, Womack, Jones, and Roos,source of economic profits, then it follows that 1991; Hayes and Clark, 1985; Barney, Spender and Reve,such issues as skill acquisition, the management 1994; Clark and Fujimoto, 1991; Henderson and Cockburn,

of knowledge and know-how (Shuen, 1994), and1994; Nelson, 1991; Levinthal and Myatt, 1994. 16 Using FTC line of business data, Rumelt showed that stablelearning become fundamental strategic issues. It industry effects account for only 8 percent of the variance inis in this second dimension, encompassing skill business unit returns. Furthermore, only about 40 percent of

acquisition, learning, and accumulation of organi-the dispersion in industry returns is due to stable industry effects. zational and intangible or ‘invisible’ assets (Itami 17 In this regard, this approach has much in common with recent work on organizational ecology ( e.g., Freeman and Boeker, 1984) and also on commitment ( Ghemawat, 1991: 17–25). 18 Capability development, however, is not really analyzed.

Dynamic Capabilities 515

and Roehl, 1987), that we believe lies the greatest are influenced by past choices. At any given point in time, firms must follow a certain trajectory orpotential for contributions to strategy. path of competence development. This path not only defines what choices are open to the firm

The dynamic capabilities approach: Overview today, but it also puts bounds around what its internal repertoire is likely to be in the future.The global competitive battles in high-technology

industries such as semiconductors, information Thus, firms, at various points in time, make long- term, quasi-irreversible commitments to certainservices, and software have demonstrated the need

for an expanded paradigm to understand how domains of competence.19

The notion that competitive advantage requirescompetitive advantage is achieved. Well-known companies like IBM, Texas Instruments, Philips, both the exploitation of existing internal and

external firm-specific capabilities, and developingand others appear to have followed a ‘resource- based strategy’ of accumulating valuable tech- new ones is partially developed in Penrose

(1959), Teece (1982), and Wernerfelt (1984).nology assets, often guarded by an aggressive intellectual property stance. However, this strat- However, only recently have researchers begun

to focus on the specifics of how some organiza-egy is often not enough to support a significant competitive advantage. Winners in the global tions first develop firm-specific capabilities and

how they renew competences to respond to shiftsmarketplace have been firms that can demonstrate timely responsiveness and rapid and flexible prod- in the business environment.20 These issues are

intimately tied to the firm’s business processes,uct innovation, coupled with the management capability to effectively coordinate and redeploy market positions, and expansion paths. Several

writers have recently offered insights and evi-internal and external competences. Not surpris- ingly, industry observers have remarked that com- dence on how firms can develop their capability

to adapt and even capitalize on rapidly changingpanies can accumulate a large stock of valuable technology assets and still not have many use- environments.21 The dynamic capabilities

approach seeks to provide a coherent frameworkful capabilities. We refer to this ability to achieve new forms which can both integrate existing conceptual and

empirical knowledge, and facilitate prescription.of competitive advantage as ‘dynamic capabili- ties’ to emphasize two key aspects that were not In doing so, it builds upon the theoretical foun-

dations provided by Schumpeter (1934), Penrosethe main focus of attention in previous strategy perspectives. The term ‘dynamic’ refers to the (1959), Williamson (1975, 1985), Barney (1986),

Nelson and Winter ( 1982), Teece (1988), andcapacity to renew competences so as to achieve congruence with the changing business environ- Teeceet al. (1994). ment; certain innovative responses are required when time-to-market and timing are critical, the rate of technological change is rapid, and theTOWARD A DYNAMIC CAPABILITIES

FRAMEWORKnature of future competition and markets difficult to determine. The term ‘capabilities’ emphasizes

Terminology the key role of strategic management in appropri- ately adapting, integrating, and reconfiguring In order to facilitate theory development and

intellectual dialogue, some acceptable definitionsinternal and external organizational skills, resources, and functional competences to match are desirable. We propose the following. the requirements of a changing environment.

One aspect of the strategic problem facing 19 Deciding, under significant uncertainty about future states

an innovating firm in a world of Schumpeterianof the world, which long-term paths to commit to and when to change paths is the central strategic problem confrontingcompetition is to identify difficult-to-imitate the firm. In this regard, the work of Ghemawat (1991 ) isinternal and external competences most likely to highly germane to the dynamic capabilities approach to

support valuable products and services. Thus, asstrategy. 20 See, for example, Iansiti and Clark (1994 ) and Hendersonargued by Dierickx and Cool (1989), choices ( 1994).about how much to spend (invest ) on different 21 See Hayeset al. ( 1988), Prahalad and Hamel ( 1990),

possible areas are central to the firm’s strategy.Dierickx and Cool (1989 ), Chandler ( 1990), and Teece ( 1993).However, choices about domains of competence

516 D. J. Teece, G. Pisano and A. Shuen

to which a core competence is distinctive depends Factors of production

on how well endowed the firm is relative to its competitors, and on how difficult it is for com-These are ‘undifferentiated’ inputs available in

disaggregate form in factor markets. By undiffer- petitors to replicate its competences. entiated we mean that they lack a firm-specific component. Land, unskilled labor, and capital are

Dynamic capabilities typical examples. Some factors may be available for the taking, such as public knowledge. In the We define dynamic capabilities as the firm’s

ability to integrate, build, and reconfigure internallanguage of Arrow, such resources must be ‘non- fugitive.’22 Property rights are usually well and external competences to address rapidly

changing environments. Dynamic capabilities thusdefined for factors of production. reflect an organization’s ability to achieve new and innovative forms of competitive advantage

Resources23 given path dependencies and market positions (Leonard-Barton, 1992).Resources are firm-specific assets that are difficult

if not impossible to imitate. Trade secrets and certain specialized production facilities and engi-

Products neering experience are examples. Such assets are difficult to transfer among firms because of trans- End products are the final goods and services

produced by the firm based on utilizing the com-actions costs and transfer costs, and because the assets may contain tacit knowledge. petences that it possesses. The performance

(price, quality, etc.) of a firm’s products relative to its competitors at any point in time will depend

Organizational routines/competences upon its competences (which over time depend on its capabilities).When firm-specific assets are assembled in inte-

grated clusters spanning individuals and groups so that they enable distinctive activities to be

Markets and strategic capabilities performed, these activities constitute organiza- tional routines and processes. Examples include Different approaches to strategy view sources of

wealth creation and the essence of the strategicquality, miniaturization, and systems integration. Such competences are typically viable across mul- problem faced by firms differently. The competi-

tive forces framework sees the strategic problemtiple product lines, and may extend outside the firm to embrace alliance partners. in terms of industry structure, entry deterrence,

and positioning; game-theoretic models view the strategic problem as one of interaction between

Core competences rivals with certain expectations about how each other will behave;25 resource-based perspectivesWe define those competences that define a firm’s

fundamental business as core. Core competences have focused on the exploitation of firm-specific assets. Each approach asks different, often com-must accordingly be derived by looking across

the range of a firm’s (and its competitors) prod- plementary questions. A key step in building a conceptual framework related to dynamic capa-ucts and services.24 The value of core com-

petences can be enhanced by combination with bilities is to identify the foundations upon which distinctive and difficult-to-replicate advantagesthe appropriate complementary assets. The degree can be built, maintained, and enhanced.

A useful way to vector in on the strategic22 Arrow ( 1996) defines fugitive resources as ones that can move cheaply amongst individuals and firms. elements of the business enterprise is first to 23 We do not like the term ‘resource’ and believe it isidentify what is not strategic. To be strategic, a misleading. We prefer to use the term firm-specific asset. We use it here to try and maintain links to the literature on the resource-based approach which we believe is important. 24 Thus Eastman Kodak’s core competence might be con- sidered imaging, IBM’s might be considered integrated data25 In sequential move games, each player looks ahead and

anticipates his rival’s future responses in order to reason backprocessing and service, and Motorola’s untethered communi- cations. and decide action, i.e., look forward, reason backward.

Dynamic Capabilities 517

capability must be honed to a user need26 ( so to coordinate activity.28 The very essence of most capabilities/competences is that they cannot bethere is a source of revenues), unique (so that

the products/services produced can be priced readily assembled through markets (Teece, 1982, 1986a; Zander and Kogut, 1995). If the abilitywithout too much regard to competition) and

difficult to replicate (so profits will not be com- to assemble competences using markets is what is meant by the firm as a nexus of contractspeted away). Accordingly, any assets or entity

which are homogeneous and can be bought and (Fama, 1980), then we unequivocally state that the firm about which we theorize cannot be use-sold at an established price cannot be all that

strategic (Barney, 1986). What is it, then, about fully modeled as a nexus of contracts. By ‘con- tract’ we are referring to a transaction undergirdedfirms which undergirds competitive advantage?

