If a new system of innovation is to fully emerge and prosper, federal science and technology policies--which helped create the broad institutional contours of the post- war R&D system--must now be adapted to support the new institutional relationships between industry, universities, and government. Fostering and expanding these new alliances are particularly critical if the nation is to develop long-term approaches to supplement the increasingly short-term investments companies and institutions now make. In short, the key enabler to strengthening innovation, and its dissemination and absorption, may lie as much in increasing the social capital of our productive sectors as in direct investments in science and technology.

How Social Capital Increases Innovation

Adam Smith and other classical economists recognized that firms require an underlying fabric of shared values and understanding to make division of labor work. Smith's observation is even more true today. When research partnerships and consortia succeed, the "glue" that holds them together is not simply contracts that detail every aspect of these complex and dynamic relationships (although contracts are, of course, important). Nor even is it the information systems that link networks of organizations (although these networks facilitate information sharing). Rather, the glue that makes collaboration feasible in the New Economy is composed of trust and a norm of reciprocity, or enlightened self-interest, among decision makers in networks.

Successful innovation networks rely on collaboration and information sharing to work. Networks absent collaboration--when they function at all--result in unwieldy transaction costs, duplication of effort, lack of coordination, and costly contractual disputes. Research on how manufacturing firms have implemented electronic data interchange (EDI) illustrates this. Firms that successfully implemented EDI were those that could effectively cooperate to formulate shared goals, resources, and incentives ¹². Success was dependent on "interpersonal, organizational, and cultural" factors rather than technical ones. Moreover, it appears that firms that effectively cooperate and communicate internally (by means of cross-functional teams and systematic, decentralized problem-solving at all levels) are precisely the ones that can best collaborate with other organizations. A network culture inside an organization facilitates the creation of networks outside.¹³

This "glue" or social capital is a critical component of the value created by these cooperative relationships in terms of economic performance and innovative capacity ¹⁴. Social capital is as important as physical capital (plant, equipment, and technology) and human capital (intellect, character, education, and training) in driving innovation and growth. The "stock" of social capital is increased when a network of organizations develops the ability to work in collaboration to promote mutual productive gain ¹⁵. Social capital indicates those properties of organization, notably networks, shared norms, and trust, that lower the costs of coordination and cooperation and thereby increase collective productivity. Thus, well- functioning partnerships, consortia, and networks are in and of themselves a form of social capital ¹⁶. Trust develops over time as individuals gain confidence in the reliability ofothers through a series of interactions ¹⁷. Norms of appropriate behavior develop as a social contract is negotiated largely through interactions among actors in a network. The norm of reciprocity is fundamental to productive relationships. Closely linked to reciprocity is the norm that actors will forgo their immediate self-interest to act not only in the interest of the group but also in their own long-term self-interest. "Favors" that must be repaid quickly and accounts that are monitored closely in a network indicate a lack of social capital.

Networks are more efficient organizational forms for innovation in the New Economy because they facilitate enhanced learning. Compared to large, hierarchical structures, networks more effectively scan the environment for potentially significant events, more accurately interpret environmental change, and more creatively and adaptively craft responses to change. Superior ability to scan the business environment for new developments and interpret them accurately translates into heightened capacity for timely and responsive innovation. Dense social networks can encourage experimentation and entrepreneurship among actors because they represent an optimal mix of collaboration and competition. Network members compete fiercely but also collectively process and share information about environmental changes including markets, regulations, technologies, and opportunities.

In other words, actors in a collaborative network learn interactively. They learn of new technologies, opportunities, challenges, and the outcome of transactions more quickly because of the density of interaction within the network. Learning is of a higher quality because it is subject to discussion and debate among counterparts whose perspectives and backgrounds may differ. And innovation itself is stimulated. Studies have found that successful innovation is characterized by the easy flow of ideas, technology, and people across institutional boundaries. Organizations that exhibit these characteristics have been called learning organizations. Similarly, geographic regions that include highly adaptive industry networks have been termed "learning regions."¹⁸

In contrast, vertically organized firms and institutions tend toward characteristics that adversely affect learning or information-processing capacity. Such firms typically possess: inward, insular foci; unproductive levels of hierarchy within large firms and bureaucracies; unproductive levels of secrecy and organizational loyalty that dampen information sharing within and across professions; value placed on institutional stability and organizational autonomy that fit poorly with a turbulent economic, technological, Appendix 1: Collaboration Drives Silicon Valley globally competitive environment; authority centralized at unproductive levels; and predominantly vertical flows of information which tend to be slower, biased, and thus less reliable.

Countries and regions in which such organizations dominate cannot innovate effectively. In fact, attempts at partnerships and consortia under these conditions of low trust lead either to complete failure or to poorly performing, non-collaborative networks. A key challenge of government in the New Economy is not only to transform itself into a learning network organization, but to also gain the ability to support and engage as a partner in fully functioning collaborative networks, in part to allow it to effectively pursue its own goals, including mission-related research.

These are powerful factors in explaining innovation. Experts on science and technology policy have used the concept to explain why innovation varies so much among countries with similar endowments of financial, physical, and human capital ¹⁹. Yet, traditional economic perspectives that focus on short-term self interest and individual transactions have failed to recognize the importance of social capital as a key to innovation.

The Federal Role in Promoting Collaborative R&D

Because collaboration and social capital are critical for innovation in the New Economy, federal research policies need to focus explicitly on building social capital and promoting collaborative research. PPI proposes three policy tools to lead to a higher share of federal support for research being translated into innovation and economic growth. All three recognize the importance of the development of an iterative and non-sequential process of innovation, the need for closer ties between developers and users of science and technology, and the importance of social capital.

1. Collaborative Research and Experimentation (R&E) Tax Credit

Since 1981, companies conducting research in the United States have been eligible for a 20 percent tax credit on research and experimentation investments above what was invested during a "base period." The rationale for such a tax credit is that firms are not able to capture all the benefits from R&D as a share of the benefits "spill over" to other firms or to customers. This inability to capture all the benefits leads firms to underinvest in research relative to what would be socially optimal. Studies by economists suggest that the rates of return for society from a company's R&D spending are significantly higher than the company's own rate of return.

The social benefits from collaborative research are likely to be significantly higher than company-specific proprietary research. In addition, the social rate of return from company-funded basic research is even higher (over 150 percent) than company applied research and development. Collaborative research, whether in partnership with a university, national laboratory, or industry consortium, is more likely to be exploratory and earlier stage than research conducted by a single company. Moreover, because the research is shared from its inception, the benefits are less likely to be fully captured by an individual firm. As a result, because spillovers from collaborative research are greater, firms will tend to underinvest in this even more than in individual research. This suggests that the federal government should support collaborative research through a more generous R&E tax credit.

PPI proposes that all company expenditures, not just incremental expenditures, on collaborative R&E be eligible for a 20 percent R&E tax credit. Investments in collaborative research consortia of at least 5 companies or investments by a company for research conducted at a U.S. university or federal laboratory would qualify as collaborative R&E. A bill introduced by Congressman Amo Houghton in May 1998 with bipartisan support (H.R. 3857), would amend the IRS code to allow a 20 percent research credit for expenses attributable to collaborative research consortia. PPI advocates extending the credit to also include industry funding of university and federal laboratory research. A bill introduced by Senator Jeff Bingaman (S. 2268) applies a 20 percent tax credit to both kinds of research. It is not clear how much this more expansive tax credit would cost in immediate revenue loss but it appears that it could be in the range of $150-200 million per year.

While this would be a significantly higher tax credit than currently exists, it is still not as generous as the policies of some other countries or state governments in the United States. For example, in Canada, all R&D is subject to a 20 percent flat R&D tax credit. Massachusetts provides a 10 percent credit for industry R&D, but a 15 percent credit for industry-sponsored university research.

A collaborative R&D tax credit would have numerous benefits. It would:

2. Industry Research Alliances

The debate over science and technology policy has tended to oscillate between those who argue that the federal government should fund industry to conduct generic pre-competitive R&D and those who maintain that money should be spent on curiosity- directed basic research at universities. But this is a false dichotomy. There is no reason why some share of university basic research cannot be oriented toward problems and technical areas that are more likely to have economic or social payoffs to the nation. Science analyst Donald Stokes has described three kinds of research: purely basic research (work inspired by quest for understanding, not by potential use); purely applied (work motivated only by potential use); and strategic research (research that is inspired by both potential use and fundamental understanding). Moreover, there is widespread recognition in the research community that drawing a bright line between basic and applied research no longer makes sense. ²¹

As Stanford economist Paul Romer states:

[Companies are] finding that while there are a billion haystacks in which there will be some very valuable needles, it is an enormously expensive proposition to go about looking underneath every one. So what they are asking themselves is, how do you allocate the resources that you devote to research most effectively? They are realizing that they cannot just give scientists lots of money and let them follow their curiosity. If they do, this form of tax-and-subsidy system runs the risk of dissipating efforts by looking in too many different directions that don't necessarily lead to the highest returns for shareholders.²²

As a result, companies have restructured their programs so that exploratory research is more connected to technology usage and commercial opportunity. The same principle needs to apply to at least a share of federal investments in research if the nation is to increase competitiveness, productivity, and living standards, and to solve pressing environmental and health challenges.

One way to better link economic goals with scientific research is to encourage the formation of industry research alliances that fund academic research. There are numerous examples of successful university-industry partnerships. For example, 18 wireless communications companies have formed a research consortium with the University of California-San Diego Engineering Department to work on advanced research related to the industry. Industry invests because research is performed in areas that are too risky, too long term, and too generic for any one company to invest in. The university invests to ensure that its scientists remain at the cutting edge of their scientific disciplines and work on scientifically and technically demanding tasks.

Similarly, the Semiconductor Research Corporation (SRC)--a non-profit research consortium of 36 companies and federal government agencies--plans, invests in, and manages a low-overhead, industry-driven, pre-competitive research program that addresses the needs of the Semiconductor Industry Association's National Technology Roadmap for Semiconductors. By continuously refining the industry-driven research agenda, the SRC provides a valuable forum for industry, university, and government scientists to collaborate and share knowledge, which is essential to advancement in the semiconductor industry's highly competitive environment. A new subsidiary of SRC, the Microelectronics Advanced Research Corporation (MARCO), is establishing research centers at American universities to focus on long-term, more basic research identified in the semiconductor technology roadmap. Jointly funded by government and industry, the program will invest up to $20 million in 1998.

The National Center for the Manufacturing Sciences (NCMS), located in Ann Arbor, Michigan, is a collaborative research network that includes approximately 50 large corporations and hundreds of medium and small firms. NCMS uses the collaborative model to develop a variety of distinct manufacturing processes. The center's annual budget of $80-100 million comes from industry as well as $20-25 million in DOD funds focused on particular manufacturing processes. Partnerships developed under the umbrella of the center have been responsible not only for new technological applications but also for process improvements such as rapid prototyping using computer simulation.²³

Collaborative efforts to decrease time to market, or product development cycles, will become increasingly important to U.S. competitiveness during the next decade. Working with the Massachusetts Institute of Technology (MIT), a network of firms including Xerox, Polaroid, IT&T Industrial, and Ford have developed the Center for Innovation in Product Development. The center, an engineering research center funded by the National Science Foundation, draws academic expertise in both engineering and management from MIT. It also takes university-industry collaborations to new levels by hosting research at both industrial and university venues and by requiring students to conduct part of their research at participating firms under the guidance of both university and industry researchers and engineers.²⁴

In order to both explicitly link industry objectives to research and build social capital for innovation, PPI proposes that Congress establish an Industry Research Alliance Challenge Grant Fund to match industry consortia funds invested in research at universities and federal labs. To be eligible for matching funding, firms would have to:

Such a process would definitely not entail "picking winners and losers" because industry, in conjunction with academic partners, would identify the broad technology areas critical for research. In fact, because the policy uses market mechanisms to fund R&D, it prevents government from picking winners and losers and scientists from pursuing research interests in isolation from societal needs and benefits. Government would fund proposals based on the amount of corporate matching funds and the quality of the proposal, but not according to any specific technology or industry. The proposed process would not constitute "corporate welfare" because funding would not be directed to industry but to universities or federal laboratories for R&D on generic, shared technology needs. In fact, because government would leverage industry funds, more money, not less, would flow to universities and national laboratories.

Finally, the effort would draw university and industry researchers closer together to understand common challenges, and in so doing would build the nation's innovation-based social capital. As Paul Romer states:

For the nation as a whole, an effective institutional arrangement for supporting technological advance must therefore support a high level of exploration and research in both private firms and in universities. Moreover, it must support a high degree of interaction between these two domains. Both people and ideas must move readily between them. If they do not, the university research can become sterile and irrelevant. Private sector efforts can lose the steady flow of new talent and new ideas that sustain its creatively.²⁵

This initiative would increase the share of federally funded university and laboratory research that is market relevant, and in so doing better adjust the balance between curiosity-directed research and research more directly related to societal need.

This initiative could be funded out of proposed increases in federal support for R&D. The Clinton Administration has proposed increasing R&D by 32 percent over the next five years. In addition, Sens. Bill Frist (R-TN), Jay Rockefeller (D-WV), Joe Lieberman (D-CT), Pete Domenici (R-NM), Phil Gramm (R-TX), Conrad Burns (R-MT), Jeff Bingaman (D-NM), and John Breaux (D-LA) have co-sponsored the Federal Research Investment Act (S. 2217), which would double federal expenditures for basic scientific, medical, and engineering research by 2010 (an increase of $32 billion per year). Rep. Joseph Kennedy (D-MA) has introduced a similar bill (HR 3660) in the House with bipartisan support. A share of this increase, perhaps 5-to-10 percent, could be allocated to industry research alliances.

3. State Technology Innovation Challenge Grants

The federal innovation system historically has focused on larger firms (typically multinational firms with large R&D units) and the approximately 30 first-tier, large research universities (which receive 47 percent of federal research funds to universities), usually in a small number of states (67 percent of federal support for R&D goes to 10 states). Both sets of institutions have played key roles in driving innovation and technological development in the United States. However, in the New Economy small- and medium-sized firms have become much more important to the nation's innovation system. Moreover, many of the nation's non-top-tier research universities and colleges have developed significant science and technology strengths, often in particular fields. These universities and colleges often play key roles working with industry in their region. Such regional technology collaboration between firms and higher education is increasingly important in the New Economy.

Yet it is hard for federal policy to focus explicitly on either smaller firms or smaller research universities, and especially on collaboration between the two. There are simply too many small and medium size firms, and universities and colleges for the federal government to engage meaningfully with them. In contrast, it is easier for states to work with small- and medium-sized firms and second-tier colleges and universities. As a result, any federal policy that seeks to stimulate collaborative R&D among small- and medium-sized firms and non-top-ranked universities should strengthen and support state efforts.

Most states have initiatives underway that could be expanded. Starting in the early 1980s, several states began to refocus their economic development efforts to promote technological innovation. They realized that R&D and innovation are drivers of the New Economy, and specifically that state economies prosper when they maintain a healthy research base closely linked to commercialization of technology. All 50 states have initiatives of some kind to promote technology-based economic development.

Most state programs focus on small and mid-size firms and second-tier universities. For example, under the leadership of Governor Richard Thornburgh, Pennsylvania established the Ben Franklin Partnership Program which provides matching grants primarily to small- and medium-sized firms to work collaboratively with Pennsylvania universities.

More recently, California created a partnership among state government, industry, and universities to promote economic development through support for research and workforce education. Called the Industry-University Cooperative Research Program, a major focus has been the development of commercial biotechnology in California. A key effort of this program, the University of California's Biotechnology STAR Project, invested $5 million of state and university funds and $7.5 million of industry funds during its first year, 1996, to develop 47 research partnerships. Projects include gene therapy to combat disease, the development of new agricultural products, conversion of municipal solid waste to fuel, and biomaterials research to develop new medical implants. Eighty percent of the firms involved in the program in 1997 were small businesses, of which half had fewer than 50 employees ²⁶. Independent evaluations of similar state programs show substantial economic gains and favorable cost-benefit ratios.

Yet, in part because of the healthy national economy and because some states have pursued high cost "smokestack chasing," state support for these technology programs has generally remained flat since the late 1980s. In addition, just as private R&D has focused more on product development and less on earlier, riskier research, state initiatives have tended to focus on later stage product development and technology diffusion, and less on earlier stage industry-university partnerships. In part, this is because it is more difficult for states to capture all the benefits from collaborative research within their borders. Research results flow to universities, suppliers and customers in other states. As a result, most states invest less in collaborative R&D efforts than is warranted.

As a result, PPI proposes that Congress establish a $250 million annual competitive matching grant fund for states to invest in university-industry and other technology and innovation network programs. States would be required to match the federal funds at a ratio of at least one-to-one and invest in joint university- industry or other collaborative industry-based innovation projects. This could take several forms, including university-industry research centers (such as those in New York, North Carolina, Ohio, and Utah) or matching grant university research projects (modeled perhaps after existing programs in states such as Arkansas, Maryland, Pennsylvania, and Rhode Island). Industry would be required to match all public funds one to one. Thus, $250 million in federal funds would be leveraged into at least $1 billion in additional funding, much of it conducted at non-top tier universities with small- and medium-sized firms as partners. In contrast to the Industry Research Alliances, these funds would be managed at the state level and focused on regional innovation partnerships.

Conclusion

If productivity and per-capita income in the United States are to grow at the levels experienced in the 1950s and 1960s, increased innovation will be key. However, in the New Economy innovation is increasingly based on collaboration, rapid learning, and networks--all three supported by strong investments in "social capital" for innovation.

The responsibility of the federal government to invest in research is clear and relates importantly to the ability of researchers in science and technology to collaborate in a variety of ways. While supporting scientific and technical advance is key, the return on investments will be greater if they also support institution-building to facilitate the creation of a rich and widespread system of collaborative innovation networks within industry, and between industry, universities, and federal laboratories. The proposals detailed here are steps to move our nation more firmly in this direction.

Appendix 1.
Collaboration Drives Silicon Valley

One of the best-known learning regions is Silicon Valley, California. The professional culture is both highly collaborative and intensely competitive. Non-proprietary professional and technical information typically is shared among employees and companies. Firms display low levels of vertical integration because of the high degree of outsourcing for inputs that occurs. A rich system of interconnections links producers, suppliers, and customers within and across related sectors.

Educational systems at all levels ensure a supply of skilled labor and provide training, retraining, and development of technical staff. At the graduate school level, these systems produce basic and applied research and researchers. But Stanford University plays a much more central role. Stanford spawned Hewlett-Packard, one of the Valley's central firms, as well as many other firms. Firm leaders and several venture capitalists have strong ties to the university and, thus, to one another. In many ways, Stanford University is the chief influence on the culture of Silicon Valley. Having a first-rate research university or laboratory as an integral part of a regional network rather than as an isolated "ivory tower" greatly strengthens the potential for scientific innovation within both the university and industry.

Innovation benefits from the strong supporting role played by institutions of higher learning. State and community college systems, through strong, focused engineering and technical training programs, supply and sustain high-quality technical employees able to function effectively in a networked environment. Strong networks or partnerships among government, universities, and industry help to ensure the supply of these experts. These partnerships also ensure the continued high quality of scientists and technical experts by translating new ideas, technologies, and methods from universities to industry and from industry to universities.

Source: Anna Lee Saxenian, Regional Advantage: Culture and Competition in Silicon Valley and Route 128 (Cambridge, MA: Harvard University Press, 1994).

Endnotes

1. See PPI Policy Briefing,"The Case for Technology in the Knowledge Economy: R&D, Economic Growth, and the Role of Government, Kenan P. Jarboe and Robert D. Atkinson, June 1998.

2. Booz, Allen and Hamilton, cited in John Micklethwalt and Adrian Wooldridge, The Witch Doctors (New York, NY: Time Books, 1996).

3. National Science Foundation, Science and Engineering Indicators, 1996. The Industrial Research Institute also reports that since 1993, the number of companies surveyed who reported that they were increasing total R&D more than 5 percent was more than twice as high as those reporting similar increases in basic research. The average length of an industry basic research project dropped from 21.6 months in 1991 to 16.7 months in 1996.

4. H. Brooks and L. Randazzese, "University-Industry Relations," chap. 14 in L. Branscomb and J. Keller, eds. Investing in Innovation: Creating A Research and Innovation Policy that Works (Cambridge, MA: MIT Press, 1998).

