Economic Perspectives on Public Support for Research
This chapter examines the economic logic of public subsidies for research and development (R&D) activities in general. The first section notes a number of serious theoretical objections that can be raised against public support of R&D, and it reviews empirical considerations that reaffirm the general presumption that, without government support, market failures will result in too few resources being allocated to expanding scientific and technological knowledge. The second section takes up the special considerations that bear on the economic case for public support of exploratory, open research—the sort that is usually designated as basic science, however unsatisfactory that label may be. That discussion emphasizes the complementarities and guidance that such research creates for private-sector, applications-oriented, proprietary R&D, rather than the possibilities of spin-off products that may compete with results targeted by industrial research organizations. It also highlights the contribution federal funding makes to the education of the scientific and engineering workforce.
The Economic Rationale for Public Support of Civilian R&D
During the past 30 years, economists have worked out cogent reasons why the price system and competitive markets should not be expected to do a good job in producing or distributing knowledge and information—certainly not by comparison with markets' performance in similarly allo-
cating resources in more conventional, tangible commodities such as fish or chips (of both the computer and the potato varieties).1 This conclusion rests on the fundamental insight that ideas—especially ideas tested and reduced to codified scientific and technological information through R&D activities—have some important attributes found in public goods, goods that are widely available to individuals whether or not they paid for them. Correspondingly, they may be better understood by studying other public goods, such as a smog-free environment or defense against nuclear missile attack.
Information and Knowledge as Commodities
An idea is a thing of remarkable expansiveness: it can spread rapidly from mind to mind without any reduction in its meaning and significance for those into whose possession it comes. Thomas Jefferson remarked upon this attribute, which permits the same knowledge to be used jointly by many individuals at once: ''He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine receives light without darkening me. . . ." Economists have pointed out that the potential value of an idea to any individual buyer generally would not match its value to the social whole. The latter value, however, is not readily expressed in a willingness to pay on the part of all who would gain from the illuminating idea. Once a new bit of knowledge is revealed by its discoverer(s), some benefits will instantly spill over to others who are therefore able to share in its possession. Commodities that have the property of expansibility, permitting them to be used simultaneously for the benefit of a number of agents, are sometimes described as being nonrival in use: although the cost of the first instance of use of new knowledge may be large, in that it includes the cost of its generation, further instances of its use impose at most a negligible incremental cost.2
This formulation ignores the cost of training potential users to be able to use new information. Although it is correct that there can be fixed costs of access to the information, these costs do not invalidate the proposition that reuse of the information will neither deplete it nor impose further costs. It may be costly to teach someone how to read the table of the elements or use differential calculus, but any number of individuals thus instructed can go on using that knowledge without incurring further costs.
The second feature of ideas is that it is difficult, indeed costly, to retain exclusive possession of them while putting them to use. Another disadvantage of exclusivity is that results obtained by methods that are not or cannot be revealed often are felt to be less reliable. Of course, it is possible to keep a piece of information or a new idea secret. Producing results not achievable otherwise, however, indicates the existence of a
method for doing so. Even a general explanation of the basis for achieving the observable result jeopardizes the exclusivity of its possession, for knowing that something can be done is an important step toward discovering how it may be done.
The dual properties of nonrival usage and costly exclusion of others from possession define what is meant by a pure public good. The term "public good" does not imply that such a commodity cannot be supplied privately, nor does it mean that government must produce it. But competitive market processes will not do an efficient job of allocating resources for producing and distributing pure public goods, because such markets work well when the incremental costs and benefits of using a commodity are assigned to the users.
Capturing the Benefits of Research Investments
One may see the problem posed by the public goods characteristics of knowledge by asking how ideas can be traded in competitive markets, except by having aspects of their nature and significance disclosed before the transactions are consummated. Rational buyers of ideas, no less than buyers of coal, and of fish and chips, first want to know something about what they will be getting for their money. Even if the exchange fell through, the potential purchaser would enjoy (without paying) some benefits from what economists refer to as transactional spillovers. These occur because there may be significant commercial advantages from acquiring even general information about the nature of a discovery, or an invention—especially one that a reputable seller has thought it worthwhile to bring to the attention of people engaged in a particular line of business.