To answer this, one must first make some by a legal agreement, or some other arrangement which clearly spells out rights, rewards, andfundamental distinctions between markets and

internal organization (firms). The essence of the responsibilities. Moreover, the firm as a nexus of contracts suggests a series of bilateral contractsfirm, as Coase (1937) pointed out, is that it

displaces market organization. It does so in the orchestrated by a coordinator. Our view of the firm is that the organization takes place in a moremain because inside the firms one can organize

certain types of economic activity in ways one multilateral fashion, with patterns of behavior and learning being orchestrated in a much morecannot using markets. This is not only because

of transaction costs, as Williamson (1975, 1985) decentralized fashion, but with a viable head- quarters operation.emphasized, but also because there are many

types of arrangements where injecting high-pow- The key point, however, is that the properties of internal organization cannot be replicated byered ( market like) incentives might well be quite

destructive of cooperative activity and learning.27 a portfolio of business units amalgamated just through formal contracts as many distinctiveInside an organization, exchange cannot take

place in the same manner that it can outside an elements of internal organization simply cannot be replicated in the market.29 That is, entrepre-organization, not just because it might be destruc-

tive to provide high-powered individual incen- neurial activity cannot lead to the immediate rep- lication of unique organizational skills throughtives, but because it is difficult if not impossible

to tightly calibrate individual contribution to a simply entering a market and piecing the parts together overnight. Replication takes time, andjoint effort. Hence, contrary to Arrow’s (1969)

view of firms as quasi markets, and the task the replication of best practice may be illusive. Indeed, firm capabilities need to be understoodof management to inject markets into firms, we

recognize the inherent limits and possible counter- not in terms of balance sheet items, but mainly in terms of the organizational structures andproductive results of attempting to fashion firms

into simply clusters of internal markets. In parti- managerial processes which support productive activity. By construction, the firm’s balance sheetcular, learning and internal technology transfer

may well be jeopardized. contains items that can be valued, at least at original market prices (cost). It is necessarily theIndeed, what is distinctive about firms is that

they are domains for organizing activity in a case, therefore, that the balance sheet is a poor shadow of a firm’s distinctive competences.30nonmarket-like fashion. Accordingly, as we dis-

cuss what is distinctive about firms, we stress competences/capabilities which are ways of28 We see the problem of market contracting as a matter of

coordination as much as we see it a problem of opportunismorganizing and getting things done which cannot in the fact of contractual hazards. In this sense, we arebe accomplished merely by using the price system consonant with both Richardson ( 1960) and Williamson ( 1975, 1985). 29 As we note in Teeceet al. ( 1994), the conglomerate offers few if any efficiencies because there is little provided by26 Needless to say, users need not be the current customers

of the enterprise. Thus a capability can be the basis for the conglomerate form that shareholders cannot obtain for themselves simply by holding a diversified portfolio of stocks.diversification into new product markets.

27 Indeed, the essence of internal organization is that it is a30 Owners’ equity may reflect, in part, certain historic capabili- ties. Recently, some scholars have begun to attempt to meas-domain of unleveraged or low-powered incentives. By unlever-

aged we mean that rewards are determined at the group or ure organizational capability using financial statement data. See Baldwin and Clark (1991 ) and Lev and Sougiannisorganization level, not primarily at the individual level, in an

effort to encourage team behavior, not individual behavior. ( 1992).

518 D. J. Teece, G. Pisano and A. Shuen

That which is distinctive cannot be bought and focuses on replication and imitation, as it is these phenomena which determine how readily a com-sold short of buying the firm itself, or one or

more of its subunits. petence or capability can be cloned by competi- tors, and therefore distinctiveness of its com-There are many dimensions of the business

firm that must be understood if one is to grasp petences and the durability of its advantage. The firm’s processes and positions collectivelyfirm-level distinctive competences/capabilities. In

this paper we merely identify several classes of encompass its competences and capabilities. A hierarchy of competences/capabilities ought to befactors that will help determine a firm’s distinc-

tive competence and dynamic capabilities. We recognized, as some competences may be on the factory floor, some in the R&D labs, some in theorganize these in three categories: processes, po-

sitions, and paths. The essence of competences executive suites, and some in the way everything is integrated. A difficult-to-replicate or difficult-and capabilities is embedded in organizational

processes of one kind or another. But the content to-imitate competence was defined earlier as a distinctive competence. As indicated, the key fea-of these processes and the opportunities they

afford for developing competitive advantage at ture of distinctive competence is that there is not a market for it, except possibly through the mar-any point in time are shaped significantly by the

assets the firm possesses (internal and market) ket for business units. Hence competences and capabilities are intriguing assets as they typicallyand by the evolutionary path it has

adopted/inherited. Hence organizational processes, must be built because they cannot be bought. shaped by the firm’s asset positions and molded by its evolutionary and co-evolutionary paths,

Organizational and managerial processes explain the essence of the firm’s dynamic capa- bilities and its competitive advantage. Organizational processes have three roles:

coordination/integration ( a static concept); learn- ing (a dynamic concept); and reconfiguration (a

Processes, positions, and paths transformational concept). We discuss each in turn.We thus advance the argument that the competi-

tive advantage of firms lies with its managerial and organizational processes, shaped by itsCoordination/integration. While the price sys-

tem supposedly coordinates the economy,32 man-(specific) asset position, and the paths available to it.31 By managerial and organizational proc- agers coordinate or integrate activity inside the

firm. How efficiently and effectively internalesses, we refer to the way things are done in the firm, or what might be referred to as its routines, coordination or integration is achieved is very

important ( Aoki, 1990).33 Likewise for externalor patterns of current practice and learning. By position we refer to its current specific endow- coordination.34 Increasingly, strategic advantage

requires the integration of external activities andments of technology, intellectual property, com- plementary assets, customer base, and its external technologies. The growing literature on strategic relations with suppliers and complementors. By paths we refer to the strategic alternatives avail-

32 The coordinative properties of markets depend on pricesable to the firm, and the presence or absence of being “sufficient” upon which to base resource allocationincreasing returns and attendant path depen- decisions.

dencies. 33 Indeed, Ronald Coase, author of the pathbreaking 1937 article ‘The nature of the firm,’ which focused on the costsOur focus throughout is on asset structures for of organizational coordination inside the firm as compared towhich no ready market exists, as these are the across the market, half a century later has identified as critical

only assets of strategic interest. A final sectionthe understanding of ‘why the costs of organizing particular activities differs among firms’ ( Coase, 1988: 47). We argue that a firm’s distinctive ability needs to be understood as a reflection of distinctive organizational or coordinative capabili-31 We are implicitly saying that fixed assets, like plant and

equipment which can be purchased off-the-shelf by all industry ties. This form of integration ( i.e., inside business units ) is different from the integration between business units; theyparticipants, cannot be the source of a firm’s competitive

advantage. In asmuch as financial balance sheets typically could be viable on a stand-alone basis ( external integration). For a useful taxonomy, see Iansiti and Clark (1994 ).reflect such assets, we point out that the assets that matter

for competitive advantage are rarely reflected in the balance34 Shuen ( 1994) examines the gains and hazards of the tech- nology make-vs.-buy decision and supplier codevelopment.sheet, while those that do not are.

Dynamic Capabilities 519

alliances, the virtual corporation, and buyer– gest that productive systems display high interde- pendency, and that it may not be possible tosupplier relations and technology collaboration

evidences the importance of external integration change one level without changing others. This appears to be true with respect to the ‘leanand sourcing.

There is some field-based empirical research production’ model (Womacket al., 1991) which has now transformed the Taylor or Ford modelthat provides support for the notion that the way

production is organized by management inside of manufacturing organization in the automobile industry.36 Lean production requires distinctivethe firm is the source of differences in firms’

competence in various domains. For example, shop floor practices and processes as well as distinctive higher-order managerial processes. PutGarvin’s (1988) study of 18 room air-condition-

ing plants reveals that quality performance was differently, organizational processes often display high levels of coherence, and when they do,not related to either capital investment or the

degree of automation of the facilities. Instead, replication may be difficult because it requires systemic changes throughout the organization andquality performance was driven by special organi-

zational routines. These included routines for also among interorganizational linkages, which might be very hard to effectuate. Put differently,gathering and processing information, for linking

customer experiences with engineering design partial imitation or replication of a successful model may yield zero benefits.37choices, and for coordinating factories and

component suppliers.35 The work of Clark and Fujimoto (1991) on project development in the

36 Fujimoto (1994: 18–20) describes key elements as they automobile industry also illustrates the role playedexisted in the Japanese auto industry as follows: ‘The typical

volume production system of effective Japanese makers ofby coordinative routines. Their study reveals a the 1980s (e.g., Toyota ) consists of various intertwinedsignificant degree of variation in how different elements that might lead to competitive advantages. Just-in-

firms coordinate the various activities required toTime ( JIT), Jidoka (automatic defect detection and machine stop), Total Quality Control ( TQC), and continuous improve-bring a new model from concept to market. These ment ( Kaizen) are often pointed out as its core subsystems.differences in coordinative routines and capabili- The elements of such a system include inventory reduction

ties seem to have a significant impact on suchmechanisms by Kanban system; levelization of production volume and product mix (heijunka ); reduction of ‘muda’performance variables as development cost, devel- ( non-value adding activities), ‘mura’ (uneven pace ofopment lead times, and quality. Furthermore, production) and muri ( excessive workload); production plans

Clark and Fujimoto tended to find significantbased on dealers’ order volume (genyo seisan); reduction of die set-up time and lot size in stamping operation; mixedfirm-level differences in coordination routines and model assembly; piece-by-piece transfer of parts betweenthese differences seemed to have persisted for a machines (ikko-nagashi ); flexible task assignment for volume

long time. This suggests that routines related tochanges and productivity improvement ( shojinka); multi-task job assignment along the process flow (takotei-mochi); U-coordination are firm-specific in nature. shape machine layout that facilitates flexible and multipleAlso, the notion that competence/capability is task assignment, on-the-spot inspection by direct workers

embedded in distinct ways of coordinating and( tsukurikomi); fool-proof prevention of defects (poka-yoke ); real-time feedback of production troubles ( andon); assemblycombining helps to explain how and why seem- line stop cord; emphasis on cleanliness, order and disciplineingly minor technological changes can have on the shop floor ( 5-S); frequent revision of standard operating

devastating impacts on incumbent firms’ abilitiesprocedures by supervisors; quality control circles; standardized tools for quality improvement ( e.g., 7 tools for QC, QCto compete in a market. Henderson and Clark story); worker involvement in preventive maintenance (Total( 1990), for example, have shown that in- Productive Maintenance ); low cost automation or semi-auto-

cumbments in the photolithographic equipmentmation with just-enough functions ); reduction of process steps for saving of tools and dies, and so on. The human-resourceindustry were sequentially devasted by seemingly management factors that back up the above elements includeminor innovations that, nevertheless, had major stable employment of core workers ( with temporary workers

impacts on how systems had to be configured.in the periphery); long-term training of multi-skilled ( multi- task) workers; wage system based in part on skill accumu-They attribute these difficulties to the fact that lation; internal promotion to shop floor supervisors; coopera-systems-level or ‘architectural’ innovations often tive relationships with labor unions; inclusion of production

require new routines to integrate and coordinatesupervisors in union members; generally egalitarian policies for corporate welfare, communication and worker motivation.engineering tasks. These findings and others sug- Parts procurement policies are also pointed out often as a source of the competitive advantage. 37 For a theoretical argument along these lines, see Milgrom35 Garvin ( 1994 ) provides a typology of organizational

processes. and Roberts (1990 ).

