Future R&D Environments: A Report for the National Institute of Standards and Technology (2002)

Scientific research and technology development do not develop in a vacuum. They are shaped by contextual factors, such as institutional and funding patterns and social values. Although these factors are discussed generally in this section of the report, they would affect the three fields of focus in this report somewhat differently. Each field has a distinctive organizational pattern, a different emphasis on science and technology development, and a different mode of operation. The biomedical field, for example, has a substantial university component and is driven the most by basic research. The computer and information technology field, on the other hand, consists mostly of small firms and is largely technology driven. Materials occupies the middle ground. The work is more likely to be carried on by large and established firms, and there is more of a balance between basic research and hands-on technological effort.

The R&D enterprise in the United States is carried out as a public-private partnership, with roles for each sector in funding the effort and in carrying out the actual work. The system has evolved since the end of World War II, and not only does the committee expect it to continue to do so, but there are indications that the rate of change will accelerate as a result of significant changes in the environment for R&D. Federal funding of curiosity-driven research is under pressure, changes in corporate structure and the emergence of new industries are giving rise to changes in the organization of industrial research, patterns of private investment in R&D are changing, the globalization of R&D is giving rise to new competitive pressures, universities are becoming more entrepreneurial, and the career choices

of undergraduate and graduate students are changing (there is less interest in science and technology careers, and among those who do go into science or engineering, there is more interest in jobs outside academia). In this chapter of the report, the committee addresses what it believes will be among the most salient elements in shaping the environment over the next several years for science and technology in the United States.

A useful starting point is a brief summary of what the system looks like today and how it evolved. The R&D enterprise in the United States had grown to an estimated $265 billion a year in 2000 from a starting level of less than $5 billion a year in the years shortly after World War II.¹ Until 1980, federal funding exceeded industrial funding, but by 2000, industry accounted for 68 percent of R&D expenditures and carried out 75 percent of the work, measured in dollars spent. Most of those expenditures, however, could be categorized as development; basic research comprises only 18 percent of the total spent nationally on R&D and the patterns of funding for and performance of basic research were and continue to be entirely different from those of development.²

Federal support of R&D is provided by a number of agencies. Much of the funding, especially for applied research and development, is provided by the mission agencies that are users of the results, for example, the Department of Defense and the National Aeronautics and Space Administration. Some, including the National Science Foundation and the National Institutes of Health, conduct or support research for its own sake, including basic long-range research, although funding is generally predicated on the historical contributions of research to national well-being. Some R&D agencies exist to support the private sector, such as the Cooperative State Research, Education, and Extension Service in the Department of Agriculture and NIST in the Department of Commerce. NIST not only supports technology development, it is also responsible for metrology, the science of weights and measures, which underlies the development of technical standards relied on by industry.

In 2000, the federal government provided 49 percent of the funding for basic research, industry contributed 34 percent, and another 18 percent was provided by the universities, other nonprofits, and nonfederal governments. Although the federal government was the largest funder of basic research, it was not the largest performer. Universities and university-based federally funded R&D centers (FFRDCs) carried out 49 percent of all basic research in the United States in

2000, industry and industry FFRDCs 34 percent, nonprofit research institutions and nonprofit FFRDCs 10 percent, and federal laboratories about 7 percent.³

Basic research activity is far from uniform across the more than 3000 post-secondary institutions in the United States. Most university basic research is carried out at just 60 universities, both public and private.⁴ Industry provides 8 percent of funding for academic basic research, a level that changed very little in the last decade. In contrast, the federal government provided about 58 percent of the funding for academic basic research during the last 10 years (the rest is funded by the universities themselves and other nonprofits). Thus, research universities continue to rely very heavily on the federal government.

There have been clear trends in recent years in the research fields supported by the federal government.⁵ In the mid-1990s, expenditures on nondefense R&D moved ahead of defense expenditures for the first time. That trend has continued, at least until now. Spending on health dominates, accounting for almost half of the federal nondefense research budget. Health research has also shown by far the fastest growth, compared with other fields of research, which as a group have had nearly flat appropriations in recent years, as measured in constant dollars. Nonhealth research increased by less than 0.6 percent from 1993 to 1999, while health research increased by 28 percent, in constant dollars. Federal funding of math and computer sciences increased by 45 percent, but funding of the physical sciences was 18 percent less in 1999 than in 1993. Several fields of engineering also received less. A recent study of these trends by the Board on Science, Technology, and Economic Policy (STEP) of the National Research Council (from which the data in this paragraph were taken) concluded that there is cause for concern about reduced federal funding of most of the physical sciences and engineering, because it is being driven by changes in the mission of individual departments, especially the Department of Defense, not from a systematic analysis of national needs and priorities. STEP recommended that the White House Office of Science and Technology and the Office of Management and Budget lead an

evaluation of the federal research portfolio to ensure that research related to industrial performance and other national priorities is adequately supported.⁶

Finally, it should be noted that the past couple of decades have seen a sea change in industrial laboratories, which, as noted below, appears likely to continue. Many of the great industrial central laboratories—Bell Labs, GE, Exxon, to name just a few—have been significantly downsized, or have reoriented their research to shorter-term goals, or both. There have also been changes in university laboratories, or at least in the strategies of universities, following the passage of the Bayh-Dole Act in 1980, which allowed universities to acquire intellectual property rights on products and processes even if they were developed with government funding.

Given the present state of affairs and the trends evident at the moment, the question is what the effect will be on research and development activities in the next 10 years. The following sections address that question.

To comprehend the likely changes in the organization of research in the next decade, it is necessary to sort through the trends in the levels of public and private support for R&D as well as the kinds of R&D that public and private entities are likely to provide. As corporate research centers continue to decline because of mergers or pressure to maximize short-term profits, even more basic research will have to shift to universities. Since there is nothing in recent trends to indicate that industry will increase its support for university research to a substantial amount (despite a few well-publicized individual university-industry linkages), it is likely that the government will remain the principal source of funds.⁷

There are some serious challenges and uncertainties in this respect. Although federal funding for research has been growing, the growth has been far from uniform across the various fields of science. Biomedical science support has grown substantially, which will probably continue, but other areas have not fared as well. Since the thesis of this committee is that exploiting synergies between fields is especially important in achieving scientific progress, success in the next decade is likely to depend on either redressing the imbalances in funding or changing the approach to research support within specific agencies. In regard to the latter option, the National Institutes of Health (NIH) is particularly important. As was pointed out in Chapter 2, there are enormous opportunities to advance the

biomedical sciences, but both materials science and information technology are key to many of those advances. Although NIH is clearly the main source of federal support for research in the biomedical sciences (88 percent in 1999),⁸ the organization of its institutes around specific diseases and its general expectation that research questions will be structured around specific hypotheses about the cure of those diseases make it difficult to pursue the kind of multidisciplinary research described earlier, which tends to be structured around questions of physicochemical mechanisms, new approaches to virtual experimentation, or device and instrument development. The institute structure also fragments research on cross-cutting problems such as medical error, nutrition, and risky behavior.

A second source of uncertainty about how the federal research support system will function in the next decade is the sporadic but continuing debate in Congress over the value of establishing explicit goals for research or expected outcomes from it. For the past several years, the idea that the best guarantee of progress in science and technology is to support curiosity-driven research has been questioned by a number of people in Congress, including some of the strongest supporters of science. The argument is that the high cost of research requires that investments be guided by clear goals, which may range from pursuing areas deemed ripe for progress to identifying social needs and encouraging focused efforts to meet those needs, and progress toward those goals should be assessed on a regular basis.

