1
Not Your Stepping Stone: Collaboration and the Dynamics of Industry Evolution in Biotechnology
Kenneth Koput Walter W. Powell
University of Arizona Stanford University
Rough Draft for
Organization Science Winter Conference
February, 2000
1
INTRODUCTION
Contemporary research on organizations and industrial performance is replete with reports of a marked upsurge in various forms of interorganizational collaboration, including research consortia, joint ventures, strategic alliances, extensive subcontracting and/or outsourcing of key functions. Astute observers of these developments, such as Richard Rosenbloom and William Spencer (1996), suggest that industrial competition today resembles less a horse race and more a rugby match in which players frequently change uniforms.
A recent National Research Council analysis of trends in industrial research and development (R&D) reports that the innovation process has undergone a significant transformation over the past decade (Merrill and Cooper, 1999). This transformation appears to be both “substantial” in magnitude and consequential to economic performance, both positively and negatively (p. 104). The four components of this reorienting of R&D are: 1.) A shift in the industries and sectors that dominate R&D towards new emerging technologies and nonmanufacturing industries; 2.) A change in the time horizons of R&D, with industry focusing more on shorter-term development and relying more on universities for basic research; 3.) A change in the organizational structure of R&D, with greater decentralization of research activities and increased reliance on both outsourcing and collaboration among firms, universities, and government laboratories; and 4.) Changes in the location of R&D, with successful research increasingly dependent on geographic proximity to clusters of related organizations.
A companion National Research Council survey of eleven industries, purposefully diverse in character and technology but all resurgent in the 1990s, notes that common to each industry is: 1.) Increased reliance on such external sources of R&D as universities, consortia, and government laboratories; 2.) Greater collaboration with domestic and foreign competitors, as well as customers, in the development of new products and processes (Mowery, 1999: 7). National Science Foundation data show a marked increase in the number of international alliances between U.S. and Western European countries between 1980 and 1994; but by the mid-1990’s, the formation rates for intranational alliances linking U.S. firms with their domestic competitors outpace international linkages (National Science Foundation, 1998). Similarly, there is now ample research showing the growing links between U.S. firms and universities (Cohen et al 1994), a greater involvement by firms and government labs in research joint ventures (Link, 1996; 1999), and a much greater foreign presence in U.S. R&D through collaboration with U.S. universities (NSF, 1998). In the realm of science, Hicks and Katz (1996) find that research papers are more and more likely to be co-authored and involve authors with multiple institutional affiliations spanning the public and private sectors. In short, as Mowery (1999: 9) observes, “the diversity of institutional actors and relationships in the industrial innovation process has increased considerably, even as the investments by U.S. firms now appear to focus on shorter time horizons.” Complex networks of firms, universities, and government labs now play a critical role in many industries, and especially so in a number of newer industries such as computers, semiconductors, pharmaceuticals and biotechnology (Powell and Owen-Smith, 1998; Mowery and Nelson, 1999).
Divergent Accounts
Taken together, these developments in industry and science suggest a significant restructuring of organizations, work arrangements and the organization of innovation.
There is, however, considerable debate about both the causes and consequences of this restructuring, and much scholarly and popular discussion about the purported new economy. We cannot resolve these large issues here, rather we focus on one aspect: the role of the small science-based startup firm. The startup firm, whether as a spinoff out of existing companies or universities or as a stand-alone entity, has played a greater role in the U.S. economy over the past two decades. Moreover, these smaller units appear to operate according to a different organizing logic, with extensive linkages to other organizations and partnerships with outside parties for key business functions. But beyond these observations, there is disagreement about both the contributions and the developmental path of smaller firms.
While small firms may be bountiful, some dismiss them as trivial and controlled by larger firms. Harrison (1994) argues that we are witnessing the growth of decentralized power, where there is growing concentration of corporate power but without centralization. He dubs this phenomena the “lean and mean” strategy, and provides anecdotal data drawn selectively from a few firms such as Nike to portray an intensive spider’s-web world of outsourcing, contract labor, demanding work practices, but growing control at the center where corporate power is lodged. In his view, small firms are largely dependent on corporate giants. Another perspective on recent developments is offered by ecologist Glenn Carroll (1985; 1994; Carroll and Hannan, 1995: 215-21), who suggests a resource-partitioning process that involves the simultaneous expansion in the number of small firms and a contraction in number but a concentration in size of large firms. Carroll (1994) notes that there has been a proliferation of smaller organizational units, with the average size of a firm declining by roughly 30 – 40% over the period 1960-1989. Some of this transformation is due to compositional shifts in the U.S. economy, reflected in service sector growth and manufacturing decline. But Carroll and his colleagues offer a resource-based argument that applies especially to fields in which production or marketing activities exhibit economies of scale and price competition. As competition for scale economies increases, only a few generalist organizations survive, and they do so by offering fairly homogeneous products or services to a mass market audience. But the interesting twist is that the concentration of generalists on the mass market opens small pockets of resources on the periphery of the market, where smaller specialist firms emerge and thrive. Thus, in the words of Carroll and Hannan (1995: 217) “increasing market concentration enhances the life chances of specialist organizations.” Such a process has been observed in the newspaper, beer, wine, microprocessor, music, and book industries. In this account, small firms proliferate by catering to specialized tastes
An alternative view is found in recent studies of the innovation process, cited above. Rather than dependency or specialism, these analysts suggest a refashioning of the division of labor in which smaller firms, and the linkages between them and centers of innovative activity play a much more prominent role, especially in research-intensive fields. A somewhat discordant chorus of voices are found in this camp, with some suggesting that networks represent an alternative means for governing economic exchange (Powell, 1990), while others point to a profound blurring of organizational boundaries and a remaking of the production process (Sobel, 1991), and still others contending that new information technologies allow more disaggregated, and flexible means for organizing production and delivering services (Morton, 1991; Brynjolfsson et al, 1994; Shapiro and Varian, 1999).