520 D. J. Teece, G. Pisano and A. Shuen

The notion that there is a certain rationality or which repetition and experimentation enable tasks to be performed better and quicker. It also enablescoherence to processes and systems is not quite

the same concept as corporate culture, as we new production opportunities to be identified.39

In the context of the firm, if not more generally,understand the latter. Corporate culture refers to the values and beliefs that employees hold; cul- learning has several key characteristics. First,

learning involves organizational as well as indi-ture can be ade facto governance system as it mediates the behavior of individuals and econo- vidual skills.40 While individual skills are of rel-

evance, their value depends upon their employ-mizes on more formal administrative methods. Rationality or coherence notions are more akin ment, in particular organizational settings.

Learning processes are intrinsically social andto the Nelson and Winter (1982) notion of organi- zational routines. However, the routines concept collective and occur not only through the imi-

tation and emulation of individuals, as withis a little too amorphous to properly capture the congruence amongst processes and between teacher–student or master–apprentice, but also

because of joint contributions to the understand-processes and incentives that we have in mind. Consider a professional service organization like ing of complex problems.41 Learning requires

common codes of communication and coordinatedan accounting firm. If it is to have relatively high-powered incentives that reward individual search procedures. Second, the organizational

knowledge generated by such activity resides inperformance, then it must build organizational processes that channel individual behavior; if it new patterns of activity, in ‘routines,’ or a new

logic of organization. As indicated earlier, rou-has weak or low-powered incentives, it must find symbolic ways to recognize the high performers, tines are patterns of interactions that represent

successful solutions to particular problems. Theseand it must use alternative methods to build effort and enthusiasm. What one may think of as styles patterns of interaction are resident in group

behavior, though certain subroutines may be resi-of organization in fact contain necessary, not discretionary, elements to achieve performance. dent in individual behavior. The concept of

dynamic capabilities as a coordinative manage-Recognizing the congruences and complemen- tarities among processes, and between processes ment process opens the door to the potential

for interorganizational learning. Researchers (Dozand incentives, is critical to the understanding of organizational capabilities. In particular, they can and Shuen, 1990; Mody, 1993) have pointed out

that collaborations and partnerships can be ahelp us explain why architectural and radical innovations are so often introduced into an indus- vehicle for new organizational learning, helping

firms to recognize dysfunctional routines, andtry by new entrants. The incumbents develop distinctive organizational processes that cannot preventing strategic blindspots. support the new technology, despite certain overt similarities between the old and the new. TheReconfiguration and transformation.In rapidly

changing environments, there is obviously valuefrequent failure of incumbents to introduce new technologies can thus be seen as a consequence in the ability to sense the need to reconfigure

the firm’s asset structure, and to accomplish theof the mismatch that so often exists between the set of organizational processes needed to support necessary internal and external transformation

(Amit and Schoemaker, 1993; Langlois, 1994).the conventional product/service and the require- ments of the new. Radical organizational re- This requires constant surveillance of markets and

technologies and the willingness to adopt bestengineering will usually be required to support the new product, which may well do better practice. In this regard, benchmarking is of con- embedded in a separate subsidiary where a new set of coherent organizatonal processes can be39 For a useful review and contribution, see Levitt and

March ( 1988).fashioned.38 40 Levinthal and March, 1993. Mahoney (1992) and Mahoney and Pandian (1995) suggest that both resources and mental

Learning. Perhaps even more important thanmodels are intertwined in firm-level learning. 41 There is a large literature on learning, although only aintegration is learning. Learning is a process by small fraction of it deals with organizational learning. Relevant contributors include Levitt and March ( 1988), Argyris and Schon (1978 ), Levinthal and March ( 1981), Nelson and Winter (1982 ), and Leonard-Barton ( 1995).38 See Abernathy and Clark ( 1985).

Dynamic Capabilities 521

siderable value as an organized process for Prior commercialization activities require and enable firms to build such complementaritiesaccomplishing such ends (Camp, 1989). In

dynamic environments, narcissistic organizations (Teece, 1986b ). Such capabilities and assets, while necessary for the firm’s established activi-are likely to be impaired. The capacity to recon-

figure and transform is itself a learned organiza- ties, may have other uses as well. These assets typically lie downstream. New products and proc-tional skill. The more frequently practiced, the

easier accomplished. esses either can enhance or destroy the value of such assets (Tushman, Newman, and Romanelli,Change is costly and so firms must develop

processes to minimize low pay-off change. The 1986). Thus the development of computers enhanced the value of IBM’s direct sales forceability to calibrate the requirements for change

and to effectuate the necessary adjustments would in office products, while disk brakes rendered useless much of the auto industry’s investmentappear to depend on the ability to scan the

environment, to evaluate markets and competitors, in drum brakes. and to quickly accomplish reconfiguration and transformation ahead of competition. Decentrali-Financial assets. In the short run, a firm’s cash

position and degree of leverage may have stra-zation and local autonomy assist these processes. Firms that have honed these capabilities are tegic implications. While there is nothing more

fungible than cash, it cannot always be raisedsometimes referred to as ‘high-flex’. from external markets without the dissemination of considerable information to potential investors.

Positions Accordingly, what a firm can do in short order is often a function of its balance sheet. In theThe strategic posture of a firm is determined not

only by its learning processes and by the coher- longer run, that ought not be so, as cash flow ought be more determinative.ence of its internal and external processes and

incentives, but also by its specific assets. By specific assets we mean for example its special-Reputational assets.Firms, like individuals,

have reputations. Reputations often summarize aized plant and equipment. These include its difficult-to-trade knowledge assets and assets good deal of information about firms and shape

the responses of customers, suppliers, and com-complementary to them, as well as its reputational and relational assets. Such assets determine its petitors. It is sometimes difficult to disentangle

reputation from the firm’s current asset and mar-competitive advantage at any point in time. We identify several illustrative classes. ket position. However, in our view, reputational

assets are best viewed as an intangible asset that enables firms to achieve various goals in theTechnological assets.While there is an emerg-

ing market for know-how (Teece, 1981), much market. Its main value is external, since what is critical about reputation is that it is a kind oftechnology does not enter it. This is either

because the firm is unwilling to sell it42 or summary statistic about the firm’s current assets and position, and its likely future behavior.because of difficulties in transacting in the market

for know-how ( Teece, 1980). A firm’s techno- Because there is generally a strong asymmetry between what is known inside the firm and whatlogical assets may or may not be protected by

the standard instruments of intellectual property is known externally, reputations may sometimes be more salient than the true state of affairs, inlaw. Either way, the ownership protection and

utilization of technological assets are clearly key the sense that external actors must respond to what they know rather than what is knowable.differentiators among firms. Likewise for com-

plementary assets. Structural assets. The formal and informal structure of organizations and their external link-Complementary assets.Technological inno-

vations require the use of certain related assets ages have an important bearing on the rate and direction of innovation, and how competencesto produce and deliver new products and services. and capabilities co-evolve (Argyres, 1995; Teece, 1996). The degree of hierarchy and the level of42 Managers often evoke the ‘crown jewels’ metaphor. That

is, if the technology is released, the kingdom will be lost. vertical and lateral integration are elements of

522 D. J. Teece, G. Pisano and A. Shuen

firm-specific structure. Distinctive governanceOrganizational boundaries. An important dimension of ‘position’ is the location of a firm’smodes can be recognized (e.g., multiproduct, inte-

grated firms; high ‘flex’ firms; virtual corpora- boundaries. Put differently, the degree of inte- gration (vertical, lateral, and horizontal) is oftions; conglomerates), and these modes support

different types of innovation to a greater or lesser quite some significance. Boundaries are not only significant with respect to the technological anddegree. For instance, virtual structures work well

when innovation is autonomous; integrated struc- complementary assets contained within, but also with respect to the nature of the coordination thattures work better for systemic innovations. can be achieved internally as compared to through markets. When specific assets or poorly protectedInstitutional assets. Environments cannot be

defined in terms of markets alone. While public intellectual capital are at issue, pure market arrangements expose the parties to recontractingpolicies are usually recognized as important in

constraining what firms can do, there is a ten- hazards or appropriability hazards. In such cir- cumstances, hierarchical control structures maydency, particularly by economists, to see these

as acting through markets or through incentives. work better than pure arms-length contracts.44

However, institutions themselves are a critical element of the business environment. Regulatory

