Capitalizing on Investments in Science and Technology (1999)

The United States' leadership in research across the spectrum of science and engineering is well known. Less well known to the general public is its effectiveness in transforming the results of research into concrete national benefits. This transforming process is called capitalization: Utilizing the results of research to advance national goals, such as maintaining a high standard of living, creating high-paying jobs, improving education, protecting the environment, enhancing personal and public health, ensuring national security, and deepening human understanding.

Innovation, technology transfer, and the commercialization of research have long been a focus of academic study and policy debate. In undertaking this study, the working group sought to better understand how science and technology investments by government and industry are transformed into national benefits.

The long-term role of science and technology in raising living standards has become more apparent in recent years, as the growth of science and technology based companies and industries has lifted U.S. economic performance. A striking example is the continuing explosion of innovation and business opportunities surrounding the Internet.

The groundwork for the Internet was laid by basic technology research in university and other laboratories, and was funded largely by the federal government, especially the Department of Defense (DoD) through the Defense Advanced Research Projects Agency (DARPA) (see Box 2-1). Starting in the late 1960s, DARPA helped to create the ARPANET so that universities could share expensive research computers.¹ During the 1970s a growing number of researchers used this network as a communications medium within the science and engineering community, and the National Science Foundation (NSF) became its primary sponsor. Software was developed to use the new "Internet": Physicists in Geneva invented the World Wide Web to share data via hypertext, and graduate students at a federally funded computer center in Illinois added a browser called Mosaic. Finally, private firms leapt into the field, offering to help isolated desktop

computers "get connected," and the Internet exploded into homes and businesses around the world.

Over three decades, the Internet has moved from a government-sponsored to a market-driven network, with much of the development taking place outside the realm of science. To illustrate the power of capitalization, consider that DARPA's original 1969 contract with Bolt, Beranek, and Newman, to set up the first nodes on the ARPANET, was for only $1 million. As this report was being prepared, the combined market value of just five networking firms, none of which existed before 1980, approached $150 billion.² One survey firm estimates that there were

102 million Internet users worldwide as of January 1998, and that the number of users has been almost doubling every year for the past decade (Matrix Information and Directory Services, 1998).

The economic benefits of capitalization are vital, and obvious. For example, most econometric studies of research and development (R&D) investments have found that the private returns on these investments exceed 20 percent and the social returns exceed 50 percent.³

In the case of the Global Positioning System (GPS), research investments made over a long period of time in a number of different fields resulted in a capability that first enhanced U.S. national security, and is now producing an explosion of civilian uses (Beyond Discovery, 1996b). In the case of monoclonal antibodies, research breakthroughs during the 1970s led directly to the development of tests that enable the United States and other countries to eliminate the risk of transmitting AIDS through blood transfusions (see Appendix A). This contribution to public health is now being amplified by significant economic returns.

As Richard Nelson (1998) notes, technological advance involves uncertainty in a fundamental way. The process is full of surprises, and it generally is not possible to predict the outcomes of research programs. For years, experts have been trying to develop useful models and definitions to categorize various part of the research and innovation process. R&D statistics and policy discussion often reflect assumptions of a linear model, by which innovation proceeds from fundamental discovery to applied research, and then to development and marketing. However, there is widespread recognition that this model is not adequate to describe most real-world innovations.

The late Donald Stokes (1997) explored this question in depth, dividing research activities into four "quadrants" according to whether they are performed through "considerations of use" or a "quest for fundamental understanding." Stokes succeeds in showing that the progress of research is as complex as the motivations and abilities of the people who perform it. For example, sometimes major advances in fundamental knowledge are made by those working on practical, short-term problems.

In developing the idea of capitalization as a process of realizing returns on investments in research, the working group has adopted some conventional terms, such as basic (or fundamental) and applied research.⁴ Yet the examination of specific examples reveals the problematic nature of such categories. Even the simplest research project involves complex flows of information between applications, under-

lying principles, and work already done by others. The capitalization process in its entirety is a field of complex interactions and feedback loops between individuals, institutions, and the environment in which it occurs.

Some radical research breakthroughs lead to quick capitalization by existing companies and industries; a number of significant advances in catalysis were capitalized upon quickly. Other radical innovations prompt new companies to emerge; this was the case with monoclonal antibodies. The life-sciences and health-oriented research communities have produced a number of breakthroughs leading quickly and directly to capitalization. In other cases, capitalization results from incremental improvements in the design or manufacturing process of proven products. Although the semiconductor industry has benefited from several major breakthroughs, much of the work responsible for the development of successively more powerful integrated circuits over the past 30 years has been incremental. The improvement of chemical processes through advances in catalysis has benefited from both radical and incremental advances (see Appendix A).

Other successful examples of capitalization examined by the working group have depended upon accumulated advances in several different fields. In the case of speech recognition, advances in the modeling of human speech and software design, combined with the vastly improved performance of computers, enabled the first commercial applications during the 1980s. This occurred after decades of government and industry research.