This analysis leads to the conclusion that the findings of scientific research, being new knowledge, would be seriously undervalued were they sold directly through perfectly competitive markets. Some degree of exclusivity of possession of the economic benefits derived from ideas is necessary if the creators of new knowledge are to derive any profit from their activities under a capitalist market system. Firms can protect their knowledge either by seeking patent or copyright protection or by trying to keep it secret. Patents and copyrights provide legally enforceable means of protecting knowledge, but they require that inventors publicly disclose the workings of their inventions (e.g., through a patent application), enabling others to learn from their work and to find alternative means of achieving the same end (i.e., reverse engineering a particular device). Keeping a trade secret (if done effectively) avoids public disclosure, but offers little means for legal resource if others learn the secret (unless they use unlawful means to do so). Industries vary in the degree to which firms prefer to seek intellectual property protection versus keep-
ing trade secrets. Patents tend to be very important in pharmaceuticals, for example, but less so in computing. Regardless of the mechanism chosen for protection, imposing restrictions on how ideas may be used saddles society with the inefficiencies that arise when monopolies are tolerated, a point belabored by economists ever since Adam Smith.
Technical Standards as Public Goods
Technical standards also demonstrate characteristics of public goods in that competitive markets often fail to produce them without public assistance. Technical standards acquire economic value for their possessors only as a consequence of being publicly disclosed and jointly used, and they actually grow in utility for the individual user in proportion to the degree of universality in their adoption. Many technological and engineering reference standards, such as those for the thread sizes of nuts and bolts, or the diameter of optical fiber (to permit splicing without degrading the light signal that is propagated through the inner core), benefit buyers and vendors by reducing transactions costs and permitting economies of scale in production, especially when they are widely adopted.
It should be noted that many other reference standards have emerged from the work of scientific communities, such as the units in which electrical current, resistance, and power are measured. The ampere, ohm, and watt, like the joule, angstrom, and countless other precisely specified units, provide a standardized terminology that facilitates scientific communications. They thus enable individuals in a distributed research network to work together (i.e., become interoperable) in the way that compatibility standards enable interacting components of systems to achieve greater functionality. As is the case with other standards, market incentives are weak for producing and distributing scientific reference standards.
Firms that know of and wish to use technical standards would have every incentive to freely share that information, in order to encourage others to follow suit. Hence, an adequate supply of reference standards and related technologies may not be forthcoming through individual private enterprise, as it may not be worthwhile for any single firm to undertake the cost of designing a reference standard that would be useful for the industry as a whole and redistributed freely.
Governmental support for the collaborative development of reference standards, or direct funding of agencies such as national standards institutes that undertake such work, constitutes a mechanism for rectifying this market failure. The alternative of using intellectual property rights protection to grant monopoly privileges to private developers of such standards has a perverse effect. It tends to restrict the extent of the
standards' use, and therefore deprives even those who do pay the monopolist's charges, imposed by licensing of patent-protected standards, from enjoying the added benefits that would accrue to all users from enlarging the user community. This is a generic problem with standards for systems, such as telephone and other communication systems, whose value to individual subscribers is enhanced by being able to connect with, and be contacted by, a larger number of network members.
Secrecy and Intellectual Property Rights
Some suggest that the problems of incomplete appropriability of benefits from research are overstated, or indeed nonexistent, because industrial secrecy is sufficient to protect against some firms free-riding on the R&D investments of others. But other factors must also be considered. First, one has to consider what costs a strategy of secrecy imposes upon private enterprise, and whether such practices can be totally effective in the face of the mobility of technical personnel and reverse engineering. Second, one must look at the matter from the societal viewpoint. On the supposition that extensive secrecy was a viable policy for firms engaged in research, what is the potential for wasting R&D resources by duplicating research, not to mention potential injury to consumers, were the developers of new products and processes actually able to maintain indefinite secrecy about their research results?
The economic logic of providing intellectual property rights in science and technology is that this is a better choice, from the societal standpoint, than secrecy. Modern economic analysis has come to view the granting of patent and copyright monopolies as a sacrifice of short-run consumer interests that may be justified by far greater long-run gains derived from giving creators of new, useful knowledge more secure pecuniary incentives to reveal it rapidly to the public. Still, in order to pursue research profitably, it is necessary for firms to be able to control the flow of information about work that is in progress, and to build an inventory of potential future projects that they can expect to exploit, rather than seeing these walk out the door with their research personnel.
