Increasing the productivity of the American workforce is the key to higher living standards and stronger economic growth in the future. Evidence indicates that investments in research and development (R&D) have large payoffs in terms of growth. R&D yields new products, improving the quality of life, and new processes, enabling American firms to reduce costs of production and become more competitive. Indeed, investments in R&D are estimated to account for half or more of the increase in output per person. Maintaining or increasing this country's R&D effort is essential if we are to increase the rate of productivity growth and improve American living standards.
The largest part of R&D in the United States is funded by private industry. Small entrepreneurs see an opportunity, raise funds any way they can, and take their chances on an innovative idea. Large companies spend billions on R&D labs to develop a stream of new products and processes. Private companies know the markets they serve and the workers who must produce the products. Risking their own funds gives them a strong incentive to avoid costly failures.
Since the founding of this country, the Federal government has had an important role in the promotion of science and technology. Indeed, the Constitution gave Congress the right to grant patents to "promote the progress of science...." But in today's complex and competitive world economy, promoting the progress of science goes beyond simply the granting of patents. First, successful R&D in private companies depends upon the flow of new ideas and trained people stemming from basic research and pre- commercial R&D. Federal support for these activities is vital. Second, the Federal government sponsors much applied research to improve its own capabilities in such areas as national security, health, and transportation. The government can then help transfer technologies developed for its own use to the private sector.
This paper describes U.S. expenditures on R&D, how they have been changing over time, and how they compare with other countries. It then examines the rationale and role for government involvement in R&D and documents the high returns to R&D investments. Finally, it projects the results of the Congressional budget resolution on R&D expenditures and contrasts that projection with Japanese plans.
An alternative comparison is R&D expenditures as a percentage of GDP, in order to account for differences in the size of economies. Using this comparison, the United States is just behind Japan and slightly ahead of (unified) Germany and France (see Chart 2).
But this does not really tell the whole story. We must look not only at how much we spend, but also at what we spend it on. Aggregate R&D expenditures can be broken down into defense and non-defense R&D expenditures. The United States falls behind Germany, even further behind Japan, and remains just ahead of France in terms of non-defense R&D expenditures (see Chart 3).
As seen in Chart 4, the United States consistently has lagged behind in this measure over the past two decades.
Although total expenditures on non-defense R&D have remained relatively constant as a share of GDP in the last 10 years (at a level well below those of Germany and Japan), Federal expenditures on non-defense R&D in the United States actually have fallen as a percentage of GDP over the last three decades (see Chart 5).
In the United States in 1994, the Federal government provided approximately 36 percent of all R&D funds and industry provided about 59 percent, with the balance coming from universities and colleges and other non-profit organizations. Industry primarily funds product-related applied research and development, as these areas are most likely to yield immediate payoffs. Government funds most basic research, since the results of this type of research are the most uncertain and applications may not be realized for quite some time, as well as more than one-third of all applied research. Table 1 details the breakdown of support for different types of research.
|Universities & Colleges||5.3||3||10||4||*|
Examples abound. Lasers and transistors are now a part of everyday life. The inventors of the laser probably had no idea that it would eventually be used for removing cataracts or for playing music in a compact disc player. Likewise, the American physicists who invented the transistor at Bell labs in 1948 could not have imagined that their invention would be used today in radios, computers, spaceflight and guided missiles, and countless other electronic devices. In both cases, even if the inventors' imaginations did reach such heights, today they receive no additional monetary benefit for the large advantages that society reaps from their insights.
Sometimes the spillovers are far more subtle. The discovery of nylon showed that it was possible to create artificial fibers with remarkable properties -- and this knowledge affected the direction of research efforts applied by thousands of other researchers.
The consequences of the existence of important spillovers is that private firms will not invest enough in R&D from a national perspective. This point is not merely theoretical: many studies have demonstrated that investments in R&D yield high returns to investors and even higher returns to society. One recent review of econometric studies concluded that the average private rate of return to an innovation seems to be between 20 and 30 percent, while the social rate of return is closer to 50 percent.[10,11] An earlier, extensive, case-study approach found that the median private return to the innovations studied was 25 percent, while the median social rate of return was 56 percent. While estimates of the rates of return are just that -- estimates -- a wealth of studies over the past two decades have confirmed these high private returns and even higher social returns. Table 2 highlights the results of some of these studies.