In light of the complexity and unpredictability of the capitalization process, the working group decided not to attempt a particular model of capitalization, although it has made use of definitions and categories developed by others. The working group also has confirmed that capitalization, despite its complexity, is amenable to increased understanding and improvement through effective public policies and private strategies. Four elements of the process are discussed in this report: (1) research and research investments, (2) the environment for capitalizing, (3) human resources, and (4) partnerships and other cooperation between sectors.

Investments in research, development, and commercialization create fuel for the engine of capitalization. Because investments in research must compete with other national priorities, research funding is a topic of perennial debate among policy makers. The federal government provides over one-third of total research funding (the rest comes primarily from industry), and so, the debate is a public one, flavored by the pressures and demands of the political process.⁵ What kind of research should be funded? Why should it be funded by the government? How do we know it will bring the results we want?

Research never fits easily into discrete categories. Much basic science and engineering research is performed with specific uses in mind, especially uses imagined by the funder; both applied research and technology development often provide new questions and tools for basic research and advance the frontiers of fundamental understanding. Many of the most interesting and useful science and engineering questions lie in the gray zone between the quest for fundamental knowledge and the development of specific products.

Just as it is not easy to classify research activities, it is not easy to decide what kinds of research deserve support. Many products that today serve the public good also are sold commercially and grew out of a complex ferment of investigation. For example, early in its development, research into magnetic resonance imaging (MRI) was led by academic scientists based in the United Kingdom.⁶ Some of them were physicists interested in imaging phenomena, whereas others were medical researchers interested in clinical diagnostics. The British government supported both kinds of research, which have resulted in one of the most important recent advances in medical technology. MRI produces a two- or three-dimensional image of internal body structures that previously required invasive surgery or arthroscopic procedures. Once the potential uses of MRI became clear, commercial firms in several countries took over problem-solving and development.

The examples examined by the panel show that the federal government and industry have vital and complementary roles in funding research. Long-term and stable federal support has maintained U.S. capabilities to perform research at the frontiers of all major fields, and has been critical to capitalization. A diverse funding portfolio is characterized by multiple funders supporting research in industry, universities, and government laboratories, and by competition between researchers and institutions. Such a portfolio is vital to sustain the search for new knowledge, the growth in our stock of scientific and engineering human capital, and the necessary infrastructure.

It generally is understood that public investment in research is critically important to achieving societal goals. The federal government funds a large portion of U.S. basic research, and the size and allocation of funding are perennial topics of debate, featuring considerable swings and variations. Figures 2-1 through 2-4 show the longer-term trends in public and private support.

The federal component of research funding, because of its size and stability, is essential for several reasons: it can support complex laboratory facilities unavailable anywhere else, it can sustain long-term research that leads to technologies unimagined when the research was initiated, it educates the nation's scientists and engineers, it helps universities to maintain free access to knowledge, and it can pay for infrastructure and instrumentation technologies essential to research.

Figure 2-1

U.S. Research and Development Expenditures (Current Dollars) . Note: industry figures before and after 1991 may not be directly comparable due to changes in the survey. Source: National Science Foundation, Science and Engineering Indicators, 1998. Appendix Table 4-5.

For example, federal government support played an integral role in the development of the GPS.⁷ GPS satellites continually send out radio signals giving their exact position and time. Military and civilian users with GPS receivers can pinpoint their position on Earth's surface with a high degree of accuracy using these signals. The development and deployment of GPS required $12 billion in DoD funding and years of effort on the part of DoD and its contractors. In addition to

Figure 2-2

U.S. Research and Development Expenditures (Constant Dollars). Note: Industry figures before and after 1991 may not be directly comparable due to changes in the survey. Source: National Science Foundation, Science and Engineering Indicators, 1998. Appendix Table 4-6.

R&D efforts directly related to the system, GPS draws on a range of science and engineering advances generated in large part through federal support, including satellite launch and control technologies, microwave communication, and microelectronics. Atomic clocks, which are carried on every GPS satellite, were enabled by fundamental insights in quantum physics prior to World War II, with some important development work during the 1950s supported by the National Bureau of Standards [now the National Institute of Standards and Technology (NIST)].

Originally developed and deployed as a tool for the U.S. military, GPS is increasingly important as an infrastructure for civilian travel and navigation. To-

Figure 2-3

R&D Spending as a Percentage of GDP. Note: Industry figures before and after 1991 may not be directly comparable due to changes in the survey. Source: National Science Foundation, http://www.nsf.gov/sbe/srs/natpat97/start.htm, Table 7.

Figure 2-4

Basic Research Spending by Source. Note: Industry figures before and after 1991 may not be directly comparable due to changes in the survey. Source: National Science Foundation, Science and Engineering Indicators, 1998. Appendix Table 4-10.

day, sailboats, crop dusters, automobiles, and backpackers can all carry GPS receivers. Combined with computerized "yellow pages," the GPS will allow travelers everywhere to find a local restaurant, gas station, or hospital in an instant. The worldwide market for products and services enabled by GPS is expected to surpass $30 billion in the next decade. GPS illustrates that capitalization on research often occurs in complex, unpredictable, and nonlinear ways. The case also shows that

important science and engineering applications often rely on research advances in a number of fields and supported by a variety of funders.