Paths systems, as well as intellectual property regimes, tort laws, and antitrust laws, are also part of the environment. So is the system of higher educationPath dependencies.Where a firm can go is a

function of its current position and the pathsand national culture. There are significant national differences here, which is just one of the reasons ahead. Its current position is often shaped by

the path it has traveled. In standard economicsgeographic location matters (Nelson, 1994). Such assets may not be entirely firm specific; firms of textbooks, firms have an infinite range of technol-

ogies from which they can choose and marketsdifferent national and regional origin may have quite different institutional assets to call upon they can occupy. Changes in product or factor

prices will be responded to instantaneously, withbecause their institutional/policy settings are so different. technologies moving in and out according to

value maximization criteria. Only in the short run are irreversibilities recognized. Fixed costs— suchMarket (structure) assets.Product market po-

sition matters, but it is often not at all determina- as equipment and overheads—cause firms to price below fully amortized costs but never constraintive of the fundamental position of the enterprise

in its external environment. Part of the problem future investment choices. ‘Bygones are bygones.’ Path dependencies are simply not recognized.lies in defining the market in which a firm com-

petes in a way that gives economic meaning. This is a major limitation of microeconomic theory.More importantly, market position in regimes of

rapid technological change is often extremely The notion of path dependencies recognizes that ‘history matters.’ Bygones are rarelyfragile. This is in part because time moves on a

different clock in such environments.43 Moreover, bygones, despite the predictions of rational actor theory. Thus a firm’s previous investments andthe link between market share and innovation has

long been broken, if it ever existed (Teece, 1996). All of this is to suggest that product market

44 Williamson (1996: 102–103) has observed, failures of coor-position, while important, is too often overplayed. dination may arise because ‘parties that bear a long termStrategy should be formulated with regard to the bilateral dependency relationship to one another must recog-

more fundamental aspects of firm performance,nize that incomplete contracts require gap filling and some- times get out of alignment. Although it is always in thewhich we believe are rooted in competences and collective interest of autonomous parties to fill gaps, correctcapabilities and shaped by positions and paths. errors, and affect efficient realignments, it is also the case that the distribution of the resulting gains is indeterminate. Self-interested bargaining predictably obtains. Such bargaining is itself costly. The main costs, however, are that transactions43 For instance, an Internet year might well be thought of as

equivalent to 10 years on many industry clocks, because as are maladapted to the environment during the bargaining interval. Also, the prospect of ex post bargaining invites exmuch change occurs in the Internet business in a year that

occurs in say the auto industry in a decade. ante prepositioning of an inefficient kind.’

Dynamic Capabilities 523

its repertoire of routines (its ‘history’) constrain gically through technology-sponsoring activities.46

The first type of competition is not unlike biologi-its future behavior.45 This follows because learn- ing tends to be local. That is, opportunities for cal competition amongst species, although it can

be sharpened by managerial activities thatlearning will be ‘close in’ to previous activities and thus will be transaction and production spe- enhance the performance of products and proc-

esses. The reality is that companies with the bestcific (Teece, 1988). This is because learning is often a process of trial, feedback, and evaluation. products will not always win, as chance events

may cause ‘lock-in’ on inferior technologiesIf too many parameters are changed simul- taneously, the ability of firms to conduct mean- (Arthur, 1983) and may even in special cases

generate switching costs for consumers. However,ingful natural quasi experiments is attenuated. If many aspects of a firm’s learning environment while switching costs may favor the incumbent,

in regimes of rapid technological change switch-change simultaneously, the ability to ascertain cause–effect relationships is confounded because ing costs can become quickly swamped by

switching benefits. Put differently, new productscognitive structures will not be formed and rates of learning diminish as a result. One implication employing different standards often appear with

alacrity in market environments experiencingis that many investments are much longer term than is commonly thought. rapid technological change, and incumbents can

be readily challenged by superior products andThe importance of path dependencies is ampli- fied where conditions of increasing returns to services that yield switching benefits. Thus the

degree to which switching costs cause ‘lock-in’adoption exist. This is a demand-side phenom- enon, and it tends to make technologies and is a function of factors such as user learning,

rapidity of technological change, and the amountproducts embodying those technologies more attractive the more they are adopted. Attractive- of ferment in the competitive environment. ness flows from the greater adoption of the prod- uct amongst users, which in turn enables them toTechnological opportunities. The concept of

path dependencies is given forward meaningbecome more developed and hence more useful. Increasing returns to adoption has many sources through the consideration of an industry’s techno-

logical opportunities. It is well recognized thatincluding network externalities ( Katz and Shapiro, 1985), the presence of complementary assets how far and how fast a particular area of indus-

trial activity can proceed is in part due to the( Teece, 1986b ) and supporting infrastructure ( Nelson, 1996), learning by using (Rosenberg, technological opportunities that lie before it. Such

opportunities are usually a lagged function of1982), and scale economies in production and distribution. Competition between and amongst foment and diversity in basic science, and the

rapidity with which new scientific breakthroughstechnologies is shaped by increasing returns. Early leads won by good luck or special circum- are being made.

However, technological opportunities may notstances (Arthur, 1983) can become amplified by increasing returns. This is not to suggest that be completely exogenous to industry, not only

because some firms have the capacity to engagefirst movers necessarily win. Because increasing returns have multiple sources, the prior posi- in or at least support basic research, but also

because technological opportunities are often fedtioning of firms can affect their capacity to exploit increasing returns. Thus, in Mitchell’s (1989) by innovative activity itself. Moreover, the recog-

nition of such opportunities is affected by thestudy of medical diagnostic imaging, firms already controlling the relevant complementary assets could in theory start last and finish first.

46 Because of huge uncertainties, it may be extremely difficult In the presence of increasing returns, firms canto determine viable strategies early on. Since the rules of the

game and the identity of the players will be revealed onlycompete passively, or they may compete strate- after the market has begun to evolve, the pay-off is likely to lie with building and maintaining organizational capabilities that support flexibility. For example, Microsoft’s recent about- face and vigorous pursuit of Internet business once the Net- Scape phenomenon became apparent is impressive, not so much because it perceived the need to change strategy, but because of its organizational capacity to effectuate a stra-45 For further development, see Bercovitz, de Figueiredo, and

Teece, 1996. tegic shift.

524 D. J. Teece, G. Pisano and A. Shuen

organizational structures that link the institutions approaches to the firm and to strategy.49 More- over, the agency theoretic view of the firm as aengaging in basic research (primarily the

university) to the business enterprise. Hence, the nexus of contracts would put no weight on proc- esses, positions, and paths. While agencyexistence of technological opportunities can be

quite firm specific. approaches to the firm may recognize that oppor- tunism and shirking may limit what a firm canImportant for our purposes is the rate and

direction in which relevant scientific frontiers are do, they do not recognize the opportunities and constraints imposed by processes, positions, andbeing rolled back. Firms engaging in R&D may

find the path dead ahead closed off, though break- paths. Moreover, the firm in our conceptualization isthroughs in related areas may be sufficiently close

to be attractive. Likewise, if the path dead ahead much more than the sum of its parts—or a team tied together by contracts.50 Indeed, to someis extremely attractive, there may be no incentive

for firms to shift the allocation of resources away extent individuals can be moved in and out of organizations and, so long as the internal proc-from traditional pursuits. The depth and width of

technological opportunities in the neighborhood esses and structures remain in place, performance will not necessarily be impaired. A shift in theof a firm’s prior research activities thus are likely

to impact a firm’s options with respect to both environment is a far more serious threat to the firm than is the loss of key individuals, as individ-the amount and level of R&D activity that it

can justify. In addition, a firm’s past experience uals can be replaced more readily than organiza- tions can be transformed. Furthermore, theconditions the alternatives management is able to

perceive. Thus, not only do firms in the same dynamic capabilities view of the firm would sug- gest that the behavior and performance of parti-industry face ‘menus’ with different costs associa-

ted with particular technological choices, they cular firms may be quite hard to replicate, even if its coherence and rationality are observable.also are looking at menus containing different

choices.47 This matter and related issues involving repli- cation and imitation are taken up in the section that follows.

Assessment

The essence of a firm’s competence and dynamic Replicability and imitatability of

capabilities is presented here as being resident in organizational processes and positions

the firm’s organizational processes, that are in turn shaped by the firm’s assets (positions) and its Thus far, we have argued that the competences

and capabilities (and hence competitive advantage)evolutionary path. Its evolutionary path, despite managerial hubris that might suggest otherwise, of a firm rest fundamentally on processes, shaped

by positions and paths. However, competencesis often rather narrow.48 What the firm can do and where it can go are thus rather constrained can provide competitive advantage and generate

rents only if they are based on a collection ofby its positions and paths. Its competitors are likewise constrained. Rents (profits) thus tend to routines, skills, and complementary assets that are

difficult to imitate.51 A particular set of routinesflow not just from the asset structure of the firm and, as we shall see, the degree of its imitability, can lose their value if they support a competence

which no longer matters in the marketplace, orbut also by the firm’s ability to reconfigure and transform. if they can be readily replicated or emulated by

competitors. Imitation occurs when firms discoverThe parameters we have identified for de- termining performance are quite different from and simply copy a firm’s organizational routines

and procedures. Emulation occurs when firmsthose in the standard textbook theory of the firm, and in the competitive forces and strategic conflict

49 In both the firm is still largely a black box. Certainly, little or no attention is given to processes, positions, and paths.47 This is a critical element in Nelson and Winter’s (1982 )

view of firms and technical change. 50 See Alchian and Demsetz (1972 ). 51 We call such competences distinctive. See also Dierickx48 We also recognize that the processes, positions, and paths of

customers also matter. See our discussion above on increasing and Cool (1989 ) for a discussion of the characteristics of assets which make them a source of rents.returns, including customer learning and network externalities.