|Author (year)||Estimated Rates of Return|
|Nadiri (1993)||20 - 30||50|
|Terleckyj (1974)||29||48 - 78|
|Sveikauskas (1981)||7 - 25||50|
|Bernstein-Nadiri (1988)||10 - 27||11 - 111|
|Scherer (1982, 1984)||29 - 43||64 - 147|
|Bernstein-Nadiri (1991)||15 - 28||20 - 110|
In addition, some firms -- especially small ones that lack funds -- may not invest enough in R&D even from their own perspective. To make R&D investments, a firm may need to go to capital markets for funding, and to provide these funds, financiers must have sufficient information to be able to assess the risks of the investments. Firms may not want to provide this information for fear of losing future private gains if somebody else were to use that information. Moreover, R&D cannot be collateralized, in the way that an investment in a building or a machine can be. Thus, the firm must either pay higher interest rates for loans or use its own funds to pay for the research. In fact, evidence suggests that small firms' investments in R&D are limited by their internal cash-flow.
The inadequacy of firms' incentives to invest in R&D creates an important role for the Federal government. The goal of technology policy, however, is not to substitute the government's judgment for that of private industry. Rather, the point is to correct a genuine and significant problem -- underinvestment in basic research and in pre-commercial R&D resulting from the divergence between private and social returns to those activities. A complementary goal is to design the technology investments that the government itself makes in public goods -- national security, public health, education, a clean environment, an efficient transportation system -- in ways that maximize the potential external benefits for the Nation's commercial technology base. In both cases, support for technological innovation enhances the Nation's economic and social welfare.
Expanding the R&D tax credit provides an additional incentive to the private sector to ameliorate the underinvestment problem discussed in this paper. Indeed, the tax credit can be effective in increasing private sector R&D expenditures, and is an important component of a comprehensive technology policy.
While the tax credit is important in promoting increased R&D expenditures, alone it is not sufficient. As a recent Congressional study noted, the tax credit does not alter the composition of R&D expenditures. It is not designed to encourage research in areas subject to particularly severe underinvestment problems, which include basic and pre-commercial research.
What can the government do? The Federal government has a long history of involvement in science and technology. For example, in 1842 the government appropriated $30,000 for Samuel Morse to build a telegraph line from Washington to Baltimore to demonstrate the feasibility of his new technology. In 1862 the Federal government passed the Morrill Act, which gave states land with which to establish land-grant colleges to teach agriculture and the mechanical arts. Government also has a long history of involvement in direct funding of agricultural research dating back to the nineteenth century: many studies over the past 30 years have found rates of return to public investments in agricultural research of over 35 percent.
The fact that government has a role in promoting science and technology clearly has long been recognized. The earliest and most widely used government incentive for encouraging innovation is the granting of patents, which essentially gives an innovator temporary monopoly rights on a new product or process. While important, patents alone are not a solution to the underinvestment problem. Even with strong patent protection, inventors capture only a small fraction of the benefits to society that accrue from their innovations, so that they will still underinvest. Underinvestment will be particularly severe for R&D with large spillovers and for research that yields results only far in the future or is extremely risky.
Investments in R&D are inherently risky, and some government-supported explorations, like those in the private sector, will be unsuccessful. Successful R&D investments -- from the jet engine to transistors to lasers -- can and have changed the whole economy. Government support was crucial in areas such as computers and integrated circuits, jet engines and airframes, and biotechnology and medical equipment. The result has been entire fields of productive wealth-enhancing, job-creating economic activity.
R&D provides the basis of America's competitive advantage in the many sectors in which the United States leads the world. Our strength, reflected in the large number of Nobel Prize winners in science -- most of whom have received government support -- is based on our research universities, the best in the world, all of which depend in large measure on government support. Students come from all over the world to learn from U.S. scientists and engineers.
Funding basic research. Most people recognize the need for government funding of basic, or fundamental, research. Indeed, as shown earlier in Table 1, the Federal government funds close to 60 percent of all basic research. Basic research is, by definition, not directed at solving an immediate problem or at inventing a particular product. While basic research has immediate returns in adding to our knowledge base and in educating scientists and engineers, economic returns from investments in basic research may be many years away, and may not have applications bearing any similarity to what the researcher originally thought. Since so much of the returns to basic research are not appropriated by the innovator (and indeed, in many cases, the output of basic research is not patentable), the gap between social and private returns is particularly large, and therefore the problem of underinvestment is particularly severe. Firms are typically reluctant to invest much in basic research.
Basic research ultimately can yield extraordinary returns to society. For example, two physicists in 1946 discovered nuclear magnetic resonance as the result of basic research. While they had no idea how this knowledge would eventually be used, others soon realized the potential applications of this knowledge. Today, most major hospitals have magnetic resonance imaging (MRI) machines for use in noninvasive scanning of patients' internal organs. The MRI is a direct outgrowth of earlier basic research.