Another field that depends on publicly funded research is human gene testing, where a half-century of basic biological research has led to discovery of over 50 disease genes. Today, this knowledge gives doctors a better chance of detecting disorders early and developing treatments. Prenatal genetic testing for fatal or debilitating conditions, such as Tay-Sachs disease, is reducing their incidence in the general population (Beyond Discovery, 1996b).

In terms of economic value and many Americans' work habits, federal support of computer science and telecommunications has been of utmost importance. A survey by the Computer Science and Telecommunications Board of the National Research Council found that federally funded university research underlies many commercially important technologies that evolved between 1965 and 1994, including time-sharing, graphics, networking, workstations, windows, RISC, VLSI design, and parallel computing. It also presented a model that graphically demonstrates how complex and nonlinear is the innovation process [CTSB (1997); see also CTSB (1998)].

In several of the examples considered by the working group, the ultimate success in capitalization has depended on industrial and other private investments. The invention of the transistor at AT&T Bell Laboratories over 50 years ago and its subsequent application is a well-known example. Similarly, in catalysis, large chemical and petrochemical companies in the United States and Europe produced many of the major advances in their own labs or funded the work at universities (see Appendix A).

The importance of government research investments in developing the computer industry was discussed in the preceding section; industry investments have been just as critical. For example, the Xerox Palo Alto Research Center (PARC), founded in 1970, has played a key role in developing laser printers, graphical user interfaces, object-oriented programming languages, and Ethernet local area networks.⁸ Although Xerox itself did not capitalize on many of PARC's advances, particularly those underlying the personal computer, they were commercialized by Apple, IBM, and a number of start-up companies, including Adobe Systems, that were led by PARC alumni. In 1988, Xerox Technology Ventures was established to create entrepreneurial companies, which are owned jointly by Xerox and by employees, to capitalize on promising in-house technologies.⁹

Today, companies such as SmithKline Beecham and Merck are investing in long-term research in bioinformatics, a promising area combining tools and insights from the life sciences and computer science (Marshal, 1996).

The importance of financing for new science- and technology-based companies, a particular type of private investment, is examined in more detail in the next section.

The diversity of the U.S. research effort and its funding are sources of strength. It is a system that thrives on pluralism, with many sources of support, many performers, and a maze of linkages among funders, performers, and users of research. A diverse research culture also contributes to competition among researchers and helps to avoid overspecialization or neglect of potentially important fields.

The diversity and pluralism of the U.S. science and engineering enterprise is illustrated by the pattern of federal government support. In the United States, agencies pursuing defense, health, space exploration, and energy missions provide the bulk of federal R&D support, with the NSF also playing an important role. This contrasts with some other industrialized countries, where government agencies responsible for science and technology, economic development and education play the major role in supporting research, particularly in nondefense areas.¹⁰

In several fields examined by the working group, U.S. ability to capitalize on research has been enhanced by its ability to work at the forefront of all major science and engineering fields. In the field of monoclonal antibodies, excellence in immunology research allowed U.S. researchers and start-up companies to capitalize on research advances made abroad. Excellence in computer science and the life sciences is allowing U.S. scientists to capitalize quickly in the area of bioinformatics.

A truism holds that once the results of research are published, that knowledge becomes a public good, equally available to all. Availability, however, is no guarantee of the ability to utilize public information advantageously. To capitalize on a research result, one needs the technical ability to understand it and capture its benefits, the availability of complementary research inputs, the economic ability to finance development and commercialization, and the regulatory protection of ownership. When scientists working at IBM's laboratory in Switzerland found that superconductivity could occur at temperatures of 40 K, U.S.-based scientists were quickly able to demonstrate superconductivity at temperatures above that of liquid nitrogen, but only because they were already conducting research on superconductivity. In other words, the human capital necessary for capitalization is a "joint product" with performance of basic research. This has been the most productive way to produce this type of human capital.

The importance of research diversity is illustrated in a four-year study by Stanford Research Institute to analyze the driving forces behind crucial technologies. The first three technologies studied by SRI appear to have developed by

quite different routes. For reaction injection molding (RIM), the primary driving force appeared to be demand from the auto industry for RIM products, spurred by government safety regulations. Industry conducted much of the necessary research itself, and both industry and government supported critical academic research by University of Minnesota chemical engineer Christopher Macosko. In the case of MRI, crucial forces were basic science research on nuclear magnetic spectroscopy and educational support for graduate students. In the case of the Internet, the key forces were sustained government funding, flexible university research, and visionary leadership (Stanford Research Institute, 1997). Each of these programs allowed maximum flexibility on the part of researchers; all culminated in outcomes whose dimensions had not been imagined beforehand.

A number of the examples considered by the working group illustrate how long-term investments in research and advanced education have allowed the United States to capitalize on them and to pioneer new fields. In speech recognition, for example, industry and government both invested for a long period of time—a period of "research gestation"—before the field gained momentum.