Older, Bigger, Wiser?
We want to add some empirical flesh to these discussions of the structure of the firm, and to debates over the relevant role of small firms and the growing salience of alliances and various types of collaborations. We have the advantage of having good longitudinal data on firm arrangements over the period 1988-1999. We have the disadvantage that the data are based on only one industry --biotechnology. Nevertheless, the commercial field of the life sciences is purportedly one of the key components of the new economy, so it is a strategic site for analysis. Moreover, the data we describe below are based on formal contractual agreements and not informal ties, handshake deals, or social embedding, hence they afford a strict test of whether horizontally based external relationships are a critical feature of the contemporary science-based firm.[1]
One way to sift through the divergent arguments concerning the structure of contemporary organizations is to examine a sample of firms in the same field over a sufficient time period that changes at the firm-level and field-level, as well transformations in the larger economy, can be observed. Viewed through the lens of more than a decade’s changes, one can examine patterns of growth for individual firms, changes in the nature and number of interorganizational relationships, and industry evolution. To account for changes in the repertoire of organizational practices, we analyze whether there is persistence, decline, or expansion in the use of external collaborations. The crux of the argument for changes in strategy and structure is that interorganizational networks are an increasingly fundamental cornerstone that enables firms to both gain and hold competitive advantage, rather than a transitional stage. Thus, we attempt to measure how organizational strategy and structure have evolved in one industry.
We assess the consequences of organizational growth, aging, and success for the types of collaborative arrangements that firms in the commercial field of biotechnology employ. Each process -- growth in size, the gaining of experience, and the successful launching of a new product -- presents a challenge to a firm in terms of how it chooses to organize. Growth brings problems of communication and coordination as the number of employees grows. As an organization ages, its stock of knowledge increases and its routines for organizing core tasks become more well developed. With a successful product, an organization faces the challenge of reproducing its initial success but now has new resources to deploy. Viewed differently, growth, aging and success are mileposts for organizations, and we can look at these road markers and see whether the mix of arrangements that were used to reach a particular milepost are continued, disbanded, or diversified as movement toward the next marker occurs.
In our earlier work on biotech firms, which covered a much shorter time period than the analyses presented here, we showed that centrality in the industry network heightened a firm’s reputation and generated access to resources. Firms so positioned attracted new employees, participated in more new ventures, and developed deeper experience at collaborating with other parties. Put colloquially, a firm grew by becoming a player; it did not become a player by growing. Growth and financial success resulted from centrality in industry networks (Powell, Koput, Smith-Doerr, 1996; Powell, Koput, Smith-Doerr, Owen-Smith, 1999).
We begin these analyses by asking how achieving growth influences subsequent behavior. As firms add employees and expand their operations internally, they face a series of choices. Organizations could opt to pursue more activities internally and fewer externally as they increase internal capacity. Or they could use additional staff to assist in expanding the number of outside collaborations. Moreover, growth could result in increased differentiation internally (following Blau’s [1970] classic arguments about growth leading to an expansion of roles and structures), or promote a diversification externally into collaborations for different types of activities. Finally, added internal capacity could lead to more tangential external relations, or could allow for deeper and/or more intensive external linkages. Stated formally, we test to see whether:
As firms grow in size, they deepen their portfolio of collaborations; expand their number of external ties; diversify the types of collaborative activities they engage in; and engage in more complex arrangements. (Hypothesis 1).