Consequently, trade secrecy protections are in this respect complementary to intellectual property protection in the production process for new research findings whose benefits the firm expects to be able to appropriate. This reinforces the argument made for strengthening intellectual property rights in patents, and their enforcement, on the grounds that reliance upon secrecy is reduced thereby. Although the disclosure of codified information is augmented by patent systems, so is the inducement to curtail the transmission of tacit knowledge that might reduce the commercial value of the patents that are being sought.3
Common Pool Problems, Patent Races, and Potential Overinvestment in R&D
Market failures do not necessarily result in underinvestment in R&D by profit-seeking firms. There also is a potential for excessive private investment when expected private marginal rates of return are not matched by the marginal social value expected to result from those expenditures. Economists have been aware for some time of three main situations in which that is likely to be the case, and, although these are thoroughly treated in the technical literature, they often go unmentioned in public testimony on the subject. These overinvestment pathologies of market competition through R&D go under the labels of business stealing, common pool problems, and racing behavior.
Business stealing refers to the situation that arises when research that is directed toward displacing a competitor from the market entails developing a new product or process that largely duplicates the functions of things that already exist, but adds some distinctive additional features. Achieving a marginal improvement in quality may be sufficient to capture a rival's share of the market, and so may justify the private investment in completely redesigning a product or system to accommodate the new feature, or to overcome the barriers that an incumbent has erected through secrecy or patent protection. But the social value of the added features for consumers may be much smaller than the private benefits of a successful attack on the incumbent's market position.
Common pool problems arise because individual competitors may vie for market position based upon R&D without taking into account the effect of their entry on the expected returns on the investments that others are making. Not every entrant will get a prize, but every entrant can believe in having just as good a chance, if not a better chance, for success than the others. The result can be duplicative investment in areas in which the anticipated prizes are large. The rivalries for certain prescription drug markets in the pharmaceutical industry often are cited as a classic manifestation of this problem: billions are spent to develop the next blockbuster therapy, whereas little investment may be devoted to products of lesser commercial value.
Racing behavior is another form of duplicative investment and is driven by the desire to beat one's rival to market. The value of being a week earlier at the patent office window, or 6 months in advance of competitors to launch a new software application, can be very large in comparison with the incremental social value of letting consumers use the innovation that much sooner. Firms then have an incentive to structure their R&D programs for speed, rather than cost minimization. They try to
bluff the opposition into quitting by establishing a lead and displaying a commitment to maintain it, whatever the cost.
It is clear that such effects, like the appropriability problem, will lead to inefficiencies in the detailed allocation of private-sector research outlays: excess correlation of research strategies, excessively duplicative funding in some areas, and inattention to other areas in which the marginal social value of new technologies may be quite high. What is less clear is whether these tendencies to overinvestment are so powerful that they destroy the presumption that private markets will, on balance, fail to allocate enough to creating new scientific and engineering knowledge. Some recent analytical work suggests this is not the case—except in circumstances where the real interest rate is so high that the value of knowledge spillovers to future generations should, in fact, be heavily discounted by the present generation, and where the impact of additional R&D funding on the creation of knowledge is rather weak (Jones and Williams, 1996).
Thus, the accrued wisdom from the economics profession regarding the aggregate tendency to underinvestment, and the corresponding case for government support of research as a stimulus to economic growth, still stands. But these qualifications point to the need for greater attention to where the publicly funded research is to be directed.
The Benefits of Public Support of Research
The development of scientific and technological knowledge is a cumulative process, one that depends on the prompt disclosure of new findings so that they can be tested and, if confirmed, integrated with other bodies of reliable knowledge. In this way open science promotes the rapid generation of further discoveries and inventions, as well as wider practical exploitation of additions to the stock of knowledge.
The economic case for public funding of what is commonly referred to as basic research rests mainly on that insight, and on the observation that business firms are bound to be considerably discouraged by the greater uncertainties surrounding investment in fundamental, exploratory inquiries (compared to commercially targeted R&D), as well as by the difficulties of forecasting when and how such outlays will generate a satisfactory rate of return.
The proposition at issue here is quantitative, not qualitative. One cannot adequately answer the question ''Will there be enough?" merely by saying, "There will be some." Economists do not claim that without public patronage (or intellectual property protection), basic research will cease entirely. Rather, their analysis holds that there will not be enough basic research—not as much as would be carried out were individual businesses (like society as a whole) able to anticipate capturing all the
benefits of this form of investment. Therefore, no conflict exists between this theoretical analysis and the observation that R&D-intensive companies do indeed fund some exploratory research into fundamental questions. Their motives for this range from developing a capability to monitor progress at the frontiers of science, to identifying ideas for potential lines of innovation that may be emerging from the research of others, to being better positioned to penetrate the secrets of their rivals' technological practices (Nelson, 1990).