Dynamic Capabilities 525

discover alternative ways of achieving the same Some routines and competences seem to be attributable to local or regional forces that shapefunctionality.52

firms’ capabilities at early stages in their lives. Porter (1990), for example, shows that differences

Replication in local product markets, local factor markets, and institutions play an important role in shapingTo understand imitation, one must first understand

replication. Replication involves transferring or competitive capabilities. Differences also exist within populations of firms from the same coun-redeploying competences from one concrete eco-

nomic setting to another. Since productive knowl- try. Various studies of the automobile industry, for example, show that not all Japanese auto-edge is embodied, this cannot be accomplished

by simply transmitting information. Only in those mobile companies are top performers in terms of quality, productivity, or product developmentinstances where all relevant knowledge is fully

codified and understood can replication be col- (see, for example, Clark and Fujimoto, 1991). The role of firm-specific history has been high-lapsed into a simple problem of information trans-

fer. Too often, the contextual dependence of ori- lighted as a critical factor explaining such firm- level (as opposed to regional or national-level)ginal performance is poorly appreciated, so unless

firms have replicated their systems of productive differences (Nelson and Winter, 1982). Repli- cation in a different context may thus be ratherknowledge on many prior occasions, the act of

replication is likely to be difficult (Teece, 1976). difficult. At least two types of strategic value flowIndeed, replication and transfer are often impos-

sible absent the transfer of people, though this from replication. One is the ability to support geographic and product line expansion. To thecan be minimized if investments are made to

convert tacit knowledge to codified knowledge. extent that the capabilities in question are rel- evant to customer needs elsewhere, replicationOften, however, this is simply not possible.

In short, competences and capabilities, and the can confer value.55 Another is that the ability to replicate also indicates that the firm has theroutines upon which they rest, are normally rather

difficult to replicate.53 Even understanding what foundations in place for learning and improve- ment. Considerable empirical evidence supportsall the relevant routines are that support a parti-

cular competence may not be transparent. Indeed, the notion that theunderstanding of processes, both in production and in management, is theLippman and Rumelt (1992) have argued that

some sources of competitive advantage are so key to process improvement. In short, an organization cannot improve that which it doescomplex that the firm itself, let alone its competi-

tors, does not understand them.54 As Nelson and notunderstand. Deep process understanding is often required to accomplish codification.Winter (1982) and Teece (1982) have explained,

many organizational routines are quite tacit in Indeed, ifknowledge is highly tacit, it indicates that underlying structures are not well under-nature. Imitation can also be hindered by the fact

few routines are ‘stand-alone;’ coherence may stood, which limits learning because scientific and engineering principles cannot be as system-require that a change in one set of routines in

one part of the firm (e.g., production) requires atically applied.56 Instead, learning is confined to proceeding through trial and error, and thechanges in some other part (e.g., R&D).

55 Needless to say, there are many examples of firms rep-52 There is ample evidence that a given type of competence ( e.g., quality) can be supported by different routines and licating their capabilities inappropriately by applying extant

routines to circumstances where they may not be applicable,combinations of skills. For example, the Garvin ( 1988) and Clark and Fujimoto ( 1991) studies both indicate that there e.g., Nestle’s transfer of developed-country marketing methods

for infant formula to the Third World ( Hartley, 1989). A keywas no one ‘formula’ for achieving either high quality or high product development performance. strategic need is for firms to screen capabilities for their

applicability to new environments.53 See Szulanski’s ( 1995) discussion of the intrafirm transfer of best practice. He quotes a senior vice president of Xerox56 Different approaches to learning are required depending on

the depth of knowledge. Where knowledge is less articulatedas saying ‘you can see a high performance factory or office, but it just doesn’t spread. I don’t know why.’ Szulanski also and structured, trial and error and learning-by-doing are

necessary, whereas in mature environments where the underly-discusses the role of benchmarking in facilitating the transfer of best practice. ing engineering science is better understood, organizations

can undertake more deductive approaches or what Pisano54 If so, it is our belief that the firm’s advantage is likely to fade, as luck does run out. ( 1994) refers to as ‘learning-before-doing.’

526 D. J. Teece, G. Pisano and A. Shuen

leverage that might otherwise come from the are thus more protectable if there is no need to expose them in contexts where competitors canapplication of scientific theory is denied. learn about them.

One should not, however, overestimate the Imitation

overall importance of intellectual property protec- tion; yet it presents a formidable imitation barrierImitation is simply replication performed by a

competitor. If self-replication is difficult, imitation in certain particular contexts. Intellectual property protection is not uniform across products, proc-is likely to be harder. In competitive markets,

it is the ease of imitation that determines the esses, and technologies, and is best thought of as islands in a sea of open competition. If one issustainability of competitive advantage. Easy imi-

tation implies the rapid dissipation of rents. not able to place the fruits of one’s investment, ingenuity, or creativity on one or more of theFactors that make replication difficult also

make imitation difficult. Thus, the more tacit the islands, then one indeed is at sea. We use the term appropriability regimes tofirm’s productive knowledge, the harder it is to

replicate by the firm itself or its competitors. describe the ease of imitation. Appropriability is a function both of the ease of replication andWhen the tacit component is high, imitation may

well be impossible, absent the hiring away of the efficacy of intellectual property rights as a barrier to imitation. Appropriability is strongkey individuals and the transfers of key organiza-

tion processes. when a technology is both inherently difficult to replicate and the intellectual property systemHowever, another set of barriers impedes imi-

tation of certain capabilities in advanced industrial provides legal barriers to imitation. When it is inherently easy to replicate and intellectualcountries. This is the system of intellectual pro-

perty rights, such as patents, trade secrets, and property protection is either unavailable or inef- fectual, then appropriability is weak. Intermedi-trademarks, and even trade dress.57 Intellectual

property protection is of increasing importance in ate conditions also exist. the United States, as since 1982 the legal system has adopted a more pro-patent posture. Similar trends are evident outside the United States.CONCLUSION Besides the patent system, several other factors cause there to be a difference between replication The four paradigms discussed above are quite

different, though the first two have much incosts and imitation costs. The observability of the technology or the organization is one such common with each other (strategizing) as do the

last two ( economizing). But are these paradigmsimportant factor. Whereas vistas into product technology can be obtained through strategies complementary or competitive? According to

some authors, ‘the resource perspective com-such as reverse engineering, this is not the case for process technology, as a firm need not expose plements the industry analysis framework’ (Amit

and Schoemaker, 1993: 35). While this isits process technology to the outside in order to benefit from it.58 Firms with product technology, undoubtedly true, we think that in several

important respects the perspectives are also com-on the other hand, confront the unfortunate cir- cumstances that they must expose what they have petitive. While this should be recognized, it is

not to suggest that there is only one frameworkgot in order to profit from the technology. Secrets that has value. Indeed, complex problems are likely to benefit from insights obtained from all

57 Trade dress refers to the ‘look and feel’ of a retail establish-of the paradigms we have identified plus more. ment, e.g., the distinctive marketing and presentation style of

The trick is to work out which frameworks areThe Nature Company. 58 An interesting but important exception to this can be foundappropriate for the problem at hand. Slavish in second sourcing. In the microprocessor business, until theadherence to one class to the neglect of all introduction of the 386 chip, Intel and most other merchant

others is likely to generate strategic blindspots.semi producers were encouraged by large customers like IBM to provide second sources, i.e., to license and share theirThe tools themselves then generate strategic proprietary process technology with competitors like AMDvulnerability. We now explore these issues and NEC. The microprocessor developers did so to assure

further. Table 1 summarizes some similaritiescustomers that they had sufficient manufacturing capability to meet demand at all times. and differences.

D y n

a m

ic C

a p

a b

ilitie s

5 2

7

Table 1. Paradigms of strategy: Salient characteristics

Paradigm Intellectual Representative Nature Rationality Fundamental Short-run Role of Focal roots authors of rents assumptions units of capacity for industrial concern

addressing of managers analysis strategic structure strategic reorientation management questions

(1) Attenuating Mason, Porter (1980) Chamberlinean Rational Industries, High Exogenous Structural competitive Bain firms, conditions and forces products competitor

positioning

(2) Strategic Machiavelli, Ghemawat (1986) Chamberlinean Hyper-rational Firms, Often Endogenous Strategic conflict Schelling, Shapiro (1989) products infinite interactions

Cournot, Brandenburger and Nash, Nalebuff (1995) Harsanyi, Shapiro

(3) Resource-based Penrose, Rumelt (1984) Ricardian Rational Resources Low Endogenous Asset perspectives Selznick, Chandler (1966) fungibility

Christensen, Wernerfelt (1984) Andrews Teece (1980, 1982)

(4) Dynamic Schumpeter, Dosi, Teece, and Schumpeterian Rational Processes, Low Endogenous Asset capabilities Nelson, Winter (1989) positions, accumulation, perspective Winter, Prahalad and paths replicability

Teece Hamel (1990) and Hayes and inimitability Wheelwright (1984) Dierickx and Cool (1989) Porter (1990)

528 D. J. Teece, G. Pisano and A. Shuen

in special circumstances, too much ‘strategizing’Efficiency vs. market power can lead firms to underinvest in core competences and neglect dynamic capabilities, and thus harmThe competitive forces and strategic conflict

approaches generally see profits as stemming long-term competitiveness. from strategizing—that is, from limitations on competition which firms achieve through raising

