8
The Division of Innovative Labor in Biotechnology

ASHISH ARORA AND ALFONSO GAMBARDELLA

The growing use of general and abstract knowledge, based upon an increasing understanding of principles governing phenomena and the tremendous growth in computational capabilities, has opened up new possibilities for specialization (Arora and Gambardella, 1992). The breadth of downstream applications of abstract principles and the feasibility of representing them in a universal and codified form implies that information can be divided into "pieces" that can be usefully recomposed together to form "larger pieces" of information, provided a general and comprehensive framework for integrating the information exists. With suitable contracts, the individual pieces of information and knowledge can be bought and sold. This lays the basis for economic agents to specialize according to comparative advantage, thereby giving rise to the division of inventive labor.

A number of authors have commented upon the growing importance of networks and strategic alliances in innovation (see, for example, Mowery, 1988). Some have gone so far as to argue that it is now difficult to identify the innovator with a particular organization. Imai (1988), inter alia, has suggested that the locus of innovation is shifting from the individual firm to a network of firms, where the network itself is the innovating institution. Instead, we submit that networks and strategic alliances are best viewed as important special cases of the more general phenomenon of a division of innovative labor.

To highlight the use of general and abstract knowledge in innovation, we focus here upon one sector of the increasingly diverse "biotechnology industry," the human diagnostics and therapeutics sector, often referred to as "biopharmaceuticals" (Burrill, 1989). Biotechnology is an example par excellence of the use



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 188
Sources of Medical Technology: Universities and Industry 8 The Division of Innovative Labor in Biotechnology ASHISH ARORA AND ALFONSO GAMBARDELLA The growing use of general and abstract knowledge, based upon an increasing understanding of principles governing phenomena and the tremendous growth in computational capabilities, has opened up new possibilities for specialization (Arora and Gambardella, 1992). The breadth of downstream applications of abstract principles and the feasibility of representing them in a universal and codified form implies that information can be divided into "pieces" that can be usefully recomposed together to form "larger pieces" of information, provided a general and comprehensive framework for integrating the information exists. With suitable contracts, the individual pieces of information and knowledge can be bought and sold. This lays the basis for economic agents to specialize according to comparative advantage, thereby giving rise to the division of inventive labor. A number of authors have commented upon the growing importance of networks and strategic alliances in innovation (see, for example, Mowery, 1988). Some have gone so far as to argue that it is now difficult to identify the innovator with a particular organization. Imai (1988), inter alia, has suggested that the locus of innovation is shifting from the individual firm to a network of firms, where the network itself is the innovating institution. Instead, we submit that networks and strategic alliances are best viewed as important special cases of the more general phenomenon of a division of innovative labor. To highlight the use of general and abstract knowledge in innovation, we focus here upon one sector of the increasingly diverse "biotechnology industry," the human diagnostics and therapeutics sector, often referred to as "biopharmaceuticals" (Burrill, 1989). Biotechnology is an example par excellence of the use

OCR for page 188
Sources of Medical Technology: Universities and Industry of general and abstract knowledge in innovation. Technological advances based on genetic engineering and molecular biology have deeply affected the pharmaceutical industry. They have given great impetus to the possibility of understanding the "causes" of diseases and the action of drugs. Reliance upon random screening of compounds to find what may work is giving way to more selective and carefully structured experiments, guided by basic theory. As a result, now researchers can often design drugs on computers and "build" them in labs before extensive experimentation on animals and human beings.1 The biotechnology sector also provides a prototypical example of the changing patterns of specialization in inventive activity that we call the division of innovative labor. Historically, large pharmaceutical companies, which integrated activities from research to distribution, have been the primary source of innovation in the industry. The rise of biotechnology, along with a number of related economic forces, has made possible, and to a large extent forced, specialization and cooperation among large pharmaceutical firms and small, research-intensive, biotechnology enterprises. In this network of innovators, universities and research institutions occupy an important place as well. The next section describes more fully the participants in the division of labor in biotechnology. The section following that discussion examines the different strategies of external linkages from the viewpoint of the large corporations and characterizes the relationship between these different strategies. We then analyze the factors that determine the value that a firm can derive from its external linkages, and the light throws upon the nature of division of innovative labor. The penultimate section discusses whether the division of innovative labor is socially desirable and the final section provides our conclusion. PARTICIPANTS IN THE DIVISION OF INNOVATIVE LABOR IN BIOTECHNOLOGY Up to the 1980s, most new drugs stemmed from systematic investments in internal research and development (R&D) by large corporations (Thomas, 1988). Biotechnology has a "science-push" origin. The initial breakthroughs include the discovery of the double-helix structure of DNA in 1953 and the discovery of the new techniques of recombinant DNA (rDNA) and cell fusion (monoclonal antibodies) in the 1970s, all of which resulted from scientific research within the university system. In addition, there have been significant advances in molecular biology and related fields, such as computer imaging of molecules, which have greatly increased the understanding of the links between molecular structure and 1   We use biotechnology as a convenient short hand for the entire gamut of scientific and technological advances, including recombinant DNA techniques, hybridoma and PCR techniques, and the more fundamental advances in molecular biology, molecular genetics, and biochemistry.