Nevertheless, funding research is a long-term strategy, and therefore sensitive to commercial pressures to shift research resources toward advancing existing product development and improving existing processes, rather than searching for future technological options. Large organizations that are less asset constrained, and of course the public sector, are better able to take on the job of pushing the frontiers of science and technology. Considerations of these kinds are important in addressing the issue of how to find the optimal balance for the national research effort between secrecy and disclosure of scientific and engineering information, as well as in trying to adjust the mix of exploratory and applications-driven projects in the national research portfolio.
Direct Contributions to the Scientific Knowledge Base
When asked to demonstrate the usefulness of exploratory research that is undertaken to discover new phenomena, or explain fundamental properties of physical systems, scientists often point to discoveries and inventions generated by research projects that turned out to have immediate economic value. Many important advances in instrumentation, and generic techniques such as the polymerase chain reaction (PCR) and the use of restriction enzymes in gene-splicing, are such examples. These by-products of the open-ended search for basic scientific understanding also might be viewed as contributing to the knowledge infrastructure required for efficient R&D that might result in exploitable commercial innovations. Occasionally, such new additions to the stock of scientific knowledge are of immediate commercial value and yield major economic payoffs. Though few and far between, they can have far-reaching consequences.
There is no dearth of examples testifying to the practical value and commercial benefits that have followed serendipitously from exploratory, or curiosity-driven, scientific inquiries. The chance finding of bacteria surviving in and near the thermal vents in Yellowstone Park may be offered as a striking recent instance of a scientific discovery having an important and economically valuable field of application that hardly could be anticipated. The bacteria in question turned out to be crucial in the development of the PCR process for replication of specific pieces of DNA,
a generic technique that is now the basis of many commercial biotechnology applications, ranging from diagnostic kits to forensic medicine. What the developers of PCR required was an enzyme that would be stable at high temperatures, and the Yellowstone bacteria produced just what was needed.
The experience of the 20th century also testifies to the many contributions of practical value that trace their origins to large, government-funded research projects that were focused upon developing new enabling technologies for public-mission agencies (Rosenberg, 1987). Consider just a few recent examples from the enormous and diverse range that could be noted in this connection: airline reservation systems, packet switching and the Internet communication protocols, the Global Positioning System, and computer simulation methods for visualization of molecular structures.
At issue is whether a more directed search for the solutions to these applied problems would have been less costly and more expedient than waiting for scientists with quite different purposes in mind to come up with these commercially useful findings. Indeed, the theme of such spinoff stories is their unpredictability. The argument that the new applications are in some sense free requires that the research program to which they were incidental was worth undertaking for its own sake, so that whatever else might be yielded as by-products was a net addition to the benefits derived. Yet, the reason those examples are being cited is the skepticism as to whether the knowledge that was being sought by exploratory science was worth the cost of the public support it required. Perhaps this is why the many examples of this kind that scientists have brought forward seem never enough to satisfy the questioners.
The discovery and invention of commercially valuable products and processes are seen from the viewpoint of the new economics of science4 to be among the rarer of the predictably useful results that flow from the conduct of exploratory, open science. Without denying that research sometimes yields immediate applications around which profitable businesses spring up, it can be argued that those direct fruits of knowledge are not where the quantitatively important economic payoffs from basic science are to be found. Much more critical over the long run than spinoffs from basic science programs are their cumulative indirect effects in raising the rate of return on proprietary R&D performed by business firms. Among those indirect consequences, attention should be directed not only to informational spillovers, but to a range of complementary "externalities" that are generated for the private sector by publicly funded activities in the sphere of open science, where research and training are tightly coupled.
Indirect Effects of Government-Sponsored Research
Federally funded R&D provides a number of indirect benefits to private R&D beyond direct transfers of knowledge. These include intellectual assistance that can guide private R&D programs toward potentially more productive areas of inquiry and assistance in training researchers. Although resources are limited, and research conducted in one field and in one organizational mode is therefore performed at the expense of other kinds of R&D, exploratory science and academic engineering research activities support commercially oriented and mission-directed research that generates new production technologies and products. As such, public support of research in many ways complements, rather than competes with, private R&D efforts.