The effects of age have been well studied in the organizations literature. Building on Stinchcombe’s (1965) key insights, there has been a great deal of attention paid to the liability of newness (Hannan and Freeman, 1989: 245-70). Younger firms need to build credibility with consumers, suppliers, creditors and the like and overcome initial reservations toward a new entity. Given these obstacles, it is not surprising that startup companies turn to venture capital firms, law firms, and established companies for assistance, as well as rely on universities and government laboratories for key technologies. Moreover, many startup science-based firms are spin-offs from universities and, less frequently, government laboratories and established firms. Consequently, survival often hinges on location in a supportive network in order to overcome initial questions about credibility and reliability. As a small firm ages, it faces a choice. The company can become less reliant on external parties. Or as the firm matures and expands its activities downstream from its initial research focus into product development, production, and sales, it may opt to become involved with more outside partners as it takes on new tasks. More pivotally, firms may engage with a wider span of partners for key activities, selecting and being selected by a heterogeneous group of collaborators for critical functions such as R&D, production, or sales. Put simply, as companies age, they make decisions about both the volume and the span of activities done internally or externally. Hence we test to see whether:
As firms grow older, they become involved in a wider variety of external activities, with a more heterogeneous group of partners for each activity. (Hypothesis 2).
In many emerging industries with a strong science base, companies spend much of their early years burning up investors’ money on costly initial research. External support is the lifeblood during this phase. As the research program develops and moves into application and testing, the prospect of a new product looms larger. As the product cycle unfolds, companies have differential needs, and the nature of the reliance on others shifts depending on the stage of development. For the successful firm, products are eventually released and sales generated. Some firms plow revenues back into R&D, others generate profits to repay investors who have been patient during the development stages. We have argued that companies in knowledge-intensive businesses are involved in learning races, that is, in a cycle of learning in which initial success generates the resources that allow advancement to the next level (Powell et al, 1999). The key question is whether that initial success triggers changes in current organizational arrangements or “restarts” the process again in pursuit of new avenues of research.
Thus the fruits of initial success present a branching point for many young companies. The founders could choose to “take the money and run,” cashing out by selling the company or agreeing to a merger. The founders could take the money and attempt to build a vertically integrated firm, with the full range of organizational functions, and with key tasks performed only internally. The cycles of learning argument that we have posited suggests that organizations respond to the initial success of sales and/or profits by expanding their absorptive capacity (Cohen and Levinthal, 1989; 1990), that is, by enhancing BOTH their ability to generate new products internally as well as their portfolio of external research affiliations. An organization’s absorptive capacity allows it to make sense of news generated elsewhere and to make news on its own (Nelson, 1994). Seen in this light, we assess whether, as firms reach an initial “finish” line, they reinvest in the process of research exploration, choose to pursue research internally, or choose to exit. More formally stated, we test the argument that:
As companies first reach an initial plateau of earnings, they subsequently engage in more external R&D collaborations. (Hypothesis 3).
Researchers from several disciplines have stressed the importance of research prowess as an admission ticket to information networks in science-based fields (Mowery and Rosenberg, 1989; Arora and Gambardella, 1994; Powell et al, 1996). A persistent finding from a diverse set of empirical studies is that R&D intensity and the level of technological sophistication are positively correlated with the number and intensity of strategic alliances (C. Freeman, 1991; Hagedoorn, 1995). More generally, there is widespread agreement that in technologically advanced sectors, the locus of innovation is often found in networks of relationships (Powell et al, 1996) or sectoral innovation systems (Mowery and Nelson, 1999). But we do not have a clear sense of which actors bind these innovation networks together, providing them with coherent agendas.
Indeed, one would expect substantial variation in network structure based on differences in technology, the supportive institutional infrastructure, public policy, and the endowments of large and small participants. We cannot provide a complete picture of this complex process here. Rather we are interested in the role and evolution of the small science-based firm. Consider the following range of possible roles performed by startups. Small firms may provide the ideas that jumpstart the innovation process. Because these firms are both closer to the underlying basic science and have a deeper understanding of the technology, they continue to exert influence as the ideas are translated into commercial products. In this scenario, the small firm is able to use its knowledge to orchestrate production through an innovation network. In contrast, small firms could serve as the sources of new ideas or novel products, but, due to their small scale they are unable to develop, produce, or market the final product. Under this scenario, the small firm would remain a specialist research boutique that generates ideas, while larger and more established firms would pick off the most promising leads and thus reap the lion’s share of the benefits from innovation. Alternatively, an inner circle of key participants may dominate the innovation network, and these centrally positioned actors are able to reap greater rewards from their strategic location in the chain of production. This inner circle might well compromise an elite mix of large and small firms, research institutes, financiers, and providers of other key resources that play a disproportionally large role. Finally, industrial structure might be highly fluid, with different organizations taking on more central roles at different times and in different places, as well as for different stages of the product life cycle. Under this latter scenario, there is neither an inner circle of participants, nor a boutique or orchestrator role for small firms, but a changing division of labor in which industrial leadership varies according to capability and resource availability. We assess the prevalence of these divergent accounts of the divisions of labor by testing the following arguments: