Books
Technology policy is as broad as it is deep. It spans diverse industries, organizations, sectors and processes, as well as a largely disconnected set of policy mechanisms. This project began as an effort to evaluate the Clinton-Gore technology initiatives, a bold, if imperfect, effort to advance economic growth through investment in innovation. In the process of this evaluation, we have constructed a set of six high-level principles for supporting the development of research and innovation for the public good— in support of both federal agency missions and larger national economic objectives. These principles, which are the focus of this chapter, are intended to provide a politically robust platform on which policy development can be based. They are designed to strengthen what has been a difficult effort to balance between maximizing economic objectives and not interfering with private markets. These principles underlie many of the Clinton administration technology programs and policies, but they were not always borne out in the politicized context of program authorization and administration. This concluding chapter seeks to articulate and explore these principles and the tools available to carry them out as a basis for policy development, management, and evaluation.
As described in Chapter 1, the context in which technology development occurs has shifted significantly over the past decade. Globalization, restructuring of industries, and the current political and fiscal climates all give rise to a need for a fundamental rethinking of the federal role in supporting research and fostering innovation. Government policy must look beyond supporting research to providing incentives for private innovation, using the full range of policy tools that we have seen arrayed in promoting information infrastructure (described in Chapter 13 by Brian Kahin). A new, more sophisticated, and more complex role for government is emerging, one that makes better use of resources, and that shares more decision making with states and with the private sector. New policies should focus on long-term investments in knowledge-based infrastructure— the capacity of the entire system of private entrepreneurship, human resources, investment, and advancing frontiers of technical knowledge— using tools that encourage and enable rather than direct the deployment of these assets. Government should look for opportunities to enhance the social capital of our society, to make the emerging networks of firms, universities, states and federal laboratories and agencies into a dynamic, trusted efficient system of creative enterprise (see especially Chapter 4 by Jane Fountain).
The goal of this concluding chapter is to move the debate over the federal role in advancing technological development beyond partisanship, and to focus on defining a basis for federal activity, for both mission needs and economic objectives, that is market-oriented and equitable. The chapter explores the questions: what are the appropriate areas for federal investment? What are the proper roles of the various stakeholders in technology policy— states, universities, national labs, federal agencies, and industry? What are the institutional mechanisms through which technology policy can be managed? These questions are addressed in the form of six principles for a research and innovation policy which we present to guide both a reexamination and restructuring of existing policies and programs and the design and development of new technology initiatives. Following a discussion of each of these principles, we describe how these principles can be applied to guide federal funding, and will then turn to the Advanced Technology Program to give a specific example of how such an enabling, research-driven approach to innovation would look.
(1) Six Principles
This section introduces six principles we offer to assist policy makers determine useful answers to questions related to program structure and purpose. This is followed by a detailed discussion of each principle.
Encourage Private Innovation
Government should leverage private investment in innovation to spur economic growth, improve living standards, and accomplish important government missions by creating incentives for and reducing barriers to technology development and research-based innovation.
Emphasize Basic Technology Research
Government direct investment in science and technology should focus on long-range, broadly useful research— in basic technology as well as basic science— that can produce benefits far in excess of what the private sector can capture for itself.
Facilitate Access to New and Old Technologies
Policy-makers should promote use and absorption of technology across the economic spectrum, with special attention to the role of higher education, the states, and networks of firms and other institutions.
Use All Policy Tools, Not Just R&D
Policy-makers should utilize the full range of relevant policy tools, such as economic policy, regulatory reform, standards, procurement, and intellectual property rights, in varying mixes, as appropriate for different industries, technologies, and regions.
Leverage Globalization Of Innovation
The U.S. government should encourage U.S.-led innovation abroad as well as at home, enabling U.S. firms to get maximum benefit from worldwide sources of technical knowledge, while keeping the United States the preferred location for investments in research and innovation.
Improve Government Effectiveness in Policy Development
Government should work to become a stable and reliable partner in a long-range national research effort, by creating more effective institutions for policy development, strong and stable nonpartisan support, and stronger participation by the states in policy formulation and execution.
(1) Encourage Private Innovation through Public-private Partnerships
Private firms perform three-quarters of all the R&D in America. They transform new scientific and engineering ideas into the products and services from which wealth is created, and are prime sources of innovation when the government is the primary consumer of the innovation. Technology policy should be structured to foster an economic climate that favors private investment in R&D and the effective and innovative use and absorption of technology by firms and organizations. With this in mind, federal interventions must encourage private investment, rather than substitute for it. They should favor the use of market mechanisms, such as tax incentives and creating markets for non-market entities (e.g., tradable permits for sulfur emissions), and not rely solely on direct government funding of R&D. However, substantial direct investments in research by government are also required, because private firms tend to underinvest in both long-range research of economic value and research to meet public purposes.
Many of the policies for encouraging private innovation are indirect. These policy tools are discussed below. But there is an important middle ground in policy, especially when the government has a job to do that calls for innovative solutions, such as promoting good health, protecting the national security, or cleaning up the environment. In the past, government has funded the total cost of research in the captive defense industry, while using regulation to impose the cost of more benign environmental technology on commercial industry. More recently policy makers have been looking for ways to leverage private investment to get the public''s needs fulfilled in both of these areas, and perhaps in others as well. The use of public-private partnerships (as in the Advanced Technology Program described in Chapter 6, the Technology Reinvestment Project described in Chapter 7, the Partnership for a New Generation of Vehicles described in Chapter 15, the Environmental Technology Initiative described in Chapter 11, and Cooperative Research and Development Agreement programs described in Chapter 9) has begun to fill that gap.
The Defense Department''s dual-use technology acquisition strategy (discussed in Chapter 7 by Linda Cohen) is a particular form of public-private partnership in which the DoD shares the cost of development of defense-relevant technology in private firms, thus leveraging the firm''s incentive to develop the technology for competitive markets. This strategy can be applied to the technology missions of other agencies as well. Instead of government paying one hundred percent of the cost of technologies it may need to fulfill its own operational needs, government should leverage private investment in industry sectors that most nearly satisfies those needs. This can be done through cost-shared investment in technology in anticipation of both public and private use. This strategy does raise difficult problems of accounting for the distribution of public and private costs and benefits. It requires careful management of cost-sharing policy. But if America is to continue to embrace smaller, more agile government and is to rely more on private innovation for addressing public issues, it will have to be able to manage the financial arrangements in public-private partnerships.
(2) Cost Sharing Arrangements
Cost-sharing should be a basic precept of federal technology initiatives where both public and private value is produced. The division of investment between public and private sources should reflect the relative public and private interests in the work. Basic science and technology research performed in universities and in independent laboratories, which rarely share in the economic benefits, should be fully funded by government in most cases. In contrast, public-private partnerships, especially those from which the private partner expects to derive a near-term commercial benefit, should require investment by both parties. Requirements for industry cost-sharing might be reduced in those cases where firms allow or even encourage the technology to be made widely available, for example through a consortium of participating institutions or by treating research results as non-proprietary. Industry cost shares will be greater when the investment area is closer to early commercialization.
In the public-private partnerships created by various federal agencies, these arrangements have been established ad hoc, with little uniformity as to cost-sharing policies. Overall federal guidelines for the establishment and management of public-private S&T partnerships are needed to ensure that cost-sharing is equitable, that technological development goals are given the best chance for success, and that intellectual property and diffusion issues are properly addressed. These guidelines should cover appropriate conditions for use of partnerships, criteria for degrees of cost sharing, policies for recapture of public investment, criteria for partner selection, criteria for allowing federal partnerships with single firms (rather than consortia), policies for disposition of intellectual property, policies for participation of partners with significant foreign ties, and policies obligating partners to cooperate with outcomes assessments after the partnership is dissolved. These policies should be designed so that they can be tracked by an entity within the federal government, such as the Office of Management and Budget, to ensure that abuses do not occur.
The funding ratios of federal technology programs should reflect the minimum public investment needed to entice private participants to develop technology that serves the public interest. The optimum ratio will shift from project to project and in some cases during the life cycle of a project. If, as we recommend in the next section, government shifts its emphasis to more basic technology, firms may be more willing to share the results and the funding ratio may reflect the increasing public stake in the outcome.
(1) Emphasize Basic Technology Research
Long recognized as an important and appropriate role for government, investments in basic scientific research have made the United States the preeminent scientific power in the world. However, the creation of new industries and new tools, materials, processes and systems thinking takes more than science. These capabilities, and indeed all technical progress, are enabled through research in "basic technologies," a concept explored by Lewis Branscomb in Chapter 5. Creative research on new kinds of materials, new processes and ways of exploring and measuring, new ways of doing and making things: this is the world of basic technological research. It includes activities such as the research behind the blue laser, the exploration of biosensors, and the process of polymer cross-linkage. Basic technological research also creates tools and data about properties of matter, and about materials and processes on which the progress of both science and of technology depend. It defines an area for federal investment that starts early in the innovation process and leads to knowledge that is often non-proprietary and widely diffusable, though it will clearly lead, eventually, to industrial applications. Basic science and basic technology are inextricably linked and dependent upon one another.
When the political debate divides the world into "R" and "D," putting basic scientific research on the one side and lumping applied research and development together on the other side, a huge and vitally important area of basic technology is left out of consideration. Many people assume that if the work is not basic science, it must be commercial development and that therefore government has no business investing in it. This assumption leaves out a large gray area we are calling "basic technology research," in which both government and industry have an interest.
Companies have become increasingly reluctant to put resources into basic technological research that is long term, high-risk or both, even though this research might eventually pay handsome returns to the firm and to society as a whole. Increasingly, as product development cycles shorten and competitive pressures rise, companies look outside for the innovations that basic research makes possible. As they "outsource" innovation to their supply chains and downsize their corporate research laboratories, the focus shifts to nearer time-horizons, perhaps increasing short-term profits, but at the expense of intellectual assets for future growth. If the United States is not the most fertile seed bed for such innovation, companies will— and do— look overseas.
As discussed in Chapter 5, and again at the end of this chapter, government should invest in research— both scientific and technological— where the expectations for long-term public benefit exceed expectations for private returns to the research performer. This is the correct answer to the question, "When is it appropriate for the government to subsidize research?" How such research investment is allocated between more theoretical and speculative (opportunity-driven) research and need-driven (but not privately appropriable) research will depend upon the missions of the research funding agencies and the decisions of the Congress and the president about the relative urgency of those needs. The discussion of these topics will be easier to understand if, instead of debating about government roles and investments in science versus those in technology, we discussed resources and goals for publicly funded research, versus a variety of incentives for promoting innovation.
(1) Facilitating Access to New and Old Technologies
The best science in the world is of little economic value if it cannot be accessed, understood, and put to use. Thus a cornerstone of any policy to promote innovation must be helping users acquire technical knowledge and skill. Firms need access to all available technology, not just the most recently created knowledge. They also need access to a workforce capable of utilizing available innovations. The development of a capable and competitive workforce is a key factor in any nation''s business climate and is a vital area of public concern. Many small firms have only limited access to the science that would help them choose wisely among technologies and to make effective use of them. Public investments in technology utilization, usually made at the state and local level, have proved to have high economic leverage. The federal government, working closely with states and regions, should develop closer links between technology policy and workforce training and development. This will help spur the diffusion and use of technology and create strong links between technology and the creation of high-wage, high-skill jobs.
A historical strength of U.S. science and technology policy has been the decentralization of initiative both among and within federal agencies, and— more importantly— to universities, government, and private laboratories and industry. Today, state governments are better prepared than in the past to play a growing role in this decentralization of initiative and of program management (as Christopher Coburn and Duncan Brown describe in Chapter 16). As choices, guided by national goals, are made within that decentralized decision process, we urge more attention to input from the intended beneficiaries of federal policy.
Firms in specific industry sectors tend to cluster regionally, generating an opportunity for alliances between state governments and industry consortia for collaboration with federal agencies in technology-based development activities. States vary greatly in their needs and approaches to technology-based development. Federal collaborations with states should have enough flexibility to respond to these local and regional differences. The U.S. Innovation Partnership, which links federal research and innovation policy-making to states through the National Governors'' Association, provides an important new mechanism for coordinating federal and state-level policies. Increasing recognition of the significance of regional economic specialization makes the idea of geographically defined programs of development rational for economic as well as political reasons.
The Manufacturing Extension Partnership (MEP) is an example of an effective federal-state partnership in technology diffusion and regional economic development (see Chapter 10 by Philip Shapira). Federal policy-makers should learn from the success of this model when developing or redesigning other federal technology efforts, particularly those conducted in collaboration with the states. Following the principles for cost-sharing outlined above, the federal-state funding ratios and the income from user charges should reflect the benefits each sector can expect to enjoy in pursuit of national and local public interests and the private interest of the assisted firms.
(2) Employing consortia to accelerate diffusion and public benefits
Cost-shared partnerships between government and industry require a delicate balance in allocating costs to the participants, and a similar balance must be struck in the flow of information. Where one side gains almost all the benefits, the balance is easily obtained: for example, when government funds the entire cost of basic science it expects the results to be published, while a firm that develops technology for specific commercial products will normally keep the technical knowledge to itself. Where both government and the firm will benefit, both the costs and the knowledge may be shared in proportion to their interest, investments, and willingness to share results with others. Since the government''s interest is strongly biased toward diffusion and utilization of the knowledge on a broad basis, government will expect the results from its partnership investment to produce broad public benefits. This can be accomplished through sharing the results with other firms, or at least commercializing them in ways that rapidly bring benefits to society and not just to the participating firm. Consortia can be used to share costs and to stimulate the absorption and use of science and technology in a manner that mitigates the problems of market distortion and fairness.
Congress encouraged the formation of R&D consortia through the Cooperative Research Act of 1984 and provided for their participation in the Advanced Technology Program in the 1988 Act. Consortia stimulate the diffusion of technology through the movement of people and the sharing of information among firms. They may also generate competitive pressures which tend to keep research less proprietary and more long-range in nature. Networks of firms and other research institutions will be more likely to diffuse research results rapidly and will generate less concern that federal R&D expenditures might disrupt markets or respond to political pressures. Government participation in such networks also reduces the danger of anti-competitive behavior within the consortium.
Some other countries, even many that are smaller than the United States, have much more experience with industrial consortia, usually in form of Industrial Associations that operate R&D organizations for the common benefit of their members. In some highly developed smaller countries like Switzerland, firms'' membership in these organizations is virtually compulsory. In Europe, the Framework Programs encourage trans-national consortia, including university participation in many cases. There are time-honored examples of horizontal consortia in the United States (American Gas Association, Electric Power Research Institute, Portland Cement Association, etc.), and more recent cases where there is significant government involvement (SEMATECH and PNGV/USCAR), but overall the U.S. experience with such organizational structures is limited.
Horizontal consortia connect firms who have similar needs for base technology but who compete in the end product market; vertical consortia connect firms in a chain of innovation. Horizontal consortia are useful for diffusing and sharing the costs of research leading to more efficient uses of new technology. Vertical consortia are useful for accelerating diffusion and can be a powerful mechanism for engaging the high-tech, first and second-tier, small and medium-sized firms. In both types, the sharing of information among firms helps to ensure that the government is not funding the development of products.
Apart from the practical advantages of engaging consortia in partnership programs, there is a political advantage of some importance. When the government contracts with single firms in projects like ATP, it is very difficult to reveal for purposes of evaluation the public benefits that may accrue. The use of an institutional arrangement, such as federal collaboration with a consortium that might include not only firms but state agencies and perhaps universities or national laboratories, itself provides the diffusion mechanism through which public values are realized.
By working within a network of firms and perhaps other institutions, the private sector can accelerate the diffusion of the innovations and assure the breadth of industry interest in the work, while the government focuses its investment on opening up new technical possibilities. There are situations in which such consortia can best be realized by relying on state authorities to assemble all of the relevant institutional commitments to maximize the likelihood of economic benefits where economic development is the objective. We discuss an example of this principle below where we discuss the ATP program.
(1) Research and Innovation Policy is More Than R&D
A "one size fits all" research and innovation policy is almost certain to be unsuccessful. Every industry is different and government agencies must be sensitive to these differences, which may call for different policy tools and different mixes of science, technology, and systems research. For example, strong patent protection is essential to business success in pharmaceuticals, but less so in the computer industry, where most large firms are cross-licensed internationally. Biotechnology firms draw directly on forefront basic science, while chemical and materials firms are more dependent on advanced process technology. Industry structures are also very different. In some, scale economies are vital (chemicals, energy, communications). In others, small and new firms are an essential source of vitality (software, materials, instruments). These realities require that government acquire much more effective means of informing itself about the conditions, needs, and opportunities of many sectors of service and manufacturing industry (as recommended in Chapter 15 by Daniel Roos, Frank Field and James Neely).
There are two broad types of policy tools: those involving the direct provision of federal funds for R&D and those that use indirect means, such as tax and economic policy, regulations, standards, procurement, and the like. These indirect tools may be used to provide incentives for private sector investment in research and development, and to enhance the accessibility and utility of research results, both new and old. The effective use of the full range of science and technology policy tools can stimulate and encourage private investment and make those investments more effective. A variety of policy instruments can also help ensure that technology strategies meet the distinct needs of different industries, regions and missions.
The administration has demonstrated the coordinated use of a broad set of tools in its management of the National Information Infrastructure (NII) (see Chapter 13 by Brian Kahin). The Internet originated in two decades of government-funded R&D activities at universities and other laboratories. As computer networking matured and other telecommunications services proliferated and converted to digital technology, private investment grew and the government relied increasingly on deregulation and other indirect policy tools for the further development of the NII. Today private investment is generating expectations of very large economic pay-off, and public policy for the National Information Infrastructure makes use of a wide array of direct and indirect policy mechanisms. The NII has been a unique departure from prior forms of physical infrastructure development, which were typically government-defined, government-led and, to a significant degree, government-financed. Here, the government role is cautiously defined and is comprehensive in scope. Almost every organ of government is in a position to foster (or impede) healthy NII development in the public interest.
The development of the Global Information Infrastructure is far from complete, and the administration has yet to define a mechanism to ensure that all of the federal policy tools— research, standards, trade policy, intellectual property, privacy, security, legal jurisdiction, and taxation— are properly coordinated. Despite this shortcoming, the Internet has done more for knowledge sharing and collaboration than any prior government effort. As such, it is identified by Jane Fountain (Chapter 4) as a promising element of social capital underlying U.S. innovative capacity. The Next Generation Internet Initiative can be a key element in advancing the technological edge and strengthening U.S. leadership in information infrastructure. The organizations in the president''s executive office (discussed in Chapter 17 by David Hart) will have to be strengthened if the needed policy development and coordination is to be achieved.
(2) Performance-based regulations and other regulatory alternatives
Experience has shown that prescriptive or coercive regulations are an expensive and inefficient tool for forcing private investment in technology. So too are regulations that erect barriers to competition, in the absence of compelling reasons for a regulated monopoly service. As has been demonstrated on a limited scale, output-based regulations, which specify the goal of the regulations rather than the process of achieving that goal, can produce results consistent with the public interest while at the same time encouraging creative, less-costly, market-based solutions to the problem. In environmental regulation, for example, incentives for new processes for reducing harmful emissions may be more effective than requirements to use "best practice" processes (as Chapter 11 by George Heaton and Darryl Banks explains).
(2) Competition and Anti-trust Policy
Although anti-trust policy is not seen by many people as a tool of technology policy, its importance is illustrated by the National Cooperative Research Act of 1984 . The willingness of anti-trust authorities to define markets in global terms has given more freedom to U.S. firms that, though they might be well positioned in the United States, are hard pressed by foreign competitors. There is a strong connection between anti-trust policy and public-private technology partnership in that the participation of government officials in an industry consortium can provide protection against some anti-trust abuses. At the same time, new forms of monopoly power may be arising with new technology, and government must continue to ensure that market power is not used to suppress useful innovations.
(2) Creating and mediating markets: standards policy and procurement
The government plays a unique role in creating markets where they have failed to form and in harnessing the power of the market in the public interest. Examples include issuance of tradable permits in the regulation of air pollution, and the use of market power arising from its own massive purchasing activities to influence voluntary industrial and product standards. These functions present important opportunities to influence and advance innovation. Government should, wherever possible, set performance rather than design standards. Similarly, it should use both procurement and regulatory authority to encourage innovative solutions that, in the long run, have the most promise to achieve federal objectives at minimum cost.
In almost all of private industry''s industrial standards processes, including international but non-governmental bodies such as the International Standards Organization (ISO) and its sister body the International Electrotechnical Commission (IEC), governments do not set standards but do play a facilitating role. A government can ensure that its own procurement standards are commercially compatible wherever possible, as the DoD has undertaken in its reform of its military procurement specifications (MILSPECs). It can also contribute personnel and resources directly to advance standards processes, as it did through participation in the Internet Engineering Task Force (which sets consensus standards for the Internet). This participation did much to hasten Internet growth and flexibility.
(2) Tax incentives
Tax policy has many direct and indirect effects on private sector investments in innovation, ranging from the way corporate R&D costs are allocated to foreign-source income, to depreciation rates for new scientific instruments and production tools. The most widely debated tax issue, however is the Research and Experimentation tax credit. (See Chapter 8 by Scott Wallsten). Although many consider it to be an attractive idea, others identify flaws, and it has never won permanent enactment by Congress. The R&E tax credit is a blunt and expensive instrument which seeks an overall rise in the level of R&D across all industries and regions. It rewards firms that are rapidly increasing their R&D investments, but does little for those industries most severely challenged by international competition.
A more effective use of the tax code to encourage innovation might be to design a more specifically targeted tax credit that encourages the diffusion of existing technology, encourages the purchase of research from universities, or lowers the cost of sending employees to mid-career training programs. Targeting industries and firms at specific stages of development for tax incentives could also foster innovation.
(2) Intellectual property policy as a technology driver
Intellectual property policy may be the most troublesome but important "indirect" tool in the toolbox. The way in which the United States develops its patent law to conform with other nations as required by the General Agreement on Tariffs and Trade (GATT) can have profound influence on U.S. innovation rates. The U.S. Patent and Trademark Office recognizes this fact, but tends to define innovation as synonymous with patent filing, when many of the most dramatic innovations have evolved with only modest recourse to patents. In some industries, for example computer hardware, patent policies are largely defensive; the larger firms worldwide are linked by cross-license agreements. In other areas such as biotechnology, however, rigorous patent protection is absolutely critical.
Similarly, changes are sweeping over copyright protection and trade secret and trademark protection, both because policy and enforcement internationally are not uniform and because many experts feel that some new form of protection, with features of both patent and copyright law, is needed. The executive branch and Congress must recognize that balancing intellectual property protection and acceptable use— that is, incentives to put knowledge in the public domain or license it free of charge— is an essential element of technology policy and should be developed in concert with the other tools in the toolbox.
(2) Technology Roadmaps
Since, as noted above, different industries, different technologies, and even different regions of the country follow different patterns of innovation, the federal government must have a clear understanding of these differences (as argued in Chapter 15 by Daniel Roos, Frank Field, and James Neely). This understanding can only come from outside the Washington beltway— technological and market understanding from the private sector, workforce needs from labor, and the promise of emerging science from the universities and other laboratories. But it is not enough for these groups to share their knowledge; the government must have the institutional capacity to absorb and use it.
To that end, the technology roadmap— a consensus articulation of a scientifically informed vision of attractive technology futures— offers an interesting tool with which to organize and shape the technology needs, both of government and industry. Roadmaps are informed by research capabilities, the state of technological development, market trends, and public sector and industry needs and priorities, to assist policy makers and the research community in resource allocation and agenda setting. The technology roadmap designed by SEMI/SEMATECH offers a ready template for ways to join the interests of the federal agencies, the university research structure, and the research capabilities of the private sector. More than performing a census and making surveys, the roadmap requires applied policy analysis and design, and an informed and open debate about alternative technology policies and strategies for every sector of the economy. Modification of some of the legal constraints against officials seeking advice from individuals outside of government might be appropriate in this circumstance.
Roadmaps have value not only in fostering communications and common purpose among a variety of research organizations working toward similar objectives, but can also be helpful guides as technologies evolve over time. Very few exciting innovations come to practical fruition in a single effort. Many more modest steps are taken, each raising new possibilities whose exploitation requires solution of new problems. Thus government projects of limited duration rarely create radical new technological capabilities. Indeed, both the TRP and the ATP programs suffered from unrealistic expectations in this regard, since neither program was designed for a time sequence of ever more ambitious steps towards highly valued goals. Technology roadmaps can track those development trajectories and provide more realistic bases for program evaluation than can the examination of results from single projects, one at the time.
Technology roadmaps have another useful function as guides to research opportunities of value to industrial and other user communities. As noted in Chapter 14, roadmaps tell researchers where intellectual effort is most likely to bear practical fruit, but do not constrain the researchers'' choices. The creation of such roadmaps requires the collaboration of well-informed technologists with scientists, and is a valuable role for industry consortia, working in collaboration with universities and national laboratories. Other nations, particularly Japan, the United Kingdom, and Germany, have made use of a related technique, the Delphi survey of technical opinions, to construct a map of research opportunities of value to society for use in budget planning and project selection by researchers. In the United States the analogous activity has been the identification of lists of "critical technologies," for which the Critical Technologies Institute (CTI) attached to the Office of Science and Technology was created. Technology roadmaps generated in collaboration with industry would be much more useful as a guide to government research investment priorities.
(1) Leverage Globalization Of Innovation
Innovation opportunities in the United States increasingly require access to foreign resources and markets and compatibility with the policies of other countries. Information infrastructure, where U.S. firms and institutions enjoy a commanding lead today, illustrates the need for U.S. leadership in developing a harmonious international environment. While trade conflicts tend to capture more public attention, trans-national collaboration and cooperation in the development of new science and technology can bring even bigger benefits.
U.S. technology policy should encourage and facilitate globalization and trans-national collaboration. The United States must learn to cooperate as well as compete, given the rapidly growing technical assets in other countries (assets which are in many cases the product of public investment). The criterion for participation should remain the U.S. self-interest, but not defined in a zero-sum fashion. In most cases, the United States will prefer to leave such collaborations to firms, universities, and technical associations without government involvement. But the U.S. government should consult with U.S. firms, and should take the lead in defining the forms of trans-national technological cooperation that will be most useful to both business interests and the U.S. public interest.
This strategy suggests that the United States should aim for an international code of investment policy, such as that being developed by the Organization for Economic Cooperation and Development (OECD), that allows the United States to permit foreign-owned R&D establishments in the United States to participate in domestic technology programs, such as public-private partnerships, by ensuring that foreign subsidiaries of U.S. firms enjoy equivalent and comparable access. The United States Trade Representative should take increasing responsibility to press for the multilateral elimination of foreign direct investment constraints under the World Trade Organization, and for establishment of a Multilateral Investment Agreement. Such an agreement would provide equal treatment for international investors with regard to market access and legal security.
The United States should also press for multinational agreements (discussed later in this chapter) to design, construct, and use the very large facilities required for modern science, such as accelerators, telescopes, and fusion research reactors, as well as global scientific monitoring projects, such as bio-assays of the seas and monitoring of the global environment.
(1) Improve Government Effectiveness in Policy Development
Strong, stable, and continuing federal support for a research and innovation policy is of critical importance to the contribution of research and innovation to economic growth, job generation, and rising living standards, and also to the long-term, cumulative nature of the pay-off in those areas from government investment. This is especially true under the policies recommended here, which generally move away from public funding of proprietary commercial technology (except in cases where the government expects to be the purchaser of the resulting products, as in defense procurement). Just as nonpartisan support for science is essential to give scientists time to pursue unexpected opportunities, so too basic technology research needs continuity of support to maximize public returns over short-term private gains. Multi-year continuity in appropriations for research investments are crucial, and Congress must sustain a consensus policy through the ups and downs of political change. It is an unavoidable dilemma of technology policy that for the most appropriate federal role— investing in basic technology research— the time required to realize economic benefits is longest and least visible, while it is quickest and most visible for the least desirable investments, such as federal subsidies to product development.
The American science and technology system thrives because it is pluralistic. There are many sources of support, many types of performers, and a maze of linkages amongst funders, performers, and users of science and technology. This pluralism was intended by the nation''s founders, who included a patent clause in the Constitution, rejecting monopolies protected by the Crown. Pluralism has grown with the nation, through the Morrill Act establishing the land grant colleges in 1862, the development of corporate R&D laboratories in the early decades of this century, and the defense, energy, space, and medical research complexes of the past fifty years. Pluralism is necessary because the outcomes of technological innovation cannot be predicted; they can only be discovered through real-world trials of competing ideas and institutions.
Making policy which can target and stimulate, not constrain or stifle, this diverse, loosely coupled environment requires the ongoing participation of stakeholders. There is not and should not be a single centralized technology policy process that undermines pluralism. The mix of technology policy tools must be adapted across regions and industries in accordance with particular needs and established patterns. Attempting to specify the exact mix from the center would be both foolhardy and counterproductive.
While the market provides a relatively demanding environment for selecting the most promising private sector organizational and technological innovations, the policy process in the public sector has generally not lead to equally sharp selection mechanisms. In part this is because of the difficulty of finding sufficiently measurable and objective indicators of success for public policy goals that work as well as profitability works for corporate goals (as Adam Jaffe explains in Chapter 3). Nonetheless, it is essential that policy-makers attempt to define their goals, to measure the effectiveness of technology policy in reaching them, and to use these measures as feedback for improvement. Efforts at comparison— with other nations, across missions and regions, and with private sector innovation processes— are essential. Intentional experiments, in which different arrays of tools are voluntarily employed by different agencies and their state, non-governmental, and private partners, should be encouraged. Every program should have an explicit element for evaluation and learning.
(2) Stakeholders must be involved in technology policy development and delivery
The federal government must have access to advice, data, and analysis that will help it tailor technology policy to fit particular industries and regions. Involving all the stakeholders— scientists and technologists, state and local governments, industry, labor, professional associations, and public interest groups— will bring new perspectives to the development of technology policy options and new rigor to testing them. The federal government often lacks information about actual practices and relationships on the ground as well as about scientific and technological opportunities. Especially now that the U.S. Congress Office of Technology Assessment (OTA) is no longer available as a source of analysis, a gap remains in the area of informed policy assessment. The White House Office of Science and Technology Policy (OSTP) should serve as the nerve center for this advisory process in support of the president and the NEC. The President''s Council of Advisors on Science and Technology (PCAST), if better financially supported, and a restructured Critical Technologies Institute (CTI) (a Federally Funded Research and Development Center dedicated to the support of the Executive Office of the President and other federal agencies) should be viewed as valuable resources for OSTP in this regard. Since PCAST also advises the president on science policy, it may be useful to link technology policy-making to the National Science Board''s (NSB) responsibility for both science and technology basic research by appointing the chair of the NSB to PCAST.
One important requirement is that the president have direct access to experienced technical professionals and innovators from the private sector. In addition to a better supported PCAST, this can be provided by reliance on the DOC Technology Administration''s direct channels of communication with private sector technology decision-makers on issues related to innovation and commercialization of R&D. A Research and Innovation Board, advising the secretary of commerce and staffed by the Technology Administration, could strengthen those private sector linkages, as indeed the Commerce Technology Advisory Board did in the 1960s and 1970s. The Board could oversee the development of technology roadmaps and the department''s sectoral industry studies (as recommended by Daniel Roos, Frank Field and James Neely in Chapter 15), and offer a potential venue for coordination and information sharing between government policy makers and private industry.
(2) State-Federal Relationships
State governments are increasingly responsible for the delivery of technology and training services and, more generally, for technology diffusion and utilization. Furthermore, innovation-based development strategies, especially those that are regional in nature, call for integration of innovation programs with training and education, as well as with other areas of policy for which the states are better equipped than are federal agencies. If state-federal collaboration is to combine the best capabilities of both, the states need earlier and more influential access to federal technology policy-making processes while adjustments are still possible (see Chapter 16 by Christopher Coburn and Duncan Brown).
The Executive Office of the President, in collaboration with the National Governors'' Association, has established the U.S. Innovation Partnership. This joint effort of administration technology policy officials and governors is staffed by the Technology Administration in the Department of Commerce and seeks to ensure better coordination between state and federal efforts in the encouragement of innovation. This effort should be encouraged and supported as a mechanism for policy coordination in those areas where the states have important roles and assets, such as state-federal investment in technology-based economic development, the construction by state and municipal governments of the Intelligent Transportation System, and strategies for environmental containment, monitoring, and remediation.
(2) The Role of Congress
Congress is central to the policy development process. Members of Congress have narrower constituencies and, as a group, face the electorate more regularly than does the president. It is understandable that many members will be eager for immediate results. Given the pressures members face, the administration should support them by advancing policies that recognize more fully both the importance and unpredictability of technological innovation and the necessity for the federal government to play a facilitating role in it. One key to this is establishing a basis for federal investment which clearly defines the federal role and offers a mechanism through cost-shared partnerships that does not cause inequity or market distortion. Congress has valuable knowledge to add to the technology policy process. For example, members can identify key stakeholders and ensure that their voices are heard. The Congressional support agencies perform needed program evaluations. Congress also has a central decision-making role, adjudicating among technology policy experiments. Policies that have failed must be killed; policies that succeed must be maintained or expanded. But most importantly, policy experiments have to be tried for a sufficiently long time and under sufficiently reasonable conditions in order to be judged. In its first term, the Clinton administration pursued aggressive expansion of its technology agenda. The 1994 reaction by the majority in Congress, aiming to hamstring or halt these experiments, was equally aggressive. A more patient, long-term approach by all parties can make second term efforts more stable and productive.
(2) A White House advocate for technological innovation, linked to economic policy
The president and the White House staff must serve as advocates for research and innovation policy and coordinate the array of federal S&T activities. The Executive Office of the President must also provide the locus for linking the technology agenda with broad national policy objectives, such as economic and security policy (as David Hart discusses in Chapter 17). The president has the long time-horizon and national purview that justifies an intense interest in innovation that will advance the quality of life in the United States. The president has the strongest incentive and the authority to promote change in the face of vested interests bent on preserving the status quo. This does not mean that the White House should advocate every R&D program. Instead, the White House should be the arbiter for policy options that encourage technological innovation in the pursuit of agency missions (including non-R&D options), particularly in times of tight budgets when investments in the future tend to be squeezed out. Such options should be compared to and traded off against non-R&D alternatives for encouraging innovation in the pursuit of the same mission.
The major White House policy councils— the National Economic Council, National Security Council, and Domestic Policy Council— are the most appropriate forums for framing technology policy options and integrating them into the larger issues with which they are concerned. The council staffs should play major roles in developing these options, but should rely on OSTP (and where appropriate the committees of the NSTC) for the technical judgments on which the efficacy and practicality of those options rest. OSTP should have close linkages to all of these councils, as well as to OMB, at the staff level and through the science advisor''s service as a principal member in many of their deliberations. The National Science and Technology Council, through which OSTP reaches out to the research and innovation units of federal agencies, can serve as a channel to the major policy councils for agency technology policy ideas. NSTC should coordinate multi-agency programs and identify a few key budget priorities.
(1) Applying These Principles to Determine When Government Should Invest in Research
These six principles for a research and innovation policy are intended to provide a clear and politically robust framework to aid in program and policy development. They do not lead to discrete answers, but instead are intended to inform decisions and provide a litmus test for program evaluation and improvement. Before any investment of federal moneys in technology can be made, one must ask: is there a significant public problem that needs to be addressed by the federal government? Are there no other entities, such as private firms, that are better positioned to offer a solution and are motivated to invest in it? Implied in this second question is the idea that any federal research investment should not only be of net benefit to the nation— the benefits will exceed the costs— but also that public benefits will substantially outweigh private returns when the government addresses the problem through public-private partnerships. If the problem passes these tests, one must then ask, is the federal government in a good position to solve the problem? Does it have the skills, experience, and political legitimacy to bring about the desired results in a manner acceptable to all those affected? Beyond satisfactory answers to these primary questions comes the realization that our federal government increasingly faces a future of finite resources. Priorities must be set and difficult choices made.
Two historical priorities should remain at the top of the list: (1) basic scientific and technological research: investments in the long-term future scientific and technological capacity of the people of the United States (in knowledge and human resources), for which there is no other resource than government; and (2) mission research: investments in research whose urgency derives from a high-priority public function, such as defense, health, environmental protection, and raising the standard of living, and where the work is of top priority within that function. The first, basic research, although consuming a minor fraction (about 21 percent) of federal R&D resources, often produces striking long-term benefits. Mission research, the second historical priority, receives larger investments, but it remains important in fulfilling critical social needs. Where commercial interests are aligned with the technology development interests of the federal government, public investment can and should be highly leveraged by private investment.
When the primary goal of public investment in technology is the stimulation of private sector innovation, government should concentrate most of its investment at the most creative points in the innovation process. The appropriate mix of federal investment in basic science and technology research enables creation of new technologies and informs choices among technical strategies. When developing products for its own use, as in the military and space programs, it is appropriate for government to fund not only research but development as well. However, product development is not an appropriate area for government funding when the goal is general economic development. Thus, while funding of basic science and technology research to support all areas of government activity is appropriate in principle, how far government goes beyond this toward commercialization should depend on the extent to which there is a public stake in the production of the end products. Rigorous program evaluation, based on metrics designed for the purpose, is of course indispensable. But there is a basic limitation on the quest for metrics that will quantitatively express economic outcomes in terms of appropriable returns. When such calculations can be made, that fact may suggest that private investors could, and should, capture these returns. The case for government investment may be highest in just those situations in which the benefits are hardest to quantify in dollar terms: basic research, education, and investments in social capital generally.
Finally, let us emphasize that determination of the government role is not to be made on the basis of the kind of research being performed. For this purpose it is not relevant whether the investigator is motivated by curiosity about nature or the satisfaction of solving a problem in which a sponsor may be interested. The criterion rests on the type and magnitude of returns to the public interest, in relation to any private benefits that may have been created.
(2) Choosing Research Performers
How should agencies choose among firms, their own laboratories, independent laboratories, and universities when funding research? We believe they should select the institutions most competent to perform the work at a high level of excellence and able to diffuse the new knowledge to those who can best use it, thus ensuring that public benefits will exceed private revenues. Thus the public interest in the work matters, and so too does the effectiveness of the means for diffusion of the results to society.
The paths by which public value is created must be considered when selecting performers. Firms engaged in federally-funded research only need to move the ideas from laboratories to business units to see them put in practice, but until the research reaches many other users, public benefits may not be realized. In contrast, students trained in universities take the new knowledge with them to their jobs. Government-funded national laboratories may be particularly capable of interdisciplinary work, exchanging ideas across disciplinary boundaries, and working with industry through CRADAs. There is no single right answer to the selection of performing institutions. Indeed, a mix of institutional performers may be best, since industry, universities, and national laboratories each have their unique ways of ensuring that new work reaches the users.
Where the government funds one hundred percent of the work, the university, national laboratory, private firm, or consortium that performs the research must be committed to effective diffusion of the results. Universities that refuse to let collaborating firms restrain the right to publish, and whose students are free to use all the knowledge they have gained, will meet this requirement automatically. National laboratories should, as recommended by the Galvin Commission for the DOE, restrict themselves to work conducted under university-like conditions (both basic science and technology research) and to mission-related work in which the beneficiaries (the users of the technology) are identified and adequate mechanisms for diffusion are planned and provided for. The CRADA mechanism is one such tool (as David Guston explains in Chapter 9).
The more difficult issue arises when government provides cost-shared funding to private firms, either individually or as part of consortium. In this case the consortium may be preferred, since the effective diffusion of results is more nearly assured when a number of firms collaborate. In the case of those consortia made up of potential rivals, natural competitive pressures will tend to keep the research less proprietary and more long-range in nature, thus more "basic." However, as Linda Cohen observes in Chapter 7, the failure of the Technology Reinvestment Project (TRP) flowed from just this situation. One group of competitors (defense prime contractors) objected to cost sharing work intended to enhance, at their expense, the market opportunities of commercial firms in the same consortium.
A balanced portfolio of publicly-funded research leading to new science and new technology is the right approach to federal research funding, not only for the National Science Foundation (NSF) and the National Institutes of Health (NIH), but also for the long-term investments supporting the missions of agencies such as the Department of Commerce (DOC), the Department of Energy (DOE), the Department of Defense (DoD), and NASA. It should also be a focus of much of the government''s funding for research in public-private partnerships such as the Advanced Technology Program, the Program for a New Generation of Vehicles, and the Environmental Technology Initiative, recognizing that industrial partners will also be investing in research and development addressing their specific commercial interests. Thus when government funds research in industry, government may take a share of the technical risks, but firms should take most of the market risk. If this is the division of labor, much of the concern about government R&D programs distorting markets, substituting for private investment, or causing anti-competitive impacts on individual firms should be allayed. In addition, basic technology research, while longer-range in nature than product development, will generally be less expensive than large-scale technology development and demonstration, allowing some additional resources to shift from development to research.
When federal agencies collaborate with consortia or individual firms, it is appropriate for the government to focus its investment on that part of the program which has a potentially high payoff to society, based on an attractive, if risky, technical opportunity. This has been the primary emphasis in NIST''s ATP program. Keeping the technical goals ambitious and attractive means a move away from technologies likely to be quickly commercialized. To the extent that a substantial part of the work is not proprietary, the rate of diffusion of benefits would increase, and concerns about anti-competitive effects (which may arise regardless of the rigor of the competitive selection process) might be reduced. But political expectations for quick commercialization and evidence of immediate economic returns from partnership programs provide a strong incentive to program managers to invest in just those kind of research subsidies to which the Congress quite properly objects.
(2) National Laboratories
One of the most successful institutional innovations of the Cold War period was the government-owned, contractor-operated national laboratories (GOCOs). Most of them have highly qualified staffs; many of them have pressing missions. But though they dominate the R&D funding portfolio of the federal government (receiving twice as much federal funding for R&D as the universities), they have had considerable difficulty in adjusting to the changing needs of the nation. This is not necessarily a failure of laboratory management, but more often results from the rigidity of agency authorities and Congressional committee charters. The Clinton administration attempted, with little success, to come to grips with the realignment of mission and the possible down-sizing of these laboratories during its first term in office. Managing priorities for funding of national laboratories and providing effective mechanisms for linking the laboratories to potential beneficiaries of their work requires collaboration with both Congress and the states.
Cooperative research and development agreements (CRADAs) solve some of the problems faced by national labs. These agreements and related mechanisms were initiated in the 1980s and, while already viewed favorably by most observers, can be made more effective. One way is to encourage the sharing of personnel so that the deep cultural canyons that often separate technology producers from users, and federal researchers from industrial researchers, can be crossed.
(2) Very large technology projects
In the past, technology policy often centered upon taking a few large risks, pouring large sums of money into a few technologies (e.g., synthetic fuels, the supersonic transport) in the hope that such a concentrated investment would ensure success. These projects were often politically attractive but rarely achieved their goals. A better strategy is to make numerous, smaller technology bets, and to do so with a project selection mechanism that is market-based, wherever possible. The federal government should follow the NSF/NIH model of relying primarily on relatively small grants spread out among many performers, awarded competitively but funded over multiple years. It should fund a variety of technology areas chosen with input from the technical experts from the private sector as well as from research institutions, as is done in ATP. It should achieve scale, where it is needed, by encouraging groups of institutions to collaborate in formulating plans for diversified research and the diffusion of the results.
Of course, there are some specialized research facilities— telescopes, oceanographic ships, accelerators, and the like— which, by their nature, require large capital investments, and without which science cannot progress. Many of these projects will, in the future, have to be planned and executed as international collaborations. Such large-scale international collaborations can prove effective, given sound scientific and economic— rather than political— justification, and resilient institutionalized funding mechanisms (conceptually the same type of criteria we would advance for funding of national-level, industry consortia). To address the issue of commitment and accountability, there must be an international institutional context for which the U.S. Congress and president feel ownership and in which they have confidence, perhaps by means of a multi-year appropriation of funding, subject to performance review against agreed milestones. Where the need for federal help arises from a commercial technology of such extreme cost that firms and their sources of finance are unable to accept the risks, such as might arise in a future supersonic transport aircraft, government should look to the array of policy tools discussed earlier in this chapter. One such tool could be for government to underwrite some of the capital risk, leaving it to the private investment community to evaluate the commercial merits of the venture.
Having explored in general terms the application of our six principles for a research and innovation policy, the chapter now turns to the specific case of the ATP program.
(1) Applying Research and Innovation Policy Principles to ATP
The NIST Advanced Technology Program (ATP) is a federal program that uses cost-shared, competitively selected support of advanced technological research as a stimulus to industrial innovation, and a means to benefit the economy (see Chapter 6 by Christopher Hill). Government expects to see measurable economic outcomes in due course, but must avoid interfering with competitive markets. As described earlier, this tension makes cost-shared technology investment an inherently uncomfortable area of policy development. Despite this, ATP has weathered ongoing scrutiny from policy-makers. The technical quality and applications potential of the work are the primary measures of ATP success. Government and industry goals for the program differ, but are not mutually exclusive, or even necessarily at odds. Commercialization, in which the government does not directly participate, is important to the participating firms. Diffusion of the technical achievement is of special interest to the government.
Applying the principles outlined here to the ATP program structure and funding guidelines can result in clearer definition of legitimate roles for the program participants, public and private. This could serve to strengthen the program, expand its economic benefits, provide clarity of mission to increase political support, and maximize the diffusion potential of technology outcomes. Such an effort would include an emphasis on basic technology; a commitment to consortium-based, cost-shared investment; and strong links to state and local economic development institutions. The model we suggest takes a more active economic development approach than the current program, through significant institutional reform. It contrasts with Christopher Hill''s recommendations in Chapter 6, which leave the program framework intact, and offer improvement on the margin. Both approaches are consistent with our research and innovation policy principles, as are the set of program refinements announcements by Secretary of Commerce William Daley on July 10, 1997. There is no one right structure that manifests from these guidelines. Instead, they are intended to provide a broadly agreeable conceptual basis from which program specifics can be developed.
The structure of the NSF State Systemic Initiatives Program offers a model for repositioning ATP with an emphasis on diffusion and regional development. Such a model suggests inviting states— or regional groups of states— to compete for selection by NIST as a regional technology-based economic development program. To qualify for selection, the states would identify an industry sector— perhaps including the main elements of the supply chain as well— and create a non-profit consortium of institutions representing state agencies, the selected industries, labor, finance, education, and research. State-funded technology and economic development programs might participate and share the costs. The consortium might also produce technology roadmaps to guide the selection of basic technology research opportunities to be pursued through ATP projects, enhancing the opportunity for universities, national laboratories and small to medium-sized firms to participate. The proposal would not only make a case that the chosen sector is ripe for dramatic technical progress, but would also outline the consortium''s commitments to investment, exports, worker training, and other activities that would maximize the economic leverage to be gained from ATP research support. The consortium might take responsibility for choosing specific firms for participation, subject to evaluation by NIST of technical merit. The consortium would also agree to make post-project evaluations for NIST of the economic outcomes of the total program.
This approach gets more economic leverage for NIST dollars, preserves the national standard for technical excellence, devolves to the states the task of selecting among individual firms, and creates a broadly-based constituency for the ATP program. Since ATP already invests most of its funds through "focus" areas— specific industry sectors with a compelling case for technology support— this would simply extend the principle by sharing responsibility with the private sector and the states. In short, collaboration by the states, firms, and laboratories participating in the consortium would strengthen the social capital that underlies the U.S. capacity for innovation. (See Jane Fountain''s discussion of social capital in Chapter 4.) Incentives in the ATP program to encourage this type of networking and collaboration would alone provide a valuable economic contribution.
(1) Conclusion
Just as the principles described here can be applied in a non-partisan framework to NIST''s Advanced Technology program, they can be applied one by one to the full portfolio of technology programs. As with ATP, the principles do not lead directly to specific program agendas and definitions; rather they guide program development within a politically and economically rational framework. Our six principles can be seen reflected in the specific program evaluations in each of the chapters.
The administration set the right tone in the president and vice president''s 1993 policy statement, which advocated a research-based investment strategy to help "private firms develop and profit from innovations," recognizing that the right incentives can direct the private sector''s powerful innovative capacity to broad public purposes, including a rising standard of living and quality of life. If the federal government''s policy gives new emphasis to its own participation in the network of universities, laboratories, and firms that is rapidly developing under world economic forces, its investment in research can create the world''s strongest base of shared technical knowledge.
For too long both the political and the scientific communities have squabbled over narrow distinctions between "science" and "technology" as targets for federal investment. These distinctions fail tell us what the proper roles are for the public and private sectors, and lead to neglect of investments in the nation''s knowledge base and innovation capacities. What is needed instead is a research and innovation policy that can continue to build America''s innovative capacity through good times and bad, and through the ebb and flow of political philosophy. The nation is nearing the end of the century in which it has demonstrated the most creative and productive scientific and technological enterprise in the world. The driver was the threat of war. It is now time to establish the principles that can sustain this progress, not for fear of war but in anticipation of rising quality of life, a clean environment, and a healthy citizenry. If both scientific and technological research are strongly supported and the full range of other policy tools are applied to encourage innovation, public policy can safely leave to industry the realization of economic progress.