In all of these cases, science and engineering research and technology development were intertwined. Harvey Brooks has suggested that "pure technology" is as appropriate for public investment as "pure science." He cites the example of radioisotopes and stable-isotope tracers; studies were supported by the Atomic Energy Commission for many years before the medical and biological communities learned how to use them. Today, they are vital tools for diagnosing and treating disease and for basic biological research. Only by creating a strong research infrastructure and educated human talent can the nation fully capitalize on research.

Producing and harvesting the fruits of research must be done by talented scientists and engineers who can create and transform new knowledge into uses that are aligned with national goals. The working group found that the human resource aspects of capitalization are much more important than is generally realized. To a very large extent, the most important long-term outcome of federal investments in research is to educate the next generation of scientists and engineers and to support the continuing education of those already in the workforce. ¹¹ Employers who understand the lifelong value of learning generally are willing to support on-the-job training, night courses, and even full-time courses leading to advanced degrees.¹²

Effective capitalization on science and technology requires a broad range of skills and talents. Outstanding researchers contribute by creating new knowledge. Entrepreneurs and corporate managers see new business opportunities and put research to work in order to pursue them. Policy makers and program managers in government are charged with linking research with national needs. Investors provide risk capital for new science- and technology-based enterprises.

Specialized science and engineering training is not necessary for all of these roles. Still, the contribution of scientists and engineers to capitalization is striking, in research and nonresearch roles. For example, visionary program management by DARPA and NSF scientists and engineers made key contributions to the launch and subsequent growth of the Internet (Hafner and Lyon, 1996). Life scientists who make discoveries with significant potential for application often play a role in capitalization as the founders or scientific advisers of new companies (Stephan and Everhart, 1998). Partners in venture capital firms often have science or engineering training as well as experience in business.

U.S. scientific and engineering human resources owe much of their strength to the practice of coupling advanced education with cutting-edge research, much of which is funded by the federal government.¹³ The realism, depth, and intellectual challenge of formulating and solving research problems constitute powerful means of preparing students for productive careers. The effectiveness of this system draws students from around the world, many of whom remain to teach, perform research, and contribute to capitalization in this country.

Most students gain their early research experiences through teachers and advisors. Typically, the advisor's research is supported by a public grant that includes provision to hire graduate students as research assistants. The research assistantship supports students financially and involves them in research. The advantage of the research assistantship is the intense involvement of the student in an authentic research experience under a faculty mentor.

Research assistantships have become the predominant form of federal support for graduate education in science and engineering over the past several decades (COSEPUP, 1995). Over the same period, direct support to students through fellowships and training grants has declined in importance. In a previous report, COSEPUP (1995) recommended efforts to diversify the mechanisms for student support.

There are good reasons for maintaining diverse support mechanisms for advanced science and engineering education. Although research assistantships give students valuable first-hand experience in research, exclusive reliance on this form of support by individual students can carry disadvantages. For example, the pressure on faculty to produce research results may be transferred to students. Students may become so involved in a specific research project that little time is left to gain a broad appreciation of their field or gain skills and experience needed for nonacademic career paths.

Likewise, science and engineering education can be enhanced by the introduction of other skills and experiences relevant to the capitalizing process. Diversified support in the form of traineeships and corporate internships can help students to gain these skills and experiences. For example, students may profit from exposure to nonacademic environments where they are likely to find employment, notably in industries and mission agencies. And they may benefit from learning skills that are useful in many positions, such as research planning, team building, management, product development, marketing, public policy, and business principles. Graduating students who possess such skills are powerful agents of technology transfer as they establish their own academic careers, join industry or the federal government, or start new companies.

Whereas the research environment is essential to graduate students, it is also important to undergraduate education. The undergraduate often is inspired by learning in an environment where new discoveries are occurring and where science and engineering are seen as vibrant, ever-changing professions. Many of the key roles in capitalization are played by those who receive undergraduate training in science or engineering, and then go into careers in business or other professions.

The ability of scientists and engineers to move out of the laboratory to manage and even found their own businesses is a powerful feature of the U.S. capitalizing environment present in few other countries. In this way, scientists and engineers can be effective agents of technology transfer to address real-world problems, and add greatly to the entrepreneurial strength of the United States (Roberts, 1991).

One prominent example of a scientist-engineer-entrepreneur is Alejandro Zaffaroni. Zaffaroni was born in Uruguay, where he studied medicine. After receiving a doctorate in biochemistry from the University of Rochester, New York, he went to work for Syntex, then a small firm headquartered in Mexico City. Zaffaroni's work with Dr. Carl Djerassi became the basis of the Pill, or oral contraceptive, a product that transformed Syntex into a large pharmaceutical company.

Zaffaroni went on to found several other firms, including ALZA Corporation (a leader in drug delivery technology), DNAX Ltd. (manufacturing macromolecular products for medicine), Affymax (exploiting a technology that permits the parallel synthesis and screening of compounds for pharmacological activity), and Affymetrix (managing genetic information). As a scientist/entrepreneur, he continues to work in the fields of drug delivery and drug discovery. Firms founded by Zaffaroni have been fertile training grounds for other entrepreneurs.