In fact, much federally supported research has always been in specific areas of national need, notably national defense and health. Since the end of the Cold War, at least until September 11, there has been more emphasis on R&D to support national economic innovation, in which—superficially, at least—the notion that progress can be measured seems possible. It seems likely, therefore, that an increasing fraction of research, particularly when it is multidisciplinary, will at least be described in terms of and possibly subsumed in goal-oriented projects. At the same time, history shows that important advances often come from discovery- rather than goal-oriented research. Given the uncertainty of research, a healthy research portfolio will be diversified among mechanisms as well as among fields. In this regard, the National Nanoscale Initiative of the National Science Foundation (NSF) may be a harbinger of things to come. One benefit of a goal-oriented approach is that it tends to encourage multidisciplinary efforts, which have always faced difficulties in the academic world, but it will be a challenge to ensure that high-risk, long-term, or open-ended research efforts are not seriously impeded.

The interest of universities in creating intellectual property and the growing attractiveness of entrepreneurship with its financial rewards for both faculty and

graduate students are also likely to continue. Although these are not entirely antithetical to long-term basic research—for example, the development of new instruments or new synthesis methods often occurs in the context of an ongoing research program with more fundamental goals—there is likely to be some shift in university research from long-term basic research programs to programs focused on short-term results.

Although there are some downsides to this shift, it is also likely to bolster cooperative efforts between U.S. universities and industry, creating more opportunities for technology transfer and cross-fertilization, mirroring relationships that have heretofore been more prevalent in Europe and Japan. We can certainly expect to see a continuing increase in industrial licensing of university-developed intellectual property. Industry will look to universities for students trained in the skills sets they need, and it will have professors serve as consultants or actually work on research projects. More professors will take sabbaticals to work in industry, and an increasing proportion of graduate students will spend time as interns.

The intensified interactions will require rules and guidelines related to conflict of interest and ownership of intellectual property, but there are already signs that the growing relationships between U.S. universities and industry are helping all parties to sort through those issues and to develop acceptable approaches. For example, it is now commonplace for research universities to have formal and comprehensive policies on conflict of interest, along with standing committees to monitor compliance. Although it would be premature to say that all of these issues have been resolved, it appears to this committee that the learning curve will continue and, although concerns will not entirely disappear, relatively standardized patterns and protocols for these university-industry relationships will soon be in place.

We can expect that government agencies and programs will promote and support cooperative efforts involving a number of corporations and universities in high-risk, high-reward research with specific goals. These programs will often encourage or even require multidepartmental or multidisciplinary collaborations within a university. A typical example is the Multidisciplinary Research Program of the Department of Defense’s University Research Initiative. The program specifies areas of interest (for FY 2002, for example, it lists 19 research areas, including integrated nanosensors, membranes based on active molecular transport mechanisms, fuel cells, and the behavior of scaled-up information networks), requires the involvement of researchers from more than one discipline, encourages multiuniversity collaboration, and makes clear that successful applications will have to either involve joint university-industry research or provide in some other way for transferring technology to the private sector.

Research will not only be more multidisciplinary, it will be larger in scale. This has implications for the relative roles of academic and industrial research and the need to coordinate them. The types of large-scale analysis available

through gene chips, sequencing of the genome, and proteomics are more amenable to the high-throughput, technician-intensive work of industrial laboratories than to small research laboratories in universities. An example might be the identification of all the kinases, phosphatases, and other enzymes in cells.⁹ Industrial research, however, is proprietary and the results not always published, which raises the prospect of duplication of research efforts by government-funded investigators in universities. This increases the need for government and industry to coordinate the parts of big science projects.

We can also expect, in this next decade, to see significant changes in the organization of industrial research itself. The general trend to downsizing or eliminating central research laboratories and decentralizing research units to product divisions was noted above. As a result, industrial research facilities are likely to be more geographically dispersed and culturally diversified, reflecting the growing global reach of investments. The tendency to have laboratories in multiple locations is likely to continue, driven not only by the desire to capitalize on the diversity of skills and creative thinking in different parts of the world, but also by the need to respond to political pressures that tie market entry in a country to a research presence there as well as by market pressures to tailor technologies for individual countries.

Furthermore, as economic opportunities improve in developing countries (some of them are aggressively building up their educational and scientific research capacities), individual researchers will be less willing to emigrate from them. More companies will have research facilities overseas to capture this talent, which will result in even more cultural diversity within companies and a less U.S.-centric view of technology development or, put more positively, greater ability to develop technology systems and products for foreign markets. Communication infrastructure will facilitate more efficient operation of multiple sites and increase opportunities for employees to perform physical experimentation remotely, providing nearly universal access to expensive physical facilities. Research universities are likely to react to this globalization of research facilities and activities by establishing their own global centers of excellence, as the Massachusetts Institute of Technology did when it set up Media Lab facilities in India and Europe.

Companies are likely to become increasingly willing to outsource research or to pool resources with other companies to take on long-term and very large or risky projects to the extent permitted by antitrust regulations. The outsourcing may be to universities or to for-profit research organizations like Battelle, SAIC, or Sarnoff. Pooling of resources can take the form of participation in government- or industry-sponsored research consortia or research centers like Rockwell

Scientific and HRL Laboratories, which are jointly owned by several companies. These strategies can reduce cost, increase efficiency, and shorten time to market.

Companies will also resort, more and more, to creative ways to outsource product development. Acquisition of start-up companies is one way to outsource development and to leverage existing manufacturing and market infrastructure without directly affecting earnings. Venture capital investment in start-up companies, which fosters multiple sources of technology innovation by offering an economic incentive, is an effective way to shorten product development time. In turn, acquisition of start-up companies can improve quarterly profit-and-loss reporting by shifting R&D expenditures into good will. The development of company-owned technologies may also be outsourced in cooperation with venture capitalists, spinning off some technologies for further development by the satellite company or spinning in the developed technology at an appropriate point. The aim will be to match the risk profile of the project and the investment strategies of the company and the venture capitalists. To develop complex and large projects, industry will seek the capabilities of the national laboratories of the Department of Energy. The Intel-led extreme ultraviolet lithography project in cooperation with three national laboratories is an example.

Aside from these broad changes, the field of biotechnology is generating some specific changes of its own, driven by the highly rigorous regulatory environment in which those companies must operate and their deep dependence on strong intellectual property positions. Commercial R&D directed at exploiting genomics for new diagnostics and drug discovery took hold predominantly in smaller or new businesses rather than in large existing pharmaceutical companies. Business models based on near-term development and production of diagnostics were by and large not successful. This created a fertile field for partnerships, mergers, and acquisitions. The entrance of large companies, such as Motorola, Corning, and Agilent, was inevitable, both because of their ability to market and because they possessed cutting-edge technologies relevant to the next generation of chips (e.g., Samsung and chip manufacture and Corning and chip substrates). Currently, almost every kind of relationship exists: partnerships between small and large companies and also between large companies. Superimposed on this dynamic is a large resource drain in the form of a web of legal battles over patents, together with uncertainty surrounding the ultimate resolution of questions about what is patentable in this field.

Finally, industry associations are playing, and will continue to play, an increasingly important role in guiding the direction of R&D by developing industrywide roadmaps modeled after the successful example of the semiconductor industry. The Optoelectronics Industry Development Association is one such association. The Institute of Electrical and Electronic Engineers also continues to play a critical role in the development of industry standards, and even in the codification of de facto standards where technology is moving faster than standards bodies.

The restructuring of research in the next decade will have ramifications for the scientific and technical workforce and will be affected by a number of changes in that workforce. Education, motivation, and work habits are all likely to be noticeably different in the next decade from the past few decades. Broadly speaking, we will see growing opportunities for scientific or technical entrepreneurship and an increasing need for breadth in the technical workforce—multidisciplinary training within the technical fields and broader training in the management of the technical enterprise.

For example, technical entrepreneurship will continue to flourish as more technical people either form or join start-up companies. And because of the impact of technology on almost every business, there may be more CEOs with technical training. Their success stories would increase the social status of engineers, which would draw more students to technical fields. There is likely to be more interplay between business and engineering schools to encourage entrepreneurship. A student with a technical degree will probably have some familiarity with entrepreneurship, business, and management when he or she graduates and will want to join a company that can provide exposure to business practices rather than take up pure research in academia.

Furthermore, there will be greater pressure to make financial rewards for technical people commensurate with the economic value they create. The reward systems in government laboratories, universities, large corporations, and start-up companies operate under different constraints. To attract and retain star performers, they will need to find ways to provide sufficient compensation, which could make it difficult for organizations that have rigid pay scales to compete for top technical talent. The committee believes that the recent cataclysmic decline of some high-tech sectors is likely to be a temporary situation and that the private sector will continue to be a very strong competitor for talent because of its ability to offer significant financial rewards for scientists in hot fields. Government laboratories are caught in the middle of the bidding war for talent, able to offer neither the financial opportunity of the private sector nor the freedom of the academic environment.

Technical workforce mobility will continue to increase owing to the speed of technology obsolescence and economic cycles. Although individually disruptive, the net result of such cycles is positive because they lead to technology migration and cross-fertilization of know-how. This mobility is being recognized as a major competitive advantage for the United States and is likely to become more widely accepted.

The industrial system will also require new, or at least a greater range of, skills in its technical workforce. For example, shorter time to market is an important competitive advantage. To achieve it, industry will need technical people who are able to manage projects from idea conception to product launch in a

concurrent, coordinated fashion. Engineers will have to have managerial skills as well as a clear understanding of all aspects of product engineering, manufacturing, marketing, and intellectual property strategy. Continuing education will become a way of life for engineers to maintain versatility and remain competitive in the job market.

There is an increasing awareness of the shortage of workers with basic technical skills, especially in sectors with rapid business growth. The optics and photonics industry, for example, had and still has an acute shortage of mechanical and optics engineers and optics technicians. Companies have reacted by hiring mechanical and manufacturing engineers from the disk drive industry and trained their own technicians. This speaks to the need for an agile workforce that can be easily deployed in new areas.

The multidisciplinary nature of many of the new technical developments, particularly in the physical and biological sciences, will require that scientists and engineers learn to bridge the communications gap between different disciplines, each with its own vocabulary, jargon, and acronyms. Would-be biological scientists will see substantial shifts in the undergraduate curriculum, with more grounding in mathematics and physics than is now the case. Physical scientists and engineers will need to study biology, a trend we are already seeing, particularly in engineering curricula. For example, at the California Institute of Technology biology is a required course for all students.

Clearly, these new needs will put pressure on universities. While it is unlikely that the traditional departmental structure will change, past practice, particularly in the technical fields, of filling students’ programs with required courses in their major will have to give way to new practices, giving students greater latitude and encouraging them to select courses in other departments. This transition may be eased by the increasing need in research for interdepartmental cooperation to be successful in attracting the new kinds of grant funding from either government (the Department of Defense Multidisciplinary University Research Initiative and the NSF National Nanoscale Initiative are examples mentioned earlier) or industry.

Although the fraction of foreign students in U.S. graduate programs in science and engineering has declined during the past few years, non-U.S. citizens on temporary visas continue to make up a high percentage of graduate students and Ph.D.’s. Generally, this appears to be more beneficial than not for the United States. Without this abundant supply of talent, our research universities and technical industry would see dramatic decreases in their global competitiveness. Even those who return to their homeland can enhance our sphere of influence and remain a resource for the United States. Changes, therefore, in public policy on the immigration of foreign science and engineering students to the United States could significantly affect the human resources aspect of the national innovation system.

Several factors mentioned in preceding sections of this report are reflected in and reinforced by the patterns of investment we can expect to see in the next decade. The desire of large corporations to outsource technological development, coupled with the availability of venture capital and the increasing popularity of technical entrepreneurship, suggests that the proliferation of small start-up companies will continue. This has the significant advantage of creating many parallel points of technology development, which will speed up the technologyto-value-creation process. While there will be many fewer winners than losers, more wealth will be created than lost.