OCR for page 188
Sources of Medical Technology: Universities and Industry function (and dysfunction). The increased knowledge about the structure of biological molecules such as enzymes and the isolation and cloning of human genes offer new opportunities to develop drugs that can counteract diseases (Gambardella, 1995; Office of Technology Assessment, 1984; Pisano et al., 1988). For instance, the synthesis of the nucleic acid sequence, originally performed at the City of Hope Medical Center, led to the large scale production of rDNA insulin. This was developed, produced, and commercialized by a joint venture between Genentech and Eli Lilly. Biotechnology, with its strong grounding in basic science, has changed markedly the organizational pattern of the innovation process in this industry. Three types of agents now contribute to the generation, development, and commercialization of new biotechnology products, and particularly of biopharmaceuticals: the universities; small/medium-sized research-intensive firms or the so-called new biotechnology firms (NBFs); and the large "established" chemical and pharmaceutical manufacturers. Many new drugs and therapies now stem from systematic interactions and cooperation between these three types of agents. Universities and NBFs were important sources of new techniques and products at the outset, and remain very important sources of basic and applied scientific knowledge in this field. However, while collaborations with universities have been common in the pharmaceutical industry for many years (see, for instance, Swann, 1988; Thackeray, 1982), and biotechnology has only intensified an already existing pattern of relationships, the rise of the NBFs is an entirely new phenomenon. Most NBFs were founded during the early 1980s. Their major asset consists of knowledge, especially that embodied in their researchers (Kenney, 1986; Office of Technology Assessment, 1984; Pisano et al., 1988). In many instances they were founded to exploit a discovery made by an individual or group, and they were created by professors or groups of academic scientists. Over a thousand NBFs have been founded in the past decade. Even in a "bad" year such as 1988, some 36 new NBFs were formed (Burrill, 1990). As Table 8-1 shows, a majority of NBFs are U.S. based. A number of reasons can be cited for the formation of NBFs, and particularly for the finding that a majority are U.S. based. Clearly, their growth was triggered by the rise of new scientific and technological opportunities and the fact that established corporations were late to enter into biotechnology research (see, inter alia, Pisano et al., 1988). However, economic forces at work in the U.S. system have played an important role as well. The relationship between universities and industry in this country is unusually close; and there is an availability of finance especially suited for high-risk technology businesses (venture capital and the like; see Florida and Kenney, 1990; Office of Technology Assessment, 1984). The biotech industry in the United States has also benefited considerably from the output of federally sponsored research in this field—often a new product or a method that can rapidly translate into new product opportunities. That is, federally supported research has helped produce discoveries that have, in turn, paved the way to new

OCR for page 188
Sources of Medical Technology: Universities and Industry TABLE 8-1 Composition of the Biotechnology Industry Country Small Medium Large Japan 1 4 46 United States 138 84 65 Europe 39 18 33 Other 21 15 11 NOTE: Small = up to 50 employees; medium = 51–299 employees; large = ≥ 300 employees. SOURCE: The table is based upon Chang (1992, p. 15, Table 1). It covers all companies listed in Bioscan (April 1991) that are involved in human diagnostics or human therapeutics and that provide information on employment. business opportunities. These opportunities have been exploited by the institutions performing the research, by NBFs through their academic ties, or by NBFs or other companies taking advantage of the knowledge generated by public research. Finally, the United States has followed an aggressive policy of defining and strengthening intellectual property rights. In this context, the landmark 1980 Supreme Court decision of Diamond v. Chakrabarty, which established the legality of patenting of life forms, is frequently cited as providing the security needed by small, research-intensive biotech firms whose major asset is their research capability. The basic assets of NBFs are skills and know-how relating to applied laboratory research. Their close ties to the academic system, their collegial atmosphere, and the possibility of using incentive-based compensation schemes (such as stock options for researchers) have given them a comparative advantage in research.2 A typical product of an NBF could be a new protein, obtained from genetically engineered organisms, that can be potentially used for diagnostic or therapeutic purposes. The synthesis of a new protein, however, while an important step, does not exhaust the entire innovation cycle. In pharmaceuticals, downstream development of compounds is a long and costly process, in no small measure due to the stringent regulation of clinical tests. Most NBFs lack the organizational and financial resources for undertaking such developments; nor do they have adequate commercialization capabilities (Burrill, 1989; Kenney, 1986). Furthermore, venture capitalists backing NBFs often press for relatively short-term cash 2   Lerner (1991) provides evidence (measured by citations to scientific publications) that the patents of NBFs are more strongly based in science than the patents of a matched sample of large firms active in biotechnology.

OCR for page 188
Sources of Medical Technology: Universities and Industry returns, which are incompatible with the long time cycles of pharmaceutical innovation. As a result, many NBFs end up offering their skills or potential new products to larger firms for research collaborations and joint product developments instead of undertaking development, manufacturing, and commercialization on their own.3 Large established chemical and pharmaceutical companies have the engineering know-how required to scale up from the laboratory bench to large-scale manufacturing, and further to control the industrial-scale processes. More importantly, they have the financial capabilities for conducting long and costly clinical trials. They are familiar with clinical testing procedures and the regulatory process, and they have established commercialization networks.4 A number of other factors have raised the incentives of the large firms to take advantage of the specialized expertise and product opportunities offered by the NBFs (see Grabowski, 1991, and Telling, 1992). U.S. policy on generic drugs in the 1980s (and particularly the Waxman-Hatch Act in 1984) reduced substantially the property rights of the pharmaceutical companies in their main products when patents expire. As a result, all major drug manufacturers had to undertake significant efforts to create a new portfolio of patented products. In addition, the evolution of the market from one where doctors order medicines without much attention to cost toward more conscious, expert buyers has required that products be not only new, but also superior. Large firms have, then, found resorting to NBFs for new product ideas useful for several reasons. First, as suggested earlier, the research of the NBFs is often partly subsidized by investor capital and possibly public money via university linkages. Second, royalty payments to NBFs are, in most cases, contingent on success. Finally, because biotech firms are typically pressed for cash, large companies can negotiate favorable terms. In other words, an alliance with an NBF can be beneficial to a large firm because it does not require a large financial commitment, and because it allows the large firm useful access to knowledge that has in part been supported by public funds and in part by the investing public.5 In sum, because large firms and NBFs control assets that are largely complementary, systematic collaborations between them have arisen. Moreover, as we saw, economic forces have reinforced this trend by pushing both the NBFs to take advantage of the downstream capabilities of large firms, and the large firms to avail themselves of the upstream research skills of the NBFs. 3   The total turnover of many NBFs is, indeed, still composed for the most part of research contracts to larger firms rather than actual product sales. See, for instance, Burrill (1989). 4   They also have the financial and organizational capabilities for conducting research based upon "lumpy" assets (e.g., molecular design based on expensive supercomputers). In this case, large firms can offer complementary expertise in more "upstream" research as well. 5   We would like to thank Gerald Laubach for a very useful and enlightening discussion of these points.