Certain academic communities, where many entrepreneurs with science and engineering backgrounds are found, have a large impact on the transfer of ideas to the marketplace. The Bank of Boston has measured the economic contribution of companies founded by Massachusetts Institute of Technology (MIT) graduates and faculty and found that 4000 MIT-related companies have annual world sales of $232 billion (BankBoston, 1997).

Many fast-growing, emerging fields, such as telecommunications and biotechnology, are broad and multidisciplinary. Capitalizing on research in these areas may require expertise in several disciplines that span basic, applied, and developmental activities. One way to achieve this expertise is through partnerships between public, private, and educational institutions. The working group's examination of partnerships shows that they can be valuable, but cannot ensure effective capitalization by themselves. For example, industry-university partnerships cannot be expected to serve as a substitute for the federal government as funder of most basic research and the educational component of research.

A valuable agent of capitalizing is the "permeable membrane" between universities and industry which allows a relatively free flow of people and ideas. Government often facilitates this flow in its role as a customer for research. There are two principal benefits of this flow:

1. Industry gains access to trained people, new ideas, and new processes;
2. Universities gain financial support for research and education, exciting real-world problems, intellectual feedback, consulting opportunities for faculty, and internships and employment possibilities for students.

Over the past several decades, U.S. industry has expanded its support for university research.¹⁴ Much of this support takes the form of sponsored research, technology licensing, graduate fellowships, consortia, or faculty consulting, without being institutionalized in special programs or centers. The value to industry of federally funded university research is even greater than previously suspected. A survey by Carnegie-Mellon University concluded: "The conventional view holds that the short-term impact of university research on industrial R&D is negligible except in a few industries. Accumulating evidence suggests that we revise this perception.... university research provides critical short-term payoffs in some industries (such as pharmaceuticals) and is broadly important in numerous industries" (Cohen et al., 1994). One survey of companies and academic researchers showed that academic research has made significant contributions to a range of products, and that industry and government have played complementary roles in funding this work (Mansfield, 1995). At the same time, industry-university collaboration is not always smooth, and it is important not to oversell the direct and short-term payoffs.¹⁵

One example of a long-standing university-industry partnership examined during the course of the study is the Semiconductor Research Corporation. SRC was created in 1982 when the semiconductor industry saw the need to prepare more students for careers in industry by funding silicon-related research at universities. Today, SRC's 13 full members and other participants invest $35 million a year in university research.¹⁶ In 1997, SRC provided support to 700 students at 44 universities. According to Gordon Moore, cofounder of Intel, the "consortium's funds have been successful in keeping several major universities engaged in research that is immediately germane to the integrated circuit industry" and "the industry has probably leveraged more than two to three times the money it has invested in the SRC, some $200 million over a ten-year period" (Moore, 1996).

The federal government has played an important role in fostering intersectoral cooperation and partnerships in research. One such partnership is SEMATECH, a consortium of private firms set up with government support to strengthen the domestic chip-manufacturing industry. Since SEMATECH was founded in 1987, U.S. firms are once again world-class competitors. Although the turnaround cannot be attributed only to SEMATECH, it did succeed in showing that competitors can collaborate to their mutual benefit. After a decade of operation and $800 million in federal funding, it now plans to continue without further federal support (Roos et al., 1998).

Another government-sponsored partnership, the Engineering Research Centers (ERC) program, was established in 1985 by the NSF at least partly in response to the concern that U.S. industrial competitiveness was declining. The program has sought to increase interactions between universities and industry, including the pursuit of interdisciplinary research, and to offer students a broad understanding of how products move from laboratory to market. According to one study (Ailes et al., 1998), the outcome of most value to participating firms is "knowledge exchange"—access to new ideas and people. Firms perceived ERC students as better prepared than their non-ERC-educated counterparts.

An example of a successful ERC is the Data Storage Systems Center (DSSC). In the early 1980s, there was a dearth of university research relevant to the disc drive industry. In response, Mark Kryder, a professor at Carnegie-Mellon University, gained industry support for a collaborative research center. The DSSC was designated an ERC in 1990 and it has continued to contribute to maintaining U.S. capabilities. As a result of this program, there has been a significant increase in Ph.D.s graduated in this field, and students now are better prepared to make an immediate contribution to industry upon graduation (McKendrick, 1997).

An overseas model for government-industry collaborations is the Fraunhofer Society, formed in Germany in 1949, which now runs 46 institutes (including 5 centers in the United States) to do applied research for industrial clients, primarily

in mechanical engineering and microelectronics. Each institute, subsidized by government and organized internally as a profit center, makes its results available to both industry and the public. An important element of technology transfer is patent policy. The institute tends to register the patent itself, awarding an exclusive license to the industrial partner only for a particular application. The institute then may license the technology to another company for a different application (Abramson et al., 1997).

In recent years, the U.S. federal government has participated in a number of collaborative research programs with industry and/or universities. Two major government-industry partnerships are based on private-sector initiative and investment. The Intelligent Transportation Systems Program, established in 1991 by the Department of Transportation, seeks to enhance the capacity of the surface transportation system while reducing its social costs. The National Information Infrastructure project is the product of the deregulation of the telecommunications industry and the success of the Internet. The federal government sees its primary role as catalyst and consensus seeker.

The Advanced Technology Program (ATP) of NIST in the Department of Commerce seeks to strengthen the civil technology base and to support precompetitive technologies on a cost-shared basis. Its goal is to create industrial technology that is deemed too risky for firms on their own, but which has the potential to benefit not only the firm but also the nation if developed. The technological value of the program has been difficult to assess, although it functions well administratively. ATP remains politically controversial (Hill, 1998). Some analysts argue that programs like ATP and the Small Business Innovation Research (SBIR) program may displace industry funding by supporting projects that industry would have undertaken without federal help (Wallsten, 1997). Others believe that ATP, SBIR, and other programs that fund companies can play a positive role if the possible economic impacts of proposed work are considered carefully, and the results evaluated (Jaffe, 1996).

Contrary to common assumption, capitalizing on research is neither costless nor automatic. For better or worse, if there is no incentive to transform an idea into a useful process or product, it will not happen. Sometimes the incentive is apparent market value, but often that value must be mined and extracted by hard work and adequate funding. Box 2-2 lists some of the policy changes that have affected the environment for capitalizing. Although assessing the impacts of these changes comprehensively was beyond the scope of the study, a number of them have had effects on capitalization in particular fields examined by the working group.

Capitalizing usually works best when there is wide diffusion of information, a consistent and supportive regulatory environment, and easy movement of people between institutions. These features are part of a web of institutions, regulations, markets, and laws that enable the ownership of rights, the development of technology,

and the ability to secure adequate returns on investments. Capitalization also is fueled by a pool of private capital and a culture of entrepreneurs who have economic freedom, confidence in the economic environment, and financial incentives to take the risks associated with innovation.

An important economic feature of the capitalizing environment is the regulatory and trade environment, which allows entry by new firms and high levels of interfirm competition. These policies reduce the market power of dominant firms—especially in the semiconductor, computer hardware, and computer software industries—and support diffusion of intellectual property to a degree not seen in Europe or Japan (Mowery, 1996).

In the development of the Internet, biotechnology, and other rapidly growing areas of capitalization, the interchange of people between universities and industry has been especially important. An environment allowing the free movement and communication of key individuals has allowed for the creation of networks of individuals, an "invisible college" of expertise.

The community of ambitious investors, entrepreneurs, scientists, and engineers who form these networks is viewed as a national asset (COSEPUP and STEP, 1999a). The "Silicon Valley culture" that tolerates risk and even failure is vital in allowing innovative ideas to flourish. An open culture facilitates access to ideas, people, and capital, even among competing firms. In the broader business culture, the ability to revitalize companies that have stagnated or have been challenged by global competition helps to maintain productivity and agility.

These forces come together in what Jane Fountain (1998) calls a "high-performing industry network," of which Silicon Valley is the paradigm. Firms are characterized by delayering of the chain of command, cross-functional teams, fluid division of labor, flexibility, high capacity to absorb innovation, and organization by business unit rather than function. Such a network typically has an outstanding nucleus of research and education at its center. One expert suggests that, in 1994, 100 Stanford-related companies accounted for about $53 billion out of a total of approximately $85 billion (over 60 percent) of the revenues of Silicon Valley companies (Gibbons, 1997).

A particular strength of the U.S. capitalizing environment is the availability of private financing. Several components of this system are (COSEPUP and STEP, 1999a):

• Venture capital institutions: The most prominent is the professional venture capital firm, which typically functions in partnership with entrepreneurs. The people who staff these firms are professional managers who invest funds from corporate and public pension funds (43 percent), endow-

• ments and foundations (21 percent), corporations (19 percent), and other investors.¹⁷
• Entrepreneurs: Central to the culture of innovation are those who create and drive new businesses. Venture capital funds and other investors look for particular qualities in these people, including knowledge of markets, intuition, interpersonal skills, willingness to take risks, independence, a strong desire to own their own business, and an ability to learn and bounce back from failure. The role of scientists and engineers as entrepreneurs was covered earlier.
• Angels: These individual investors are often entrepreneurs who have achieved financial success in a prior venture. According to one estimate, of the roughly two million self-made, high-net-worth individuals in the United States, approximately 250,000 are angels (Sohl, 1997 a,b). They invest $10 billion to $20 billion each year in over 30,000 ventures (vs. $3 billion to $4 billion invested annually by about 500 professional venture capital funds in about 3,000 companies).

Competition, the flow of knowledge, and the protection and management of intellectual property are important and complementary elements of a favorable environment for capitalization, as illustrated by several historical examples.

Many innovations in microelectronics originated at AT&T Bell Laboratories. Because AT&T operated as a monopoly under a consent decree until the early 1980s, it was forced to license its inventions on reasonable terms and was barred from competing as a merchant semiconductor producer. This resulted in a flow of knowledge and expertise that seeded the growth of the U.S. semiconductor industry. Balancing the need to ensure competitive markets with the danger of constraining the innovative power of strong companies is a continuing challenge for the United States and other countries.