While the proliferation of start-up companies may deplete the technical talent available for basic research, successful technical entrepreneurs can have a major impact on our technology-driven economy. Besides turning technology into businesses, some successful technical entrepreneurs become serial entrepreneurs, starting other companies or investing their profits in other start-ups. Many retain their connection to academia, either returning to universities or making substantial gifts to them. This virtuous cycle is already presaged by the huge private foundations of Gates, Packard, and other early high-tech entrepreneurs similar to those of Carnegie, Rockefeller, and Ford before them.

Investments in long-term R&D by public companies will continue to be negatively affected by the need to achieve short-term financial goals. Companies have always had disincentives to invest in long-term basic research, especially the possibility that competitors could profit from any advances, but the pressure for companies to increase earnings on a quarterly basis has increased the pressure to invest in short-term applied research and development. Central research laboratories that perform basic research and develop infrastructure technology but do not produce short-term, direct financial benefits will probably continue to decline. This leaves gaps in three places: (1) long-term basic research, (2) capitalintensive applied R&D with a long time to return, and (3) critical infrastructure development.

It would be easy to suggest that these gaps must be filled by government funding; indeed, to a great extent that is likely to happen. However, there are complications that will need to be resolved, and the outcome is not clear at this point. First, the eagerness on the part of government, universities, and industry to promote the earliest possible application and commercialization of new science, coupled with what has been a steady natural compression of the elapsed time between fundamental investigations and marketable innovation, has made it difficult to distinguish the kinds of R&D projects that are the appropriate responsibility of government from the kinds that should be supported by the private sector. The growth of government-funded precompetitive consortia, of targeted or highly goal-oriented multiple-institution R&D programs, and of high-risk, high-payoff critical technologies suggests a widening rather than a narrowing of the definition

of what is appropriate for government funding. Assuring U.S. success in an era of global technological competition is clearly a factor in this trend.

On the other hand, there have been and will continue to be backlashes. Because these programs are often viewed as helping corporations, there have been very public and often politicized discussions about what constitutes an “American” corporation, making it eligible to participate in these programs, in a world in which multinational corporations are the dominant players. A danger is that excessive restrictions on who can participate in these public programs may actually narrow the group to the point where the approach loses its advantage and the investment is diminished in value.

A second possible backlash, potentially more serious, is that blending the goals of fundamental research with those of technological innovation could offset the tendency of scientists to self-organize for more international cooperation in basic research. If Congress and the public think that a competitive advantage in technology and its application is the same as a competitive advantage in the underlying basic sciences, they may come to believe it is better to compete than to cooperate in basic science. This would not only decrease the cost-effectiveness of global investments in basic research but would also present a major problem for “big science.” Projects in astronomy, astrophysics, and high-energy physics, for example, require more rather than less international cooperation to fund equipment and facilities, and pressures to close avenues of international cooperation in R&D might well be a problem in these areas. Indeed, one of the challenges in the next decade will be to develop new and more effective institutional approaches to long-term, multinational funding of research projects.

A third concern continues to affect the public reaction to government funding of programs that blur the distinction between basic research and application—namely, concern about “corporate welfare” or the appropriateness of government intervention in areas where the market should determine winners and losers. The perennial debate over the appropriateness and effectiveness of the Department of Commerce’s Advanced Technology Program illustrates the problem well.¹⁰

A fourth concern is that the increasing emphasis on technological innovation, with its focus on short-term goals, will drive out funding for basic research, which is long-term and not focused on practical goals. The perennial efforts to maintain the basic and applied research parts of the defense R&D budget (the socalled 6.1 and 6.2 accounts) distinct from its development parts illustrate this problem.

The second gap area mentioned above—capital-intensive applied R&D—

appears likely to start at least one trend that runs counter to the proliferation of small start-up companies competing to produce new technologies. Consolidations in the aircraft industry and defense contracting exemplify a countertrend driven by capital constraints. In the information technology area, we may see a similar phenomenon. The huge difference between the investment needed to be a player in the silicon arena versus that needed in the optical communications arena may one day become an important factor. Costs for a new silicon plant are now so large that it will be increasingly difficult for more than one player to be active, whereas the optical arena may stay highly competitive.