First among the sources of this complementary relationship is the intellectual assistance that fundamental scientific knowledge (even that deriving from contributions made long ago) provides to applied researchers—whether in the public or private sector. From the expanding knowledge base it is possible to derive time-and cost-saving guidance as to how best to proceed in searching for ways to achieve some prespecified technical objectives. Sometimes this takes the form of reasonably reliable guidance as to where to look first, and much of the time it takes the form of valuable instructions as to where it will be useless to look. One effect this has is to raise the expected rates of return and reduce the riskiness of investing in applied R&D. Gerald Holton, a physicist and historian of science at Harvard University, recently has remarked that if intellectual property laws required all photoelectric devices to display a label describing their origins, "it would list prominently: 'Einstein, Annalen der Physik 17 (1905), pp. 132-148.'" Such credits to Einstein also would have to be placed on many other practical devices, including all lasers.
The central point that must be emphasized here is that, over the long run, the fundamental knowledge and practical techniques developed in the pursuit of basic science serve to keep applied R&D as profitable an investment for the firms in many industries as it has proved to be, especially during the past half-century. In this role, modern science continues in the tradition of the precious, if sometimes imprecise, maps that guided parties of exploration in earlier eras of discovery, and in that of the geological surveys that are still of such value to prospectors searching for buried mineral wealth.
Research as Training
A second and no less important source of the complementary relationship between public and private research is the nexus between university research and training. The profitability of corporate R&D is closely tied to the quality of the young researchers who are available for employment. Seen from this angle, government funding of open exploratory science in the universities today is subsidizing the R&D performed by the private business sector. Properly equipped research universities have turned out to be the sites of choice for training the most creative and most competent young scientists and engineers, as many a corporate director of research well knows. This is why graduates and postdoctoral students in those fields are sent or find their own way to university laboratories in the United States. It explains why businesses participate in (and sponsor) industrial affiliates programs at research universities. It also is part of the reason for U.S. industrial research corporations' broadly protective stance in regard to the federal budget for scientific research. Acknowledgment of it has had a great deal to do with the recent announcement by the Japanese government of a dramatic reversal of its former policies and the initiation of a vast program of support for university-based research.
A key point deserving emphasis in this connection is that a great deal of the scientific expertise available to a society at any point in time remains tacit, rather than being fully available in codified form and accessible in archival publications. It is embodied in the knowledge of the researchers about such things as the procedures for culturing specific cell lines, or building a new kind of laser that has yet to become a standard part of laboratory repertoire. This is research knowledge, much of it very technological in nature—in that it pertains to how phenomena have been generated and observed in particular, localized, experimental contexts—that is embodied in people. Under sufficiently strong incentives it would be possible to express more of this knowledge in forms that would make it easier to transmit, and eventually that is likely to happen. But, being possessed by individuals who have an interest in capturing some of the value of the expertise they have acquired, this tacit knowledge is transmitted typically through personal consultations, demonstrations, and the movement of people among institutions.
The circulation of postdoctoral students among university research laboratories, between universities and specialized research institutes, and, no less important, the movement of newly trained researchers from the academy into industrial research organizations, are therefore important aspects of technology transfer—diffusing the latest techniques of science and engineering research. The incentive in this mode of transfer is a very powerful one for ensuring that the knowledge will be successfully trans-
lated into practice in the new location, for the individuals involved are unlikely to be rewarded if they are not able to enhance the research capabilities of the organization into which they move.
A similarly potent incentive may exist when a fundamental research project sends its personnel to work with an industrial supplier from which critical components for an experimental apparatus are being procured. Ensuring that the vendor acquires the technical competence to produce reliable equipment within the budget specifications is directly aligned with the interests of both the research project and the business enterprise. Quite obviously, the effectiveness of this particular form of user-supplier interaction is likely to vary directly with the commercial value of the procurement contracts and the expected duration and continuity of the research program.
For this reason, big science projects or long-running public research programs may offer particular advantages for the collaborative mode of technology transfers, just as major industrial producers—such as the large automotive companies in Japan—are seen to be able to set manufacturing standards and provide the necessary technical expertise to enable their suppliers to meet them. By contrast, the transfer of technology by licensing intellectual property is, in the case of process technologies, far more subject to tensions and deficiencies arising from the absence of complete alignment of the interests of the involved individuals and organizations. But, as has been seen, the latter is only one among the economic drawbacks of depending upon the use of intellectual property to transfer knowledge from nonprofit research organizations to firms in the private sector.