5. Ibid.

6. National Science Foundation, Science and Engineering Indicators, 1996.

7. Daniel Roos, Frank Field, and James Neely, "Industry Consortia," chap. 15 in Branscomb and Keller, eds., Investing in Innovation.

8. Between 1981 and 1993, the number of scientific papers written by industry researchers and involving collaboration from another sector (e.g., universities) increased by 123 percent. In contrast, academic collaborations increased only 33 percent. (National Science Board, Science and Engineering Indicators, 1996.)

9. One study of the sources of innovation found that more than 50 percent of technological innovations occur at the interfaces between organizations, rather than within the organization alone. (E. Von Hippel, Sources of Innovation, New York: Oxford University Press, 1988).

10. Michael Porter, The Competitive Advantage of Nations (New York: Free Press, 1990), pp. 149-154.

11. National Science Foundation, Science and Engineering Indicators, 1996, p. 158.

12. Harvard University, John F. Kennedy School of Government, Center for Science and International Affairs, Science, Technology and Public Policy Program, "Manufacturing Partnerships in the Digital Environment: Best Practices in CALS Implementation," December 1996.

13. Scott A. Snell and James W. Dean, Jr., "Integrated Manufacturing and Human Resource Management: A Human Capital Perspective," Academy of Management Journal, Vol. 35, no. 3 (1992), pp. 467-504.

14. The concept of social capital development as a policy tool is examined in detail in Jane Fountain, "Social Capital: A Key Enabler of Innovation," chap. 4 in Lewis Branscomb and James Keller, eds., Investing in Innovation. The concept, social capital, was originally outlined in James S. Coleman, Foundations of Social Theory (Cambridge: Harvard University Press, 1990); and Robert Putnam, Making Democracy Work: Civic Traditions in Modern Italy (Princeton, N.J.: Princeton University Press, 1993).

15. Francis Fukuyama, Trust: The Social Virtues and the Creation of Prosperity (NY: The Free Press, 1995).

16. Ronald S. Burt, Structural Holes: The Social Structure of Competition (Cambridge: Harvard University Press, 1992), p. 12.

17. For micro-level explanations of the development of interorganizational arrangements, see Peter Smith Ring and Andrew Van de Ven, "Developmental Processes of Cooperative Interorganizational Relationships," Academy of Management Review, Vol. 19, no. 1 (1994); Jane E. Fountain, "Trust as a Basis for Interorganizational Forms," paper delivered at conference on "Network Analysis and Innovations in Public Programs," University of Wisconsin at Madison, September 30, 1994.

18. Peter Senge, The Fifth Discipline: The Art and Practice of the Learning OrganizationM/em> (New York: Doubleday, 1990). Richard Florida, "Toward the Learning Region," Futures, Vol. 27, no. 5 (June 1995), pp. 527- 536.

19. Lewis M. Branscomb, "Social Capital: The Key Element in Science-Based Development," Science-Based Economic Development: Case Studies around the World, Annals of the New York Academy of Sciences, Vol. 798.

20. Coopers and Lybrand, "Economic Benefits of the R&D Tax Credit," January, 1998.

21. Basic Research White Paper R&D Magazine Online.

22. Paul Romer, Beyond Classical and Keynesian Macroeconomic Policy, Policy Options, July-Aug. 1994.

23. Aris Melissaratos, AAAS S&T Policy Yearbook 1998.

24. Maurice F. Holmes, "Industry-University Collaboration in Innovative Process Management," chap. 21 in AAAS S&T Policy Yearbook 1998.

25. Paul Romer, "Beyond Classical and Keynesian Macroeconomic Policy."

26. Susanna L. Huttner and Cherisa Yarkin, "California's R&D Partnerships for a Knowledge-Based Economy," chap. 19 in AAAS Science and Technology Policy Yearbook, 1998.

Dr. Jane E. Fountain is associate professor of public policy and a faculty affiliate of the Center for Business and Government, the Belfer Center for Science and International Affairs, and the Harvard Information Infrastructure Project at the John F. Kennedy School of Government, Harvard University. Dr. Robert Atkinson is director of the Technology, Innovation, and New Economy Project at PPI.