The third gap mentioned above is the need for critical infrastructure. Here the role of government is clear and likely to grow in importance. Advances in metrology and the development of technical standards are critical infrastructure capabilities that must be in place to facilitate the efficient incorporation of innovations into the economy, whether the innovations are in the biological or physical sciences. The need for such infrastructure is even greater when the innovations come from small companies, which do not have the resources large corporations have to fund large research facilities and sophisticated equipment.

The focus on short-term financial goals might be offset in part by new corporate approaches to analyzing R&D investments, such as real options methods. The justification for investment in R&D, including basic research, might be increased by new methods of allocating internal investment within firms that take more account of the role of R&D investment in corporate innovation.

There is every indication that the United States will continue to recognize entrepreneurship as a competitive advantage for the country and will continue to promote it. It is likely, moreover, that the federal government will continue to encourage investment through tax policy—for example, tax credits for R&D, R&D partnerships, and preferential capital gains tax.

Although the United States has generally been sensitive to business interests in the area of regulation, changes in industrial organization, the global character of the economy, and rapidly emerging technologies will present a number of new challenges for the regulatory system. We are already facing some major questions related to antitrust regulation, illustrated by the government’s prosecution of Microsoft for violation of antitrust statutes and the United States-European Union controversy over the proposed GE-Honeywell merger.

The Microsoft case arose because the usefulness (and, consequently, the economic value) of networked systems strongly depends on the number of connected nodes in the network (the usefulness is approximately proportional to the square of the number of nodes). When competition reduces the number of nodes, or users, of a particular system, it also decreases the value of the system. As a result, some economists now argue that the public interest would be best protected by a

sequence of temporary monopolies rather than by a forced sharing of the market at every moment. Therefore, there is likely to be pressure in the coming years to design antitrust regulations that prevent the long-term perpetuation of information network monopolies rather than regulations that prevent their formation altogether.

The GE-Honeywell attempted merger raised another issue—the difference in views in the United States and the European Union over whether mergers that allow horizontal integration across business or technology sectors violate antitrust principles by providing the merged entity with the ability to gain excessive market share within one or more of the individual sectors of which it is comprised (which was the position of the European Union). In view of the global trend to mergers and acquisitions, discussed in an earlier section, the ultimate resolution of this issue will have a significant effect on the strategy of multinationals in the next decade.

Still another important regulatory issue has been introduced by the growth of the medical technology industry, including biotechnology, tissue engineering, and medical devices. Although the Food and Drug Administration (FDA) has had responsibility for assuring the safety and effectiveness of medical devices since the mid-1970s, most observers would argue that the regulatory procedures and institutional organization have not evolved adequately from the days in which the FDA’s responsibility was almost entirely restricted to drug assessments. Since the development of medical devices follows an entirely different path from that of drugs, the medical devices industry needs very different assessment procedures, more room for design iteration, closer attention to systems issues, and different measures of performance.

The next decade is likely to further complicate the matter of regulation as the new technologies mature. For example, a hybrid pancreas that consists of a polymer matrix protecting and supporting cultured human beta cells that produce insulin for the control of sugar level is presently viewed by the FDA as a device, a biologic, and a drug, each of which is subject to regulation by a separate division of the agency under different protocols. If the United States is to be able to take advantage of the rapid, multidisciplinary innovation that is bound to occur during the next decade, there will be great pressure to develop a regulatory system that can evolve as quickly as the technologies it is charged with evaluating. One likely development is the evolution of the present system, in which medical devices are categorized as either “experimental” or “clinical,” into a system that recognizes a continuum of development from prototype to mature technology.

Medical devices are one area in which the nation’s regulatory procedures play an important role in determining its competitiveness. Medical technology companies have been moving a substantial fraction of their R&D to Europe because they view European device regulation as more rational and efficient and more conducive to technological innovation. On the other hand, more accommo-

dating systems for regulating genetically modified organisms has kept the United States in the lead in this area.