Another advantage in the U.S. capitalizing climate is a system of patent laws that gives incentives to innovators while promoting the diffusion of knowledge. Intellectual property protection is critically important in all high-technology industries, although the role of patents varies widely by field. They often play a direct role in competition in the pharmaceutical, biotechnology, and chemical industries. In other industries where technology changes rapidly, such as semiconductors and computers, firms may be in a position to profit handsomely simply by exploiting a head start. In these industries, intellectual property protection is often important as a means for firms to establish and sustain their technological foundation or as a lever to prevent piracy.

One steady trend of the past several decades is the globalization of science and engineering research activities and capability to capitalize. This study examined the international aspects of capitalization, with the goal of generating insights on three issues. First, how effective is the United States at capitalizing on science and engineering research compared with other countries, and does it matter? Second, what can we learn about capitalization from cases in which one country leads in research but where there is a failure, delay, or geographic shift in capitalization? Third, what is the benefit of supporting expensive cutting-edge research in a world where knowledge, investment, and other assets are increasingly free to move across borders?

One issue that has been debated over the past 10 years is whether the United States is effective at capitalizing on research investments compared with other countries. Some have argued that firms based in Japan and elsewhere have been nimbler at commercializing the results of U.S. research than U.S.-based firms. This is a complex issue that the working group explored in some depth (see Appendix A for examples).

Assessing the benefits and costs to the United States when firms in other countries commercialize U.S. research is a complex exercise. In the case of significant new products such as the VCR and flat panel displays, U.S. industry and workers have not gained the income and other benefits that would have resulted from U.S. production. However, U.S. consumers have benefited from the availability of superior products. As discussed later, U.S.-based firms have capitalized on foreign research in a number of cases, demonstrating the broad value of free information flow.

Still, a consistent pattern in which U.S. researchers make valuable new discoveries, and then foreign firms jump ahead to develop and commercialize them, would reveal a need for new approaches. Fortunately, the examples examined by the working group show that this is not generally the case today.

In several important product categories, Japanese firms have taken technologies that had been demonstrated or commercialized by non-Japanese companies and created profitable new markets. These include the oxygen steelmaking process (developed in Austria), the numerically controlled machine tool (developed at MIT and first commercialized by U.S. companies), and liquid crystal displays (prototyped by RCA; see Appendix A). The working group did examine one case in which U.S. research was turned directly into products overseas: the field of fuzzy logic, which emerged at the University of California at Berkeley and was applied in Europe and Japan (see Appendix A). On the other side, U.S. companies also have capitalized on research breakthroughs and even products developed elsewhere, such as MRI, monoclonal antibodies, and the jet engine.

Although the United States has been very effective in taking research to the

first demonstration or product, in the 1970s and 1980s some U.S. companies and industries clearly faltered in later stages of the process in some fields—notably in semiconductors, automobiles, and consumer electronics, where Japanese companies forged ahead through superior product development, manufacturing, and marketing. Yet it is important to remember that capitalizing on research is a dynamic process. A technique that gains market leadership in one decade may not be effective in the next. U.S. industries were surprised by Japanese market successes in the 1970s, but they learned important lessons and since then they have become far more efficient. This year, when the United States ranks first in the world in competitiveness, Japan has fallen to eighteenth position (IMD, 1998). Of course, this situation could reverse itself again in the future.¹⁸

Since the 1980s, many U.S. companies have devised ways to capitalize on research by bringing their products to the marketplace more swiftly. Two examples of corporate retooling—Motorola's revised manufacturing process and Chrysler's new system of product development—illustrate how the application of a new strategy can pay off.

In the mid 1980s, Motorola faced severe competition in the cellular phone business. A challenge by the company's cellular chief, Ed Staiano, to rethink the entire manufacturing process, led to a new "dedication to quality, product leadership, global reach, strategic and organizational flexibility, and a management style that encouraged initiative at every level...." (Lester, 1998). By sharpening its marketing focus and improving the quality of its manufacturing process, Motorola became the second-largest manufacturer of cellular infrastructure equipment and for a number of years was the world's leading producer of cellular telephones. Like all international companies, Motorola continues to face challenges in global competition in the cellular phone market and in other areas.

The Chrysler Corporation benefited from increased attention to product development, manufacturing and marketing. In the early 1990s, profits and market share languished and Chrysler had no competitive entries in the small-car market, which was dominated by Japanese companies, and in other important segments. Chrysler increased the introduction of new models and changed its process of product development to achieve quicker timelines and greater efficiency. The introduction of autonomous "platform teams," consisting of all the people needed to design and produce a new car (manufacturing, purchasing, marketing professionals, hourly manufacturing workers) improved teamwork, saved time, and reduced last-minute changes (Lester, 1998). The new cars proved popular, and Chrysler has achieved impressive gains in market share and financial performance since 1992.¹⁹

Despite the overall positive performance of the U.S. economy during most of the 1990s, it is clear that the United States will continue to face economic challenges [see NRC, 1999]. As discussed earlier, science and technology is a dynamic arena, and capitalizing on science and technology advances requires constant adaptation to change.