Normative implications rivals’ costs and exclusionary behavior (Teece, 1984). The competitive forces approach in The field of strategic management is avowedly

normative. It seeks to guide those aspects ofparticular leads one to see concentrated industries as being attractive—market positions can be general management that have material effects on

the survival and success of the business enter-shielded behind entry barriers, and rivals costs can be raised. It also suggests that the sources prise. Unless these various approaches differ in

terms of the framework and heuristics they offerof competitive advantage lie at the level of the industry, or possibly groups within an industry. management, then the discourse we have gone

through is of limited immediate value. In thisIn text book presentations, there is almost no attention at all devoted to discovering, creating, paper, we have already alluded to the fact that

the capabilities approach tends to steer managersand commercializing new sources of value. The dynamic capabilities and resources toward creating distinctive and difficult-to-imitate

advantages and avoiding games with customersapproaches clearly have a different orientation. They see competitive advantage stemming from and competitors. We now survey possible differ-

ences, recognizing that the paradigms are still inhigh-performance routines operating ‘inside the firm,’ shaped by processes and positions. Path their infancy and cannot confidently support

strong normative conclusions.dependencies (including increasing returns) and technological opportunities mark the road ahead. Because of imperfect factor markets, or more

Unit of analysis and analytic focus precisely the nontradability of ‘soft’ assets like values, culture, and organizational experience, Because in the capabilities and the resources

framework business opportunities flow from adistinctive competences and capabilities generally cannot be acquired; they must be built. This firm’s unique processes, strategy analysis must

be situational.61 This is also true with the strategicsometimes takes years —possibly decades. In some cases, as when the competence is protected conflict approach. There is no algorithm for

creating wealth for the entire industry. Prescrip-by patents, replication by a competitor is ineffec- tual as a means to access the technology. The tions they apply to industries or groups of firms

at best suggest overall direction, and may indicatecapabilities approach accordingly sees definite limits on strategic options, at least in the short errors to be avoided. In contrast, the competitive

forces approach is not particularly firm specific;run. Competitive success occurs in part because of policies pursued and experience and efficiency it is industry and group specific. obtained in earlier periods.

Competitive success can undoubtedly flow Strategic change

from both strategizing and economizing,59 but along with Williamson (1991) we believe that The competitive forces and the strategic conflict

approach, since they pay little attention to skills,‘economizing is more fundamental than strategiz- ing . . . . or put differently, that economy is the know-how, and path dependency, tend to see best strategy.’60 Indeed, we suggest that, except

Williamson’s as it embraces more than efficient contract59 Phillips (1971 ) and Demsetz ( 1974) also made the case that market concentration resulted from the competitive success design and the minimization of transactions costs. We also

address production and organizational economies, and theof more efficient firms, and not from entry barriers and restrictive practices. distinctive ways that things are accomplished inside the busi-

ness enterprise.60 We concur with Williamson that economizing and strategiz- ing are not mutually exclusive. Strategic ploys can be used61 On this point, the strategic conflict and the resources and

capabilities are congruent. However, the aspects of ‘situation’to disguise inefficiencies and to promote economizing out- comes, as with pricing with reference to learning curve costs. that matter are dramatically different, as described earlier in

this paper.Our view of economizing is perhaps more expansive than

Dynamic Capabilities 529

strategic choice occurring with relative facility. cal subfield in its industry, following a techno- logical discontinuity. Additionally, the interactionThe capabiliies approach sees value augmenting

strategic change as being difficult and costly. between specialized assets such as firm-specific capabilities and rivalry had the greatest influenceMoreover, it can generally only occur incremen-

tally. Capabilities cannot easily be bought; they on entry timing. must be built. From the capabilities perspective, strategy involves choosing among and committing

Diversification to long-term paths or trajectories of competence development. Related diversification—that is, diversification that

builds upon or extends existing capabilities—isIn this regard, we speculate that the dominance of competitive forces and the strategic conflict about the only form of diversification that a

resources/capabilities framework is likely to viewapproaches in the United States may have some- thing to do with observed differences in strategic as meritorious ( Rumelt, 1974; Teece, 1980, 1982;

Teece et al., 1994). Such diversification will beapproaches adopted by some U.S. and some for- eign firms. Hayes (1985) has noted that American justifiable when the firms’ traditional markets

decline.62 The strategic conflict approach is likelycompanies tend to favor ‘strategic leaps’ while, in contrast, Japanese and German companies tend to be a little more permissive; acquisitions that

raise rivals’ costs or enable firms to effectuateto favor incremental, but rapid, improvements. exclusive arrangements are likely to be seen as efficacious in certain circumstances.

Entry strategies

Here the resources and the capabilities approaches Focus and specialization

suggest that entry decisions must be made with reference to the competences and capabilities Focus needs to be defined in terms of distinctive

competences or capability, not products. Productswhich new entrants have, relative to the compe- tition. Whereas the other approaches tell you little are the manifestation of competences, as com-

petences can be molded into a variety of products.about where to look to find likely entrants, the capabilities approach identifies likely entrants. Product market specialization and decentalization

configured around product markets may causeRelatedly, whereas the entry deterrence approach suggests an unconstrained search for new business firms to neglect the development of core com-

petences and dynamic capabilities, to the extentopportunities, the capabilities approach suggests that such opportunities lie close in to one’s exist- to which competences require accessing assets

across divisions.ing business. As Richard Rumelt has explained it in conversation, ‘the capabilities approach sug- The capabilities approach places emphasis on

the internal processes that a firm utilizes, as wellgests that if a firm looks inside itself, and at its market environment, sooner or later it will find as how they are deployed and how they will

evolve. The approach has the benefit of indicatinga business opportunity.’ that competitive advantage is not just a function of how one plays the game; it is also a function

Entry timing of the ‘assets’ one has to play with, and how these assets can be deployed and redeployed inWhereas the strategic conflict approach tells little

abut where to look to find likely entrants, the a changing market. resources and the capabilities approach identifies likely entrants and their timing of entry. Brittain and Freeman (1980) using population ecology methodologies argued that an organization is

62 Cantwell shows that the technological competence of firmsquick to expand when there is a significant over- persists over time, gradually evolving through firm-specific

lap between its core capabilities and those neededlearning. He shows that technological diversification has been greater for chemicals and pharmaceuticals than for electricalto survive in a new market. Recent research and electronic-related fields., and he offers as an explanation( Mitchell, 1989) showed that the more industry- the greater straight-ahead opportunities in electrical and elec-

specialized assets or capabilities a firm possesses,tronic fields than in chemicals and pharmaceuticals. See Cantwell (1993 ).the more likely it is to enter an emerging techni-

530 D. J. Teece, G. Pisano and A. Shuen

Alchian, A. A. and H. Demsetz (1972 ). ‘Production,Future directions information costs, and economic organization’, American Economic Review, 62, pp. 777– 795.

We have merely sketched an outline for aAmit, R. and P. Schoemaker (1993 ). ‘Strategic assets dynamic capabilities approach. Further theoretical and organizational rent’, Strategic Management

Journal 14(1), pp. 33–46.work is needed to tighten the framework, and Andrews, K. (1987). The Concept of Corporateempirical research is critical to helping us under-

Strategy(3rd ed. ). Dow Jones-Irwin, Homewood, IL.stand how firms get to be good, how they some- Aoki, M. ( 1990). ‘The participatory generation of infor-

times stay that way, why and how they improve, mation rents and the theory of the firm’. In M. and why they sometimes decline.63 Researchers Aoki, B. Gustafsson and O. E. Williamson (eds.),

The Firm as a Nexus of Treaties. Sage, London,in the field of strategy need to join forces with pp. 26–52.researchers in the fields of innovation, manufac-

Argyres, N. (1995). ‘Technology strategy, governanceturing, and organizational behavior and business structure and interdivisional coordination’,Journal

history if they are to unlock the riddles that lie of Economic Behavior and Organization, 28, behind corporate as well as national competitive pp. 337–358.

Argyris, C. and D. Schon (1978 ).Organizationaladvantage. There could hardly be a more Learning. Addison-Wesley, Reading, MA.ambitious research agenda in the social sciences

Arrow, K. (1969). ‘The organization of economictoday. activity: Issues pertinent to the choice of market vs. nonmarket allocation’. InThe Analysis and Evalu- ation of Public Expenditures: The PPB System, 1. U.S. Joint Economic Committee, 91st Session. U.S.ACKNOWLEDGEMENTS Government Printing Office, Washington, DC, pp. 59–73.Research for this paper was aided by support

Arrow, K. (1996 ) ‘Technical information and industrial from the Alfred P. Sloan Foundation through the structure’, Industrial and Corporate Change, 5(2 ), Consortium on Competitiveness and Cooperation pp. 645–652.