Intellectual property law, in both copyrights and patents, is another area in which technological advances are straining the system. Here, the choices made during the next decade will have a strong influence on U.S. competitiveness. For example, the new opportunities that computers and communication technologies offer for adding value to public data and information by reorganizing them and mining them are giving rise to an explosion of copyrighted databases. The government will face pressures to cut back on its efforts to refine raw data, exploiting the same technologies to make the data easier to use, and it may even be less aggressive or less accommodating in making raw data easily available, in order to avoid competing with the private sector. In the long run, this could work against the interests of small companies, which are less able to absorb the costs of accessing the information they need.

A related issue, whose resolution during the next decade will have an influence on scientific research, concerns the doctrine of fair use—the allowable, free copying of copyrighted information for a set of specifically designated purposes. In the past, fair use represented a compromise in which the owners of copyrighted information, recognizing the practical limitations on their ability to monitor limited copying and the relatively insignificant consequences of such copying, accommodated the needs of scholars and some others. With the advent of new information and communication technologies, monitoring has become practical and copying more problematical because of the ease with which enormous numbers of copies can be made and distributed. As a result, there has been great international pressure (through the World Intellectual Property Organization) to severely reduce fair use. The concern, expressed by a number of scientific and professional organizations in the United States particularly, is that this would limit the exchange and use of scientific data. More work is needed in the public policy arena to find a resolution to this issue that preserves the openness of communication in science, a vital aspect of scientific progress.

The tensions between the openness of science and the need to protect intellectual property in order to encourage innovation also reach into the area of patents, especially in biotechnology. The patenting of genes and even DNA sequences is clearly an unresolved issue, with the United States and the European Union diverging in their positions. The U.S. Patent Office has been fairly open to the patenting of genes, even where specific gene function has not yet been established. In fact, it has allowed patentees to make claims to future discoveries of gene function. Some have argued that this creates a useful incentive to companies to move aggressively to sequence genes, but to many the more likely effect is to create a disincentive to investigators to pursue other applications of a particular gene as the several proteins it expresses are discovered and understood.

Current U.S. patent and copyright laws may inhibit the development of new businesses, because disputes are usually resolved through the courts rather than

by the Patent Office. The cost of litigation is very high and therefore favors established companies. In some new technological areas, such as biotechnology, the lack of case law to help interpretation is especially problematical, because the uncertainty may inhibit investment in innovation. In others, patenting is less important than alternative forms of protection, such as trade secrecy.

As technological innovation shifts to small or medium-size companies, brand recognition becomes a less effective way of building consumer confidence and facilitating the adoption of new products and services. That will increase the role of the government as an honest broker, providing and disseminating independent assessments. The Internet is an important tool for disseminating assessments, but the vast array of information of varying quality that is available on the Internet makes it all the more difficult to establish credibility. The government may need to certify data on the Internet, especially where the information is widely used as the basis for research, product development, or manufacturing.

Even if the validation of data is judged to be an inappropriate role for the government, there is a need for formal data format standards that can address both longevity and migration issues. In addition, there is a need for standards in the presentation of genomics/proteomics sequences (in addition to the need for standards of sequence accuracy, noted earlier in this report) that is not currently being met by NIH, FDA, or NIST.

Finally, it should be noted that in the next decade, global standards will take on increasing importance for products in all areas. In much the same way as data and communication standards in telephony, fax machines, and the TCP/IP protocols of the Internet facilitate the development of global networks and the easy sharing of information, standards will be needed for many other products. Product standards can, it is true, serve as trade barriers, but for a nation committed to free trade on a global scale, they can also facilitate trade and help to gain competitive advantages. The failure of the United States to gain a significant share of the global market in cellular telephony is a good example of the price of not moving quickly. Today, product standards are voluntary in the United States but are sometimes import requirements in other countries. It may well be that as this field becomes more important, federal government will need to play a more direct role, especially as it is the critical voice in the World Trade Organization and other international venues where these issues are negotiated.