OCR for page 188
Sources of Medical Technology: Universities and Industry Universities also control assets and skills that are to some extent complementary to those of both the NBFs and the large firms (typically, upstream scientific capabilities). As a result, the growth of the industry has hinged upon network-like relationships based upon extensive collaborations and a division of labor between these three types of agents.6 COLLABORATION STRATEGIES OF LARGE FIRMS There are a number of different types of external linkages that large firms have used. Following the literature on this topic, one can identify four main types of linkages that large firms have formed with other agents in biotechnology:7 (1) they enter into research and/or joint development agreements with other firms; (2) they form research agreements with universities; (3) they invest in the capital stock of NBFs (minority participation); (4) they acquire NBFs. To a large extent, each of these four strategies enables the firm to gain access to a particular set of tangible or intangible resources necessary for innovation.8 Most agreements signed by large chemical and pharmaceutical producers with other companies tend to be project specific. These agreements, usually with NBFs, focus on "downstream" activities of the innovation cycle. They are aimed at developing and commercializing a particular discovery of the NBF (e.g., the synthesis of a new enzyme or hormone or growth factor) in the areas of specialty chemicals, agricultural biotechnology and, above all, pharmaceuticals. The agreements with universities tend to focus on more basic research objectives. Large firms finance research activities performed by academic laboratories to acquire, by interacting with university scientists, some familiarity with the basic knowledge in this field. Such agreements, between large firms and universities, are important sources of recruiting qualified scientists and researchers, and also serve as a means by which firms can engage the services of top researchers while these researchers continue to work in environments they find most congenial.9 Other than these general and "intangible" gains, the agreements with universities 6   Clearly, there is some overlap between the skills, competencies, and activities of universities, NBFs, and large firms. Many large corporations perform in-house basic and applied research. Universities often perform a good deal of applied research. Many NBFs have in-house basic research skills, and a few others now possess downstream capabilities. Nonetheless, our schematic distinction has an important element of truth and serves well as a first approximation. We shall return below to a fuller discussion of the nature of these comparative advantages. 7   See, for instance, Daly (1985), Kenney (1986), Office of Technology Assessment (1984), and Pisano et al. (1988). 8   The following discussion on the different forms of external linkages of the large firms in biotechnology draws upon Arora and Gambardella (1990). 9   Kenney (1986) suggests that a "preferential access" to trained manpower is an important reason for university–corporate linkages in biotechnology.

OCR for page 188
Sources of Medical Technology: Universities and Industry also serve a more explicit objective of the large firms. Given the short distance between science and commercialization in biotechnology, an agreement with a university can provide the large firm with the option of licensing any new discovery of that research center (within the scope defined by the agreement). This is an advantage for a firm that can rapidly translate those new discoveries into commercializable products. Minority participation in the capital stock of small biotech start-up firms is a means of monitoring the internal research activities of the NBFs. By acquiring part of the capital stock of an NBF, the large companies may also hope to establish a "preferential" linkage with that company, which may be useful to preempt rivals in the commercialization of relevant discoveries made by the NBF. Moreover, in keeping with the theoretical predictions of Williamson (1985), such investments may be useful in averting problems of moral hazard by serving as tokens of good faith. In 1986, for instance, American Home Products bought 13.5 percent of the shares of California Biotechnology, and in that same year the two companies entered into formal arrangements in the fields of cardiovascular drugs, veterinary therapeutics, and drug delivery systems. Other examples of this sort include the cases of Abbott and Amgen, American Cyanamid and Cytogen, Johnson & Johnson and Cytogen, and SmithKline-Beecham and Amgen. Similarly, British Biotechnology obtained its start-up capital in 1986 from a consortium that included SmithKline Beecham, and later the two entered into a joint venture for a line of anti-arthritic therapies based upon matrix metallo-proteinex inhibitors developed by British Biotechnology. As far as acquisitions are concerned, there seem to exist two different—and somewhat contrary—motives for acquiring a small biotech company. On the one hand, large companies that have substantial in-house capabilities, longer experience in biotechnology, and more active involvement in the field aim at acquiring NBFs specialized in particular areas of biotechnology research. The experience and in-house expertise of the large companies enables them to evaluate more accurately the likely contributions of the set of specialized resources that are being acquired. On the other hand, the direct acquisition of a biotechnology firm may also represent a way of "catching up" for late entrants.10 These remarks suggest that each of the four types of external linkages targets a separate goal of the large firms, and thus they are mutually complementary. 10   Compare the acquisitions in 1986 of Hybritech by Eli Lilly and of Genetics System and Oncogen by Bristol Myers. Eli Lilly is one of the most research-intensive pharmaceutical companies worldwide, and was an early entrant in biotechnology research. It sought to complement its generalized expertise in biotechnology with know-how in monoclonal antibodies, wherein Hybritech had specialized capabilities. Bristol-Myers has strong marketing capabilities, but it is less research-intensive. Through acquisitions, it sought generalized expertise in the new area of biotechnology. See Gambardella (1992 and 1995).