Arthur, W. B. (1983 ). ‘Competing technologies andat the University of California, Berkeley. The lock-in by historical events: The dynamics of allo-authors are grateful for helpful comments from cation under increasing returns’, working paper WP-two anonymous referees, as well as from Raffi 83-90, International Institute for Applied Systems

Amit, Jay Barney, Joseph Bower, Henry Ches- Analysis, Laxenburg, Austria. brough, Giovanni Dosi, Sumantra Goshal, PankajBain, J. S. (1959).Industrial Organization. Wiley,

New York.Ghemawat, Connie Helfat, Rebecca Henderson, Baldwin, C. and K. Clark (1991 ). ‘Capabilities andDan Levinthal, Richard Nelson, Margie Peteraf,

capital investment: New perspectives on capitalRichard Rosenbloom, Richard Rumelt, Carl budgeting’, Harvard Business School working paper

Shapiro, Oliver Williamson, and Sidney Winter. #92–004. Useful feedback was obtained from workshops atBarney, J. B. (1986). ‘Strategic factor markets: Expec-

tations, luck, and business strategy’,Managementthe Haas School of Business, the Wharton School, Science32(10), pp. 1231– 1241.the Kellogg School (Northwestern), the Harvard

Barney, J. B., J.-C. Spender and T. Reve (1994 ).Business School, and the International Institute Crafoord Lectures, Vol. 6. Chartwell-Bratt, Bromley,

of Applied Systems Analysis (IIASA) in Vienna, U.K. and Lund University Press, Lund, Sweden. the London School of Economics, and the LondonBaumol, W., J. Panzar and R. Willig (1982).Contest-

able Markets and the Theory of Industry Structure.Business School. Harcourt Brace Jovanovich, New York.

Bercovitz, J. E. L., J. M. de Figueiredo and D. J. Teece (1996). ‘Firm capabilities and managerial decision-REFERENCES making: A theory of innovation biases’. In R. Garud, P. Nayyar and Z. Shapira (eds),Innovation: Over-Abernathy, W. J. and K. Clark (1985 ). ‘Innovation: sights and Foresights. Cambridge University Press,Mapping the winds of creative destruction’, Cambridge, U.K. pp. 233–259.Research Policy, 14, pp. 3 –22.

Brandenburger, A. M. and B. J. Nalebuff (1996).Co- opetition. Doubleday, New York.

63 For a gallant start, see Miyazaki (1995 ) and McGrathBrandenburger, A. M. and B. J. Nalebuff ( 1995). ‘The et al. ( 1996). Chandler’s (1990 ) work on scale and scope, right game: Use game theory to shape strategy’, summarized in Teece (1993 ), provides some historical support Harvard Business Review, 73(4), pp. 57–71. for the capabilities approach. Other relevant studies can beBrittain, J. and J. Freeman (1980). ‘Organizational found in a special issue ofIndustrial and Corporate Change

proliferation and density-dependent selection’. In J.3(3), 1994, that was devoted to dynamic capabilities. R. Kimberly and R. Miles (eds.),The Organizational

Dynamic Capabilities 531

Life Cycle. Jossey-Bass, San Francisco, CA, Ghemawat, P. ( 1986). ‘Sustainable advantage’,Harvard Business Review, 64(5 ), pp. 53–58.pp. 291– 338.

Camp, R. (1989).Benchmarking: The Search for Indus-Ghemawat, P. ( 1991).Commitment: The Dynamics of Strategy. Free Press, New York.try Best practices that Lead to Superior Perform-

ance. Quality Press, Milwaukee, WI. Gilbert, R. J. and D. M. G. Newberry (1982 ). ‘Preemp- tive patenting and the persistence of monopoly’,Cantwell, J. (1993 ). ‘Corporate technological speciali-

zation in international industries’. In M. Casson American Economic Review, 72, pp. 514– 526. Gittell, J. H. (1995). ‘Cross functional coordination,and J. Creedy (eds.),Industrial Concentration and

Economic Inequality. Edward Elgar, Aldershot, control and human resource systems: Evidence from the airline industry’, unpublished Ph.D. thesis, Mas-pp. 216– 232.

Chandler, A.D., Jr. (1966).Strategy and Structure. sachusetts Institute of Technology. Hansen, G. S. and B. Wernerfelt ( 1989). ‘DeterminantsDoubleday, Anchor Books Edition, New York.

Chandler, A. D., Jr. ( 1990).Scale and Scope: The of firm performance: The relative importance of economic and organizational factors’,Strategic Man-Dynamics of Industrial Competition. Harvard Uni-

versity Press, Cambridge, MA. agement Journal, 10(5 ), pp. 399–411. Hartley, R. F. ( 1989). Marketing Mistakes. Wiley,Clark, K. and T. Fujimoto (1991 ).Product Develop-

ment Performance: Strategy, Organization and Man- New York. Hayes , R. ( 1985 ). ‘ Strategic planning: Forward inagement in the World Auto Industries. Harvard

Buiness School Press, Cambridge, MA. reverse ’,Harvard Business Review, 63( 6 ) , pp. 111–119.Coase, R. ( 1937). ‘The nature of the firm’,Economica,

4, pp. 386–405. Hayes, R. and K. Clark (1985). ‘Exploring the sources of productivity differences at the factory level’. InCoase, R. ( 1988). ‘Lecture on the Nature of the Firm,

III’, Journal of Law, Economics and Organization, K. Clark, R. H. Hayes and C. Lorenz (eds.),The Uneasy Alliance: Managing the Productivity–4, pp. 33– 47.

Cool, K. and D. Schendel (1988). ‘Performance differ- Technology Dilemma. Harvard Business School Press, Boston, MA, pp. 151– 188.ences among strategic group members’,Strategic

Management Journal, 9(3), pp. 207– 223. Hayes, R. and S. Wheelwright ( 1984).Restoring our Competitive Edge: Competing Through Manufactur-de Figueiredo, J. M. and D. J. Teece ( 1996). ‘Mitigat-

ing procurement hazards in the context of inno- ing. Wiley, New York. Hayes, R., S. Wheelwright and K. Clark (1988 ).vation’, Industrial and Corporate Change, 5(2),

pp. 537–559. Dynamic Manufacturing: Creating the Learning Organization. Free Press, New York.Demsetz, H. (1974 ). ‘Two systems of belief about

monopoly’. In H. Goldschmid, M. Mann and J. F. Henderson, R. M. ( 1994). ‘The evolution of integrative capability: Innovation in cardiovascular drug dis-Weston (eds.),Industrial Concentration: The New

Learning. Little, Brown, Boston, MA, pp. 161– 184. covery’,Industrial and Corporate Change, 3(3), pp. 607–630.Dierickx, I. and K. Cool (1989). ‘Asset stock accumu-

lation and sustainability of competitive advantage’, Henderson, R. M. and K. B. Clark (1990). ‘Architec- tural innovation: The reconfiguration of existingManagement Science, 35(12), pp. 1504– 1511.

Dixit, A. (1980). ‘The role of investment in entry product technologies and the failure of established firms’, Administrative Science Quarterly, 35,deterrence’,Economic Journal, 90, pp. 95–106.

Dosi, G., D. J. Teece and S. Winter (1989 ). ‘Toward a pp. 9– 30. Henderson, R. M. and I. Cockburn (1994). ‘Measuringtheory of corporate coherence: Preliminary remarks’,

unpublished paper, Center for Research in Manage- competence? Exploring firm effects in pharmaceut- ical research,Strategic Management Journal, Sum-ment, University of California at Berkeley.

Doz, Y. and A. Shuen (1990 ). ‘From intent to outcome: mer Special Issue,15, pp. 63–84. Iansiti, M. and K. B. Clark (1994). ‘Integration andA process framework for partnerships’, INSEAD

working paper. dynamic capability: Evidence from product develop- ment in automobiles and mainframe computers’,Fama, E. F. (1980). ‘Agency problems and the theory

of the firm’, Journal of Political Economy, 88, Industrial and Corporate Change, 3(3 ), pp. 557– 605.pp. 288– 307.

Freeman, J. and W. Boeker (1984). ‘The ecological Itami, H. and T. W. Roehl (1987 ).Mobilizing Invisible Assets. Harvard University Press, Cambridge, MA.analysis of business strategy’. In G. Carroll and D.

Vogel (eds.), Strategy and Organization. Pitman, Jacobsen, R. (1988). ‘The persistence of abnormal returns’, Strategic Management Journal, 9(5 ),Boston, MA, pp. 64–77.

Fujimoto, T. (1994 ). ‘Reinterpreting the resource-capa- pp. 415–430. Katz, M. and C. Shapiro ( 1985). ‘Network externalities,bility view of the firm: A case of the development-

production systems of the Japanese automakers’, competition and compatibility’,American Economic Review, 75, pp. 424– 440.draft working paper, Faculty of Economics, Univer-

sity of Tokyo. Kreps, D. M. and R. Wilson (1982a). ‘Sequential equi- libria’, Econometrica, 50, pp. 863–894.Garvin, D. (1988 ). Managing Quality. Free Press,

New York. Kreps, D. M. and R. Wilson (1982b). ‘Reputation and imperfect information’,Journal of Economic Theory,Garvin, D. (1994). ‘The processes of organization and

management’, Harvard Business School working 27, pp. 253–279. Langlois, R. ( 1994 ). ‘Cognition and capabilities:paper #94–084.

532 D. J. Teece, G. Pisano and A. Shuen

Opportunities seized and missed in the history of Mitchell, W. (1989 ). ‘Whether and when? Probability and timing of incumbents’ entry into emergingthe computer industry’, working paper, University

of Connecticut. Presented at the conference on industrial subfields’,Administrative Science Quar- terly, 34, pp. 208–230.Technological Oversights and Foresights, Stern

School of Business, New York University, 11– 12 Miyazaki, K. ( 1995).Building Competences in the Firm: Lessons from Japanese and European Optoe-March 1994.