The working group became aware of several examples of failure and delay in capitalization, as well as cases in which breakthroughs made by one country were capitalized upon by another. These examples provide insights into the capitalization process and the elements necessary to take advantage of research. Among the examples studied, there were several general reasons for failure, delay, or geographic shifts.²⁰

• Entrenched existing technologies: Excellent research cannot be capitalized upon in some applications because an existing technology already performs the same function. The new approach cannot overcome barriers of cost and investments in infrastructure that support the existing technology. This applies to several applications of optical sensing, and probably to U.S. slowness to utilize fuzzy logic.
• Need for complementary advances: Capitalization may be delayed because of a need for complementary advances across a number of fields. Speech recognition is a good example. In the 1980s, after several decades of long-term industry and government research, hardware and software advances combined to produce an environment in which applications could develop and expand.
• Lead users of research are absent or located in other countries: Examples in which a product developed in one country is capitalized upon by another include NC machine tools and flat panel displays. This appears to occur when an industry concentrated in the capitalizing country has an appropriate infrastructure and a pressing business need that can be advanced by applying the research. Similarly, capitalizing may be hindered by weak links between researchers and lead users. This appears to hinder the application of research on cognition and learning in education.
• Lack of flexible human resources and a weak environment for launching science- and technology-based companies: In several instances, the United States has been able to capitalize on research break-throughs achieved elsewhere, particularly in the biotechnology and biomedical areas. In fields such as monoclonal antibodies, where new companies

• played a significant role, the presence of a favorable environment for starting new companies and a pool of trained and mobile people have been important contributing factors.
• Weak cooperation between sectors: In several industries in which Japan has enjoyed success in the past, such as oxygen steelmaking and semiconductors, cooperation among companies and between government and industry appears to have played an important role. Likewise, examples of U.S. industry resurgence in semiconductors, data storage, and other areas have been associated with significant efforts to forge greater interfirm and intersectoral cooperation, such as SEMATECH, SRC, and DSSC.

One of the key insights underlying the economic study of scientific and technological advances is that private firms tend to invest less in research than is optimal for society as a whole. This is because research activities create large spillover benefits that the investing firm cannot appropriate completely (Nelson, 1959). The tendency for R&D activities to produce spillovers outside the performing organization is a strong rationale for public support.

As discussed further in Chapter 3, firms and governments alike face pressure to focus their science and technology investments in areas most likely to produce clear benefits in the short-term. At the same time, as technological capability becomes globalized, multinational firms are better able to capitalize on important developments wherever they occur. Does this imply that the fundamental science and engineering research that is least appropriable by individual firms is an international "free good"? Does the key to success for firms and nations lie in exploiting the world's basic research while performing as little as possible themselves? Will national governments underinvest in fundamental research in the future?

Although these are complex questions, the literature on the economics of innovation and the working group's observations suggest that support for fundamental research is a very worthwhile national investment.

As noted earlier, capitalization is not a costless activity. Nathan Rosenberg (1990) has observed that firms are often unable to capitalize on external basic research advances unless they are performing basic research themselves. Judging from examples examined by the working group, such as catalysis and monoclonal antibodies, this appears to hold true at the national level as well. It appears that organizations and countries must make a significant contribution to the world's stock of scientific and technological knowledge if they hope to take advantage of cutting-edge developments themselves.

Also, there is evidence that science and technology activities are not as globalized as some believe (Pavitt, 1991; Callan et al., 1997) Although the subject deserves additional examination, it appears that most national technological systems are still relatively self-contained. Further, firms that do a large percentage of their R&D outside their home country, as do firms based in Holland and Switzerland, may be best positioned to take advantage of increased globalization.

Finally, it is apparent that investments in research help to create many of the assets that are essential to capitalization. Perhaps foremost among these is human capital, in the form of educated scientists and engineers who perform cutting-edge research and also play important postresearch roles in capitalization. Sustaining U.S. research capabilities, including the physical infrastructure for research and advanced education, at the same time strengthens the stock of human capital. This is a national asset that is far less mobile than financial capital or disembodied knowledge transmitted by research papers or the Internet.

In short, it appears to be more difficult for countries to "ride free" on basic research than is sometimes assumed. Recently, Japan, Korea, and other countries that have enjoyed success in capitalizing on foreign technology in the past have moved to establish basic research institutes and to strengthen advanced science and engineering education. This may reflect a realization that efficiencies in manufacturing and marketing are not in themselves sufficient for effective capitalization at the leading edge of science and technology [see Department of Commerce (1997)]. Exploitation of proven foreign technologies has allowed countries in Asia and elsewhere to develop rapidly. As incomes and wealth grow, however, it appears that more advanced capabilities are required for countries to continue catching up or even stay in place. Just as U.S. companies and institutions learned from Japan's superior manufacturing practices, other countries are adapting U.S. best practices in capitalizing on science and technology to their own circumstances (Mathews, 1997).

The United States enjoys success in capitalizing on investments in science and technology because of its investments in research, its favorable environment for capitalizing, its outstanding human resources, and its ability to forge cooperation among the industry, university, and government sectors. None of these factors is sufficient in itself, and it appears that all four are increasingly complementary and mutually reinforcing. Chapter 3 describes some of the challenges that the United States faces in maintaining and improving its ability to capitalize in the future.