OCR for page 188
Sources of Medical Technology: Universities and Industry Through agreements with other firms (particularly with NBFs), the large firms can develop and commercialize new biotech products after the NBFs have undertaken the initial upstream stages. The agreements with universities provide the large firms with access to basic scientific knowledge. Minority participation enables them to acquire familiarity with the internal research activities of the NBF. Acquisitions add specialized or generalized internal knowledge to the large firms in some areas of the biotech business. In Arora and Gambardella (1990), we presented a formal model of the complementarity hypothesis and also tested for complementarity between the four kinds of external linkages for a sample of 81 large U.S., European, and Japanese chemical and pharmaceutical companies, during 1983–1989. Our empirical results support the idea that the four strategies are mutually complementary. We found that firms that tended to have a large number of one type of linkage also tended to have a greater number of the others. This suggests that the (marginal) ''value" of each of the four strategies is greater the larger the number of the other types of linkages undertaken by the firm.11In other words, firms differ more in the extent to which they seek out external linkages, rather than in the specific types of linkages that they do seek. EVALUATING AND USING TECHNOLOGICAL INFORMATION These findings raise the question: What factors determine the payoff from external linkages? We offer a different perspective from that of traditional analyses, which have been framed largely in terms of transaction-cost economics. The transaction-cost perspective would suggest that firms attempt to internalize the knowledge-based assets required for innovation. Hence, firms with strong in-house research capabilities would be less likely to enter into strategic alliances. Pisano (1990) has provided the strongest case for this perspective in the context of biotechnology. Pisano's results are based upon a sample of some 92 biotech projects, at the preclinical and earlier stage, from 30 large pharmaceutical firms. He found that the smaller the number of NBFs active in a particular therapeutic area, the greater the number of previous projects the large firm had carried out (through in-house efforts alone) in that particular therapeutic area; and, possibly, the greater the fraction of their sales accounted for by pharmaceuticals, the more likely the pharmaceutical firms were to perform the R&D exclusively in house. The interpretation of these results is that where small numbers of NBFs are bargaining and in-house capability is great, internalization of projects is favored so as to economize on transaction costs. But Pisano's empirical evidence is amenable to alternative interpretations. 11   We also found that, for these large firms, size or nationality did not affect their propensity to engage in external linkages and had relatively low explanatory power.

OCR for page 188
Sources of Medical Technology: Universities and Industry The smaller the number of NBFs in a particular area, the less likely it is that the large firm would be able to find a suitable partner with the appropriate research capabilities. Moreover, Pisano assumes that the choice or problem facing the large firm can be adequately represented by assuming that the project is initiated by the large firm, which then decides whether to use external subcontractors or not. Our understanding of the industry is that NBFs initiate the bulk of the new projects in the sector, some of which may be offered to large firms for their participation. Furthermore, firms such as Eli Lilly, Johnson & Johnson, and Monsanto, which began their in-house research efforts early (in some cases even before the start of the 1980s), were precisely the firms that led in making external collaborative alliances. In essence, the analogy with the make-or-buy decision in production can be misleading in the context of innovation. To be sure, transaction costs are important, but technological knowledge is a special type of good. It requires a great deal of specialized knowledge and skill to successfully utilize technological information. In order to "buy" such knowledge, one has to have a great deal of prior knowledge—inexperienced and unskilled buyers are at a severe disadvantage. In other words, it is precisely firms with strong in-house R&D capabilities that would derive the greatest value from external linkages. Cohen and Levinthal (1989) argue that firms invest in R&D for two purposes. On the one hand, they invest in R&D to generate innovations; on the other hand, R&D serves as a device for exploiting external research. Rosenberg (1990) argues that in-house basic research is necessary to monitor the flow of scientific information in the outside world. Mowery (1981) showed that, during 1921–1945, large U.S. firms, when starting new innovation projects, also contracted out part of the research to specialized institutes. These studies emphasize the potential synergies between external and internal knowledge. However, their discussion does not deal with the multidimensionality of knowledge, and therefore does not "unpackage" the source of the synergies. In Arora and Gambardella (1994b), we attempted to distinguish between two types of knowledge-based capabilities: The ability to utilize and the ability to evaluate information. In the present context, therefore, there are two considerations involved in entering into an external alliance. The first is concerned with the skills and competencies that the large firm has in the development and commercialization of the innovation—in other words, how capable the large firm is of utilizing the information that it is "purchasing" from an NBF. The second consideration is related but logically distinct and has to do with the ability of the firm to form judgments about the potential usefulness of the information that it is buying. Consider, therefore, a large corporation at the point of deciding whether or not to enter into a collaborative agreement with an NBF. The latter typically offers an "idea." For instance, it may have synthesized in an E. coli culture a