Learned, E., C. Christensen, K. Andrews and W. Guth lectronics. St. Martins Press, New York. Mody, A. (1993). ‘Learning through alliances’,Journal(1969 ). Business Policy: Text and Cases. Irwin,

Homewood, IL. of Economic Behavior and Organization, 20(2), pp. 151–170.Leonard-Barton, D. (1992 ). ‘Core capabilities and core

rigidities: A paradox in managing new product Nelson, R. R. (1991). ‘Why do firms differ, and how does it matter?’ Strategic Management Journal,development’,Strategic Management Journal, Sum-

mer Special Issue,13, pp. 111–125. Winter Special Issue,12, pp. 61–74. Nelson, R. R. (1994). ‘The co-evolution of technology,Leonard-Barton, D. (1995).Wellsprings of Knowledge.

Harvard Business School Press, Boston, MA. industrial structure, and supporting institutions’, Industrial and Corporate Change, 3(1), pp. 47–63.Lev, B. and T. Sougiannis (1992 ). ‘The capitalization,

amortization and value-relevance of R&D’, unpub- Nelson, R. (1996). ‘The evolution of competitive or comparative advantage: A preliminary report on alished manuscript, University of California, Berke-

ley, and University of Illinois, Urbana–Champaign. study’, WP-96-21, International Institute for Applied Systems Analysis, Laxemberg, Austria.Levinthal, D. and J. March (1981 ). ‘A model of adap-

tive organizational search’,Journal of Economic Nelson, R. and S. Winter ( 1982).An Evolutionary Theory of Economic change. Harvard UniversityBehavior and Organization, 2, pp. 307–333.

Levinthal, D. A. and J. G. March (1993). ‘The myopia Press, Cambridge, MA. Penrose, E. (1959 ).The Theory of the Growth of theof learning’, Strategic Management Journal, Winter

Special Issue,14, pp. 95–112. Firm. Basil Blackwell, London. Phillips, A. C. (1971 ).Technology and Market Struc-Levinthal, D. and J. Myatt (1994). ‘Co-evolution of

capabilities and industry: The evolution of mutual ture. Lexington Books, Toronto. Pisano, G. ( 1994). ‘Knowledge integration and thefund processing’, Strategic Management Journal,

Winter Special Issue,15, pp. 45–62. locus of learning: An empirical analysis of process development’,Strategic Management Journal, Win-Levitt, B. and J. March (1988 ). ‘Organizational learn-

ing’, Annual Review of Sociology, 14, pp. 319–340. ter Special Issue,15, pp. 85–100. Porter, M. E. ( 1980).Competitive Strategy. Free Press,Link, A. N., D. J. Teece and W. F. Finan (October

1996 ). ‘Estimating the benefits from collaboration: New York. Porter, M. E. (1990).The Competitive Advantage ofThe Case of SEMATECH’, Review of Industrial

Organization, 11, pp. 737–751. Nations. Free Press, New York. Prahalad, C. K. and G. Hamel (1990 ). ‘The coreLippman, S. A. and R. P. Rumelt (1992) ‘Demand

uncertainty and investment in industry-specific capi- competence of the corporation’,Harvard Business Review, 68(3), pp. 79–91.tal’, Industrial and Corporate Change, 1(1),

pp. 235– 262. Richardson, G. B. H. (1960, 1990).Information and Investment. Oxford University Press, New York.Mahoney, J. (1995). ‘The management of resources and

the resources of management’,Journal of Business Rosenberg, N. (1982 ).Inside the Black Box: Tech- nology and Economics. Cambridge University Press,Research, 33(2), pp. 91–101.

Mahoney, J. T. and J. R. Pandian (1992). ‘The resource- Cambridge, MA Rumelt, R. P. ( 1974 ). Strategy, Structure, andbased view within the conversation of strategic man-

agement’, Strategic Management Journal, 13(5), Economic Performance. Harvard University Press, Cambridge. MA.pp. 363–380.

Mason, E. (1949 ). ‘The current state of the monopoly Rumelt, R. P. (1984). ‘Towards a strategic theory of the firm’. In R. B. Lamb (ed.),Competitive Strategicproblem in the U.S.’, Harvard Law Review, 62,

pp. 1265 –1285. Management. Prentice-Hall, Englewood Cliffs, NJ, pp. 556–570.McGrath, R. G., M-H. Tsai, S. Venkataraman and

I. C. MacMillan (1996 ). ‘Innovation, competitive Rumelt, R. P. (1991). ‘How much does industry matter?’, Strategic Management Journal, 12(3 ),advantage and rent: A model and test’,Management

Science,42(3), pp. 389–403. pp. 167–185. Rumelt, R. P., D. Schendel and D. Teece (1994 ).Milgrom, P. and J. Roberts ( 1982a). ‘Limit pricing and

entry under incomplete information: An equilibrium Fundamental Issues in Strategy. Harvard Business School Press, Cambridge, MA.analysis’, Econometrica, 50, pp. 443–459.

Milgrom, P. and J. Roberts (1982b). ‘Predation, repu- Schmalensee, R. (1983 ). ‘Advertising and entry deter- rence: An exploratory model’,Journal of Politicaltation and entry deterrence’,Journal of Economic

Theory, 27, pp. 280–312. Economy, 91(4 ), pp. 636–653. Schumpeter, J. A. ( 1934).Theory of Economic Devel-Milgrom, P. and J. Roberts (1990). ‘The economics of

modern manufacturing: Technology, strategy, and opment. Harvard University Press, Cambridge, MA. Schumpeter, J. A. ( 1942).Capitalism, Socialism, andorganization’, American Economic Review, 80(3),

pp. 511– 528. Democracy. Harper, New York.

Dynamic Capabilities 533

Shapiro, C. (1989). ‘The theory of business strategy’, of rapid technological progress’,Journal of Economic Behavior and Organization, 18(1 ), pp. 1– 25.RAND Journal of Economics, 20(1), pp. 125–137.

Shuen, A. ( 1994). ‘Technology sourcing and learning Teece, D. J. (1993 ). ‘The dynamics of industrial capi- talism: Perspectives on Alfred Chandler’sScale andstrategies in the semiconductor industry’, unpub-

lished Ph.D. dissertation, University of California, Scope (1990)’, Journal of Economic Literature, 31(1), pp. 199– 225.Berkeley.

Sutton, J. ( 1992). ‘Implementing game theoretical mod- Teece, D. J. (1996) ‘Firm organization, industrial struc- ture, and technological innovation’,Journal of Econ-els in industrial economies’, In A. Del Monte (ed.),

Recent Developments in the Theory of Industrial omic Behavior and Organization, 31, pp. 193–224. Teece, D. J. and G. Pisano (1994). ‘The dynamicOrganization. University of Michigan Press, Ann

Arbor, MI, pp. 19– 33. capabilities of firms: An introduction’,Industrial and Corporate Change, 3(3), pp. 537– 556.Szulanski, G. (1995 ). ‘Unpacking stickiness: An empiri-

cal investigation of the barriers to transfer best Teece, D. J., R. Rumelt, G. Dosi and S. Winter (1994 ). ‘Understanding corporate coherence: Theory and evi-practice inside the firm’,Academy of Management

Journal, Best Papers Proceedings, pp. 437–441. dence’,Journal of Economic Behavior and Organi- zation, 23, pp. 1– 30.Teece, D. J. ( 1976).The Multinational Corporation

and the Resource Cost of International TechnologyTushman, M. L., W. H. Newman and E. Romanelli (1986). ‘Convergence and upheaval: Managing theTransfer. Ballinger, Cambridge, MA.

Teece, D. J. (1980). ‘Economics of scope and the unsteady pace of organizational evolution’,Califor- nia Management Review, 29(1 ), pp. 29–44.scope of the enterprise’,Journal of Economic

Behavior and Organization, 1, pp. 223–247. Wernerfelt, B. (1984 ). ‘A resource-based view of the firm’, Strategic Management Journal, 5(2 ),Teece, D. J. (1981). ‘The market for know-how and

the efficient international transfer of technology’, pp. 171–180. Wernerfelt, B. and C. Montgomery (1988). ‘Tobin’s QAnnals of the Academy of Political and Social

Science, 458, pp. 81–96. and the importance of focus in firm performance’, American Economic Review, 78(1 ), pp. 246–250.Teece, D. J. ( 1982). ‘Towards an economic theory of

the multiproduct firm’, Journal of Economic Williamson, O. E. (1975 ).Markets and Hierarchies. Free Press, New York.Behavior and Organization, 3, pp. 39–63.

Teece, D. J. (1984). ‘Economic analysis and strategic Williamson, O. E. ( 1985).The Economic Institutions of Capitalism. Free Press, New York.management’, California Management Review,

26(3 ), pp. 87– 110. Williamson, O. E. (1991 ). ‘Strategizing, economizing, and economic organization’,Strategic ManagementTeece, D. J. (1986a). ‘Transactions cost economics and

the multinational enterprise’,Journal of Economic Journal, Winter Special Issue,12, pp. 75–94. Williamson, O. E. (1996)The Mechanisms of Govern-Behavior and Organization, 7, pp. 21–45.

Teece, D. J. (1986b). ‘Profiting from technological ance. Oxford University Press, New York. Womack, J., D. Jones and D. Roos (1991).Theinnovation’, Research Policy, 15(6), pp. 285– 305.

Teece, D. J. 1988. ‘Technological change and the nature Machine that Changed the World. Harper-Perennial, New York.of the firm’. In G. Dosi, C. Freeman, R. Nelson,

G. Silverberg and L. Soete (eds.),Technical Change Zander, U. and B. Kogut (1995). ‘Knowledge and the speed of the transfer and imitation of organizationaland Economic Theory. Pinter Publishers, New York,

pp. 256– 281. capabilities: An empirical test’,Organization Science, 6(1), pp. 76–92.Teece, D. J. ( 1992). ‘Competition, cooperation, and

innovation: Organizational arrangements for regimes