Universities and colleges comprise the largest single group of performers of basic research, accounting for approximately 45 percent of all basic research in 1994. This research is funded primarily by the Federal government. Universities and colleges create "knowledge for knowledge's sake," help develop an educated population, and train the scientific and engineering workforce. However, academic research itself also plays a crucial role in industrial innovation. One recent study of 76 manufacturing firms revealed that these firms could not have developed about 11 percent of their new products and 9 percent of their new processes without research done at universities and colleges. This study estimated the median social rate of return to research done at academic institutions to be 28 percent.[21,22]
Another study notes that it is difficult to assign a particular rate of return to basic research, since its results may be used in many diverse ways. Instead, it suggests that basic research should be viewed as an input into applied research in many areas. Basic and applied research interact in many ways, increasing the productivity of both. In fact, one study of manufacturing firms found a correlation between increased spending on basic research and increased firm productivity, which may reflect increased effectiveness of a firm's applied R&D when the firm also conducts basic research.
Pre-Commercial R&D: The Changing Government Role. The government's role does not end with funding basic research. One can view R&D as a continuum, with basic research at one end, facing a huge underinvestment problem requiring substantial government involvement, and product commercialization at the other end, where most returns go directly to the firm. Pre-commercial R&D is somewhere in between these two extremes. Some types of pre-commercial research may be extremely risky or have an especially large gap between private and social returns. Government support of such pre-commercial R&D involves identifying, with the aid of scientists, engineers, entrepreneurs, economists, and business people, technologies that could yield large societal benefits but may not necessarily yield much private return to the innovator. It is this belief that drives the Administration's technology policies.
In fact, the United States implicitly began following a similar technology policy after the Second World War. The Second World War brought great technological advancements from government research, all in the name of the war effort. Many of those technological accomplishments had applications in civilian life, as well. President Franklin Roosevelt recognized the potential of the R&D machine that had been built up during the war, and requested that Vannevar Bush, director of the wartime Office of Scientific Research and Development, devise plans on how to use the wartime experience in peacetime. In response to President Roosevelt's request, Bush authored Science: The Endless Frontier in 1945, which became the guiding document for much of U.S. postwar science policy.
The United States channeled public investment into basic research at universities and government laboratories, then supported the initial application of the results in products and production processes procured by public agencies. New technologies first developed for (and procured by) the Department of Defense, the Department of Energy, or the National Aeronautics and Space Administration, or supported by the National Science Foundation or the National Institutes of Health, would then diffuse, or "spin off," into commercial use. In this manner, the Federal government supported the development and diffusion of jet aircraft and engines, semiconductor microelectronics, computers and computer-controlled machine tools, pharmaceuticals and biotechnology, advanced energy and environmental technologies, advanced materials, and a host of other commercially successful technologies.
This system worked well as long as military requirements represented the leading-edge applications of new industrial technologies. In many areas of basic research supported outside the defense establishment, including biomedical research and the development of pharmaceuticals, biotechnology, and medical diagnostic devices, the system continues to work well.
The circumstances that allowed the United States to rely primarily on a defense-led model have changed. With the end of the Cold War, demand for new defense systems is now less than it was. Commercial product spin-offs from military research have also diminished from their heyday of the 1950s and 1960s, and American companies face intense international competition from increasingly capable foreign firms. On the other hand, these changes also create exciting new opportunities: innovative defense technologies are now more likely to emerge first in commercial products and production techniques, and American companies are taking advantage of expanded opportunities in foreign markets. Accordingly, the Administration's technology initiatives are shifting the composition of Federal R&D from military to civilian concerns, and the composition of military R&D toward the development of so-called dual-use technologies -- those with applications to both military and commercial products.
Designing a successful program of technology support. The Administration's efforts to promote innovative technology contain design features meant to limit the possibility of government failure in the implementation of technology policy: in most cases, firms participating in the Administration's programs must cover at least 50 percent of the costs of the project; projects are initiated by private firms, which compete for limited funding; outside experts in the relevant scientific, technological, and economic fields evaluate competing proposals; and firms can compete for funds in a wide array of technological fields, to ensure that support for pre-commercial R&D support does not get "captured" by any particular technology or set of firms.
Even the best-designed technology program will have failures. Indeed, if it does not, then it certainly is too cautious. In the final analysis, the returns to government- funded R&D depend upon the returns to the successful projects outweighing the losses from the unsuccessful ones. By incorporating the above design features, the Administration's technology program provides the best chance for achieving high returns that benefit American living standards.
Returns to government R&D investments: It is impossible to provide a reliable quantitative estimate of the returns to publicly-supported R&D based upon historical data, primarily because such a large percentage of Federal R&D support has been defense-related, although as noted earlier the returns to other public investments have been enormous. Traditional ways of calculating private returns to R&D do not apply in situations where the government funds the R&D and then purchases the resulting output.
The real impact of government-supported R&D is not the returns to the individuals involved in the research, but the returns to society. Measuring such returns is not a simple task, since the results of public R&D weave their way through the economy in countless directions. Researchers have noted that because of such spillovers, one must examine Federal research on a case-by-case basis. Some government programs have been spectacular successes, yielding enormous social returns. The aircraft industry is a prime example. The development of the U.S. aerospace industry was largely government-funded. As late as 1986, close to 80 percent of all R&D in this industry was Federally-supported. Today this industry is a large employer and one of the largest exporters in the nation.
Other examples include:
By contrast, the Japanese government recently announced plans to double its R&D spending by the year 2000. Chart 7 highlights the effect of the Congressional plan and the Japanese plan: by 1997 Japan will overtake the United States in government support of non-defense R&D -- in total dollars, not just as a share of GDP.
Cutting Federal R&D will reduce private R&D expenditures. Many studies demonstrate that Federal spending on R&D stimulates additional private spending on R&D. This complementarity holds up in basic, as well as applied, research. In other words, an additional dollar of Federal R&D expenditures adds more than a dollar of R&D investment to the economy.
Unfortunately, complementarity also suggests that if the Federal government cuts R&D expenditures, the private sector will cut R&D expenditures, as well. Chart 8 shows a clear correlation between changes in Federal R&D expenditures and changes in private R&D expenditures one year later.
This correlation means that if Federal R&D support is cut, the nation is likely to lose future rewards not only from the Federally-supported R&D that will not be undertaken, but also from the industrial R&D that will not be undertaken as the private sector scales back in response to Federal cuts.
The competitive position of the United States -- and indeed future increases in standards of living -- depends on technological advances. These in turn depend on our entire scientific and technological infrastructure, which includes our educational institutions -- producing the scientists and engineers that will provide the creative advances of the future -- our research universities, and our nation's laboratories, both within the private and public sectors. Ideas flow from basic research, through pre-competitive development, to concrete applications, producing new products and developing new, better, and lower-cost production processes. Government has a vital role in sustaining this infrastructure -- from supporting scientists and engineers, to promoting basic research, to assisting in the development of new, high-risk technologies with significant spillovers. We have evolved an effective system that has led America to its current pre-eminent role. Changes in our world necessitate that this system, and the role of government, continue to evolve. Now is the time to renew our commitment to these advances and to continuing the adaptation of our system to the changing world. These are high-return investments that will provide the basis of the America of the twenty-first century.
 Baily, M.N. and A. Chakrabarti, Innovation and the Productivity Crisis, Brookings Institution, Washington, DC 1988.
 Pre-commercial R&D may be loosely defined as R&D that is close to yielding a new product or process, but is still far enough away from commercialization to require a firm to take on substantial risk in pushing it towards the market, and may be such that the social returns to the investment will be much higher than the private returns.
 Industry also relies on the government to support the technical infrastructure -- for example, standards for weights and measures. Research in this area is essential for advancing commerce and trade.
 Data for Charts 1 through 5 and Table 1, unless otherwise noted, are from the National Science Foundation, National Patterns of R&D Resources: An SRS Special Report. NSF 95- 304. 1995.
 Defense and non-defense expenditures: 1961 - 1979: Office of Management and Budget: Budget of the United States Government: Historical Tables. Fiscal Year 1996; 1979 - 1994: National Science Foundation, 1995; GDP figures from Council of Economic Advisers, Economic Report of the President. 1995.
 National Science Foundation. 1995.
 NSF. 1995.
 The chain from idea to usable product or process can be long. R&D is comprised, most generally, of basic research, applied research, and development. The divisions between these areas is not always clear, as they all interact in complex ways, with advances in one type of research influencing the direction of research in others. For conceptual purposes, though, The National Science Foundation (Science and Engineering Indicators, 1993) defines these terms as follows:
 Nadiri, Ishaq. "Innovations and Technological Spillovers." NBER Working Paper Series. Working Paper No. 4423. August, 1993.
 Rates of return can be estimated by computing the benefits (including discounted future benefits) and the costs of the innovation.
 Mansfield, Edwin, J. Rapoport, A. Romeo, S. Wagner, and G. Beardsley. "Social and Private Rates of Return from Industrial Innovations." Quarterly Journal of Economics. Vol. 77, pp. 221 - 240. 1977.
 Some of the studies cited in Table 2 look at industry- level data while others use a case-study approach. In some instances what is listed as a "social" rate of return is actually an indirect return to one industry resulting from the research of another industry. The point is clear: private rates of return to R&D are high, and the returns to society are even higher. The studies in the table are as follows (in addition to those already cited):
 Table adapted from: Griliches (1992), and Nadiri (1993).
 Himmelberg, Charles and Bruce Petersen. "R&D and Internal Finance. A Panel Study of Small Firms in High-Tech Industries." Review of Economics and Statistics. Vol. 76, Issue 1. pp. 38 - 51. 1994.
 The R&D tax credit, officially known as the research and experimentation (R&E) tax credit, allows firms to deduct from their income taxes a portion of their R&D expenditures beyond a certain base level.
 Office of Technology Assessment, Congress of the United States. "The Effectiveness of Research and Experimentation Tax Credits." September 20, 1995.
 National Research Council. Colleges of Agriculture at the Land Grant Universities: A Profile. National Academy Press. Washington, DC. 1995.
 USDA Economic Research Service. "The Value and Role of Public Investment in Agricultural Research." Staff Paper Number 9510. May, 1995.
 Universities and colleges actually performed close to 55 percent of all basic research when one includes work done at Federally funded Research and Development Centers located at universities and colleges.
 While the "28 percent" figure is clearly a rough estimate, it shows that the returns to academic research are high. Moreover, this estimate is likely to be too low for two reasons. First, the study used academic research done only in the 15 years prior to the innovation -- much academic research may not be used in industrial innovations until more than 15 years after the initial discovery or publication, or may continue to be used for many years thereafter. Second, the study examined only seven industries. The academic research useful for innovations in these industries likely was useful in other industries, as well. Clearly, investing in academic research is an area with high payoffs.
 Mansfield, Edwin. "Academic Research and Industrial Innovation." Research Policy. Vol. 20. pp. 1 - 12. 1991.
 David, Paul, et al. "Analysing the Economic Payoffs From Basic Research." Economic Innovations and New Technology. Vol. 2, pp. 73-90. 1992.
 Mansfield, Edwin. "Basic Research and Productivity Increase in Manufacturing." American Economic Review. Vol. 70, No. 5. pp. 863-873. December, 1980.
 In 1987, for example, about 70 percent of all Federal research was defense-oriented. Products resulting from defense R&D generally are purchased by the government and are not subject to a market test. See Hall, Bronwyn. "The Private and Social Returns to Research and Development: What Have We Learned." June, 1995 for a discussion of the difficulty of measuring the return to public R&D. One study of manufacturing firms found that increased government funding for applied research is correlated with increased productivity (Mansfield, Edwin. "Basic Research and Productivity Increase in Manufacturing." American Economic Review. Vol. 70, No 5. December, 1980. PP 863-873). The unique feature of this study is that because it was not actually focused on government funding, the results were based on firms not necessarily involved in defense contracting.
 Bartelsman, Eric. "Federally Sponsored R&D and Productivity Growth." Finance and Economics Discussion Series, No 121. Federal Reserve Board, Washington, DC. April, 1990.
 Some programs that do not meet their specified goal may officially be classified as failures. However, even some "failed" projects can yield enormous positive spillovers.
 Mowery, David and Nathan Rosenberg. Technology and the Pursuit of Economic Growth. Cambridge University Press. 1989.
 In 1994 the industry employed about 480,000 people. From 1990 to 1994, exports averaged over $30 billion per year.
 1990 - 1995 are actual expenditures; 1996 - 2002 are estimated results of Congressional proposals; deflators 1994 - 2000 are estimates from OMB, Analytical Perspectives: Budget of the United States Government. FY 1996. Assumed 3.5 percent inflation from 2000 - 2002.
 Levy, David and Nestor Terleckyj. "Effects of Government R&D on Private R&D Investment and Productivity: A Macroeconomic Analysis." The Bell Journal of Economics. Vol. 14, No. 2. pp. 551 - 561. Autumn, 1983.
 Robson, Martin. "Federal Funding and the Level of Private Expenditure on Basic Research." Southern Economic Journal. Vol. 60, No 1. pp. 63 - 71. July, 1993.
 Hill, Christopher. "Private Funds are Unlikely to Replace Cuts in Public Funds for R&D in the U.S." Mimeo. June 19, 1995. Data from NSF. "National Patterns of R&D Resources: 1992." NSF 92-330. October, 1992.