OCR for page 188
Sources of Medical Technology: Universities and Industry protein that is closely related to some physiological disorder; or it may have discovered a way of inducing a gene to express itself. In some instances, the NBF may offer a useful technique relating, for example, to a way of separating a particular protein from other products produced by a cell culture or a delivery system. The large firm funds any further research that may be needed, and then may "license-in" the idea for further development. One can think of each such link as the purchase of an option on a "project." The initial investment made in a typical external agreement is not very large relative to the total R&D budgets of these corporations. Typical sums involved in these linkages are on the order of (at most) 5–10 percent of the total expenditure involved in introducing a drug. The willingness of a firm to invest in such a linkage will depend first and foremost on whether it has the requisite development competencies in the particular therapeutic area, and on whether it has underutilized downstream capabilities for successful commercialization.12 A slightly different way of putting it is to note that firms' commercialization capabilities are fixed in the short run and are to some extent specific to particular therapeutic areas. The firm would like to be able to utilize its commercialization capabilities to the fullest extent possible. It follows, therefore, that the better (and more extensive) these downstream capabilities of commercialization (or, as we put it, the better the ability to utilize), the greater the willingness of the large firm to invest in an external alliance. However, the willingness to invest also depends upon the firm's ability to evaluate the likely commercial value of the project. Here the relationship is rather subtle. Since the investment is tantamount to the purchase of an option, firms that are better able to forecast the value of the project will be more discerning in their external collaborations. They are able to focus upon the more promising linkages. In Arora and Gambardella (1994b) we formally show that the two kinds of knowledge-based capabilities have rather different empirical predictions for the propensity of the firm to engage in such linkages. The better the ability to utilize, the greater, on average, the number of external linkages; on the other hand, all else held equal, the better the ability to evaluate, the fewer the number of external linkages (but these are of greater expected value). How does one empirically distinguish between the two types of abilities? The measures one chooses have to be rooted in the specificities of the industrial sector in question. In our study, based on a sample of the 26 largest U.S. pharmaceutical companies, we used two variables as measures of the ability to utilize. The first variable was the ratio of R&D expenditures to sales. Since these large corporations, one would expect a research-intensive firm to also have in 12   Gerald Laubach has emphasized to us the importance of replenishing product pipelines. Firms that have strong abilities to utilize and to evaluate, such as Merck, may nonetheless not have many external alliances if their product pipelines are relatively full.

OCR for page 188
Sources of Medical Technology: Universities and Industry place other assets, such as a marketing and distribution network, that are necessary for successful commercialization of research outcomes. In other words, R&D intensity can serve as a proxy for a "package" of research capabilities and complementary downstream assets. The second variable is the stock of biotechnology patents applied for in the United States. This variable is intended to capture the extent to which a firm has invested in biotechnology-related research in the past years, and therefore its ability to utilize external technological information in that area.13 What publicly available measure can one use for the ability to evaluate? It has been argued that science provides information that helps restrict the search for successful innovations at the downstream, applied R&D stages. Superior scientific capabilities enable the firm to reduce the uncertainty about the outcome of an individual project (David et al., 1988; Nelson, 1961). As noted earlier, Rosenberg (1990) has argued that in-house basic research is useful primarily for being "plugged in" to external information flows. Since a great deal of relevant information in biotechnology is science based, an in-house scientific capability is crucial for evaluating and assessing information originating outside of the firm's boundaries.14 As a measure of in-house scientific capabilities, we used the average number of scientific papers (stock) published by the personnel of the firms divided by total sales.15 In our empirical tests, we found that these measures performed well. The measures of the ability to utilize information are positively related to the number of external linkages, and the measure for scientific capability is negatively related to the number of linkages.16 These results also help us clarify an important point. Our discussion might have unintentionally suggested that within the division of innovative labor in the biotech industry, research is performed by NBFs, while large firms provide only downstream capabilities. This is not true. Large firms perform a great deal of upstream research. In fact, our results suggest that 13   The use of patents as a measure of the technological strength of pharmaceutical firms is supported by the results reported in Narin et al. (1987). 14   It should be noted that by "science" we mean abstract knowledge and representations in terms of general and universal categories. The use of such abstraction allows the researcher to delineate and characterize more carefully the set of possible outcomes, eliminate a number of other possibilities, and hence be able to focus upon a smaller set of more carefully designed experiments. Furthermore, information from diverse sources can be better integrated to throw light upon the problem at hand and thereby allow a more informed and accurate judgment about the likelihood of the success of the project. See Arora and Gambardella (1994a and 1992) for further discussion and references. 15   Halperin and Chakrabarti (1987) found that company scientific publications are highly correlated with the number of elite researchers employed by the firm (more so than patenting by the firms). 16   In Arora and Gambardella (1994b) we focused only on the agreements of large firms with NBFs, and neglected minority participation and acquisitions. We also examined the agreements with universities and other research institutions. However, alliances with universities did not appear to be related in the same way to the measures, suggesting that alliances with universities are more properly thought of as providing access to knowledge and as a means of building up in-house competencies.

OCR for page 188
Sources of Medical Technology: Universities and Industry internal research provides the necessary knowledge to extract greater rents from alliances. In-house capabilities of a more fundamental nature help in directing external investments toward those alliances with greater potential payoff.17 THE DIVISION OF INNOVATIVE LABOR: TRANSIENT AND UNDESIRABLE? Our results bear upon another important question regarding the biotechnology sector as well. As widely discussed among economists, managers, and analysts, an important question about the future of the biotech industry (and more generally of the pharmaceutical industry) is whether the new division of innovative labor is merely a transient phenomenon. We saw that transaction-cost perspectives lead to the view that once large firms acquire sufficient knowledge, they will tend to withdraw from external linkages and revert to in-house R&D alone. The available evidence is mixed, but on the whole it supports the idea that while the patterns of specialization and division of labor have changed over time, no clear trend towards extensive internalization and integration exists.18 Moreover, we submit that a modest decline in the number of alliances between large firms and NBFs is not inconsistent with the division of innovative labor. Instead it reflects greater learning about external linkages on the part of agents involved. Market forces are increasingly selecting the NBFs not only according to their ability to perform "good science," but also according to their ability to translate science into commercializable outcomes. Thus, fewer NBFs survive as more than transient start-up "gamblers," which has a negative scale effect on the number of alliances. More importantly, large firms, and especially the most innovative ones, have now become sufficiently familiar with the knowledge base of biotechnology, and particularly with its upstream scientific base. As the results discussed in the previous section would suggest, firms have therefore become more discerning in their choice of partners, and hence form fewer but potentially more valuable alliances. A final question is whether a division of innovative labor in biotechnology is socially desirable. Florida and Kenney (1990), inter alia, see the emergence of NBFs as socially undesirable. They claim that venture capital funding leads to a 17   Max Link, the CEO of Sandoz Pharma Ltd., epitomized this view: "We believe that the stronger a multinational is in a particular field, the higher the probability that a cooperation will yield positive results" (Scrip, 1991). 18   Some recent reports about the industry suggest a moderate decline in the number of alliances between large firms and NBFs (Burrill and Lee, 1991). On the other hand, Barabanti et al. (1992) find an increase in the total number of collaborations of 21 major pharmaceutical companies in biotechnology, as well as in the average number of collaborations per firm, in these three periods: up to 1983, 1984–1987, 1988–1991. Their results also indicate considerable volatility in the patterns of collaborative alliances.

OCR for page 188
Sources of Medical Technology: Universities and Industry demand for quick and high returns, which leads to a short-term focus and myopic decisionmaking. Moreover, in their view, the formation of a large number of NBFs causes a "fragmentation" of capabilities by spreading the available scientific and technological expertise too thinly. Furthermore, the "patent race" nature of NBFs suggests a misallocation of resources, with too many NBFs operating in the same niche. Hence they expect a decline in the comparative advantage of NBFs, with a given NBF losing its ''innovativeness," as well as NBFs of later vintage being less innovative than their predecessors.19 Implicitly, therefore, Florida and Kenney are of the view that the beneficial advantages of specialization and division of labor are outweighed by the costs of information exchange and integration, as well as the destructive competition for resources and markets that may accompany a division of labor.20 While a division of labor in innovation can, in fact, produce dispersion of capabilities, with implied high coordination costs, it is also true that fragmentation of capabilities is a more serious problem in instances where the knowledge base for innovation is tacit and context-specific. When the knowledge base is tacit and context-dependent, it is difficult to exchange technological information across organizational boundaries. As we have argued, effective information exchange has become easier. Increased understanding of scientific and technological principles, coupled with the use of computer-based models, makes it possible to generate and utilize information that is expressed in relatively general and abstract forms: that is, in forms that are common to different organizations and that can be applied to different contexts. This increased universality in the form of information eases information exchange for innovation, and encourages specialization and division of labor according to comparative advantage.21 Biotechnology exemplifies the increasing use of general and abstract knowledge for innovation, and the consequent possibilities that have arisen for a new division of labor in inventive activity. This perspective leads us to suspect that the division of labor in biotechnology will prove to be more enduring than traditional Chandlerian or transaction-cost perspectives would suggest. The social desirability of a division of innovative labor in general arises in part from the differences in comparative advantages that depend upon the size and nature of different organizational forms. As many authors have argued, large 19   However, Lerner (1991) finds no evidence to suggest that the research productivity of NBFs (as measured by citation-weighted patents) in his sample declines over time. Lerner also reports that the research productivity of NBFs in his sample is significantly higher than that of the matched sample of large companies. 20   It is interesting to note that Merges and Nelson (1990) take the opposite view regarding the desirability of a large number of start-ups. They argue that, from an evolutionary perspective, diversity is likely to promote the rate of growth. 21   We spell out this argument more fully in Arora and Gambardella (1994a) where we relate it to the feasibility and social desirability of a division of innovative labor more generally.

OCR for page 188
Sources of Medical Technology: Universities and Industry and small firms have "natural" comparative advantages in different stages of innovation (see, for example, Arrow, 1962). Smaller firms have flexible and informal organizational structures, which facilitate inventiveness. As Arrow suggests, the organizational distance between inventors and the people making decisions on the internal allocation of resources to innovation projects is greater in larger firms. The greater distance creates loss of information in communication and greater asymmetry of information. Smaller companies are, therefore, better able to undertake more novel (and riskier) projects, provided they can finance such projects. In contrast, larger firms have superior organizational and financial capabilities (including easier access to financial markets because of information asymmetries and the like) for large-scale development, production, and marketing. Consistent with our analysis of comparative advantage, NBFs have found "go-it-alone" strategies easier in the less lucrative market for diagnostic kits, especially in-vitro kits, because the development and regulatory procedures are simpler. Even in such markets, however, complementary assets such as the ability to supply diagnostic equipment and previous experience in dealing with hospitals have proved to be of great importance for established companies such as Abbott. Lacking these complementary assets for commercialization, NBFs such as Hybritech and Genetic Systems were forced to sell out, to Eli Lilly and Bristol-Myers, respectively. (See also Arora and Gambardella, 1990; Barabanti et al., 1992; and Orsenigo, 1989.) Indeed, most NBFs have found the transition from research to production and marketing a difficult one, and the research orientation of their founders and top management have frequently made it more so. Perhaps recognizing their comparative advantage, some NBFs have adopted the strategy articulated by Benzon Pharma, a small Danish drug company with a number of alliances with larger firms such as SmithKline Beecham and Schering-Plough. Benzon's strategy is to work on new scientific areas that are applicable to many kinds of drugs. All projects are carried out with the possibility of linking them to the programs of other, usually much larger, pharmaceutical companies (Financial Times, 1990). Only a few NBFs like Amgen and Biogen have successfully acquired the complementary assets for commercialization, the former with erythropoietin and colony stimulating factors, and the latter with the interferons. Even where NBFs have been able to acquire some downstream capabilities, they remain extremely vulnerable to adverse Food and Drug Administration (FDA) findings and other "misfortunes" in production or marketing. The tissue plasminogen activator (t-PA) disappointment led to the acquisition of 60 percent of the capital of Genentech by Hoffman La Roche.22 Similar problems led to the 22   Interestingly enough, Hoffman La Roche preserved in substantial ways Genentech's autonomy as an independent company, suggesting that this acquisition is more a match of complementary assets than a prototypical case of vertical integration.

OCR for page 188
Sources of Medical Technology: Universities and Industry merger of Cetus, one of the earliest biotech firms, with Chiron, another "large" NBF. Finally, the value of Centocor's stock dropped by 25 percent after the FDA questioned data on its new septic shock drug. Thereafter, Centocor signed an agreement with Eli Lilly, which hopes to use its expertise in development and FDA regulation, as well as its commercialization capabilities, to revamp Centocor's compound.23 A very important question, suggested by the conceptualization of the process as one of division of labor, and one that has not yet been asked in the literature, has to do with the creative destruction of competencies implicit in the downward integration of NBFs. Given that some of the most innovative NBFs, such as Amgen and Biogen, are converting themselves into full-fledged pharmaceutical companies, it would be useful to ask whether the process is inevitable and (a related question) whether it is socially efficient. In other words, even if private incentives point toward integration, one must ask whether these represent the benefits that authors such as Florida and Kenney have pointed toward, or whether these are responses to market imperfections that are widely known but whose ramifications may not be well understood. The answer to the question will have important implications for policy. For instance, strengthening the definition and enforcement of intellectual property rights may slow down or even stop the process of downward integration. At this stage we shall pose it as an important and interesting question for further research. SUMMARY AND CONCLUSIONS In this essay we have examined the changing patterns of collaborative alliances in biotechnology. Based on some previous studies on the topic, we discussed how innovation in biotechnology is the outcome of systematic interactions between universities, NBFs, and large pharmaceutical firms. We characterized the different strategies of external linkages from the viewpoint of large firms, and found that they were mutually complementary in that they target distinct but synergistic objectives of the firm. The division of innovative labor that we have studied is taking place in the context of a noticeable increase in the use of general and abstract knowledge in innovation. The breadth of applicability of universal principles offers the possibility of subdivision of the innovation process 23   In this context, it is interesting to note that, while Centocor and the other biotech firms mentioned above seek partnerships for resources and expertise downstream, Eli Lilly is moving in the other direction. It has recently taken serious steps to combat its "not invented here" syndrome. As explicitly declared by its top managers, Lilly seeks extensive alliances with partners that can supply new ideas and products to be pumped into its pipeline. In the early 1990s, Lilly signed a number of such alliances with small/medium research-intensive companies in fields such as molecular design of drugs (Agouron), serotonin-based drugs (Synoptic Pharma), and trauma infections (Business Week, 1992).

OCR for page 188
Sources of Medical Technology: Universities and Industry by allowing "pieces" of technological information to be recombined in a useful and cost-effective manner. However, for a "market" in technology to function effectively, institutions relating to intellectual property, as well as institutions for financing start-up ventures, have to adapt. We argued that "buying" technological information requires a great deal of prior competence and capabilities, very similar to those required for generating new technological information. The nature of the "market" for information and its functioning, and their implications for the organization and rate and direction of innovation, are not fully understood and remain important areas of future research. ACKNOWLEDGMENTS We would like to thank Rebecca Henderson and the other participants at the Institute of Medicine workshop, "The University-Industry Interface and Medical Innovation," for comments and discussion. We are especially grateful to Gerald Laubach for detailed comments that were very helpful. All errors are ours. REFERENCES Arora, A., and Gambardella, A. 1990. Complementarities and external linkages: The strategies of the large firms in biotechnology. Journal of Industrial Economics 37:361–379. Arora, A., and Gambardella, A. 1992. New trends in technological change. Rivista Internazionale di Scienze Sociali 3(July-September):259–277. Arora, A., and Gambardella, A. 1994a. Changing technology of technological change. Research Policy 23:523–532. Arora, A., and Gambardella, A. 1994b. Evaluating scientific information and utilizing it: Scientific knowledge, technological capability, and external linkages in biotechnology. Journal of Economic Behavior and Organization 24:91–114. Arrow, K. 1962. Economic welfare and the allocation of resources for invention. In: The Rate and Direction of Inventive Activity. Princeton, N.J.: Princeton University Press. Barabanti, P., Gambardella, A., and Orsenigo, L. 1992. The evolution of the forms of collaboration in biotechnology. Unpublished manuscript. Milan: Bocconi University. Bioscan. 1991. The Biotechnology Corporate Directory Service (April). Phoenix, Ariz.: Oryx Press. Burrill, G. S. 1989. Biotech 89: Commercialization. New York: Mary Ann Liebert, Inc. Burrill, G. S. 1990. Biotech 90: Into the Next Decade. New York: Mary Ann Liebert, Inc. Burrill, G. S., and Lee, K. B. 1991. Biotech '92: Promise to Reality . New York: Mary Ann Liebert, Inc. Business Week. 1992. Lilly looks for a shot of adrenalin. Nov. 23, pp. 70–72. Chang, L. 1992. Biotechnology and the organization of innovation. Unpublished manuscript. Stanford, Calif.: Department of Economics, Stanford University.

OCR for page 188
Sources of Medical Technology: Universities and Industry Cohen, W., and Levinthal, D. 1989. Innovation and learning: The two faces of R&D. Economic Journal 99(September):569–596. Daly, P. 1985. The Biotechnology Business. London: Francis Pinter. David, P., Mowery, D., and Steinmueller, E. 1988. The economic analysis of payoffs from basic research: An examination of the case of particle physics research. CEPR Working Paper 122 (January). Stanford, Calif.: Stanford University. Financial Times. 1990. Making for the gaps left by giants. February 14, p. 16. Florida, R., and Kenney, M. 1990. The Breakthrough Illusion: Corporate America's Failure to Move From Innovation to Mass Production. New York: Basic Books. Gambardella, A. 1992. Competitive advantages from in-house basic research: The U.S. pharmaceutical industry in the 1980s. Research Policy 21:391–407. Gambardella, A. 1995. Science and Innovation: The U.S. Pharmaceutical Industry in the 1980s. Ph.D. thesis. Cambridge, Mass.: Cambridge University Press. Grabowski, H. 1991. The changing economics of pharmaceutical research and development. In: Institute of Medicine. Medical Innovation at the Crossroads, vol. 2. The Changing Economics of Medical Technology. A. Gelijns and E. Halm, eds. Washington, D.C.: National Academy Press, pp. 35–52. Halperin, M., and Chakrabarti, A. 1987. Firm and industry characteristics influencing publications of scientists in large American companies. R&D Management 17(3):169–173. Imai, K. 1988. Japan's corporate networks. Paper prepared for the Minit-JPERC Conference, Stanford University, Stanford, Calif., August 22–24, 1988. Kenney, M. 1986. Biotechnology: The University-Industry Complex. New Haven, Conn.: Yale University Press. Lerner, J. 1991. New biotechnology firms, capital markets, and innovation. Unpublished manuscript. Cambridge, Mass.: Department of Economics, Harvard University. Merges, P., and Nelson, R. 1990. On the complex economics of patent scope. Columbia Law Review 90(4):840–916. Mowery, D. C. 1981. The Emergence and Growth of Industrial Research in American Manufacturing 1899–1945. Ph.D. thesis. Stanford, Calif.: Stanford University. Mowery, D. C., ed. 1988. International Collaborative Ventures in U.S. Manufacturing. Cambridge, Mass.: Ballinger Publishing Co. Narin, F., et al. 1987. Patents as indicators of corporate technological strength. Research Policy 16:143–155. Nelson, R. 1961. Uncertainty, learning, and the economics of parallel research and development efforts. Review of Economics and Statistics 43:351–364. Office of Technology Assessment, U.S. Congress. 1984. Commercial Biotechnology: An International Analysis. OTA-BA-218. Washington, D.C.: U.S. Congress. Orsenigo, L. 1989. The Emergence of Biotechnology: Institutions and Markets in Industrial Innovation. London: Francis Pinter. Pisano, G. P. 1990. The R&D boundaries of the firm: An empirical analysis. Administrative Science Quarterly 35:153–176. Pisano, G. P., Shan, W., and Teece, D. J. 1988. Joint ventures and collaboration in the biotechnology industry. In: D. C. Mowery, ed. International Collaborative Ventures in U.S. Manufacturing. Cambridge, Mass.: Ballinger Publishing Co. Rosenberg, N. 1990. Why do firms do basic research? Research Policy 19:165–174. Scrip. 1991. Sandoz' biotech discovery initiative. March 20, p. 7.

OCR for page 188
Sources of Medical Technology: Universities and Industry Swann, J. P. 1988. Academic Scientists and the Pharmaceutical Industry . Baltimore, Md.: The John Hopkins University Press. Telling, F. W. 1992. Managed care and pharmaceutical innovation. In: Institute of Medicine. 1992. Medical Innovation at the Crossroads , vol. 3. Technology and Health Care in an Era of Limits. A. C. Gelijns, ed. Washington, D.C.: National Academy Press. Thackeray, A. 1982. University scientists in the chemical industry. In: National Science Board, ed. University-Industry Research Relationships . Washington, D.C.: U.S. Government Printing Office. Thomas, L. G. 1988. Multifirm strategies in the U.S. pharmaceutical industry. In: D. C. Mowery, ed. International Collaborative Ventures in U.S. Manufacturing. Cambridge, Mass.: Ballinger Publishing Co. Williamson, O. E. 1985. The Economic Institutions of Capitalism. New York: Free Press.

OCR for page 188
Sources of Medical Technology: Universities and Industry This page in the original is blank.