The End of Research?
Roger A. Pielke, Jr.
Understanding the implications of slowing growth or even a future decline in federal funding for R&D requires understanding federal funding on science and technology. A number of recent statements by prominent scientists and policy makers suggest confusion about trends in federal support for R&D, and that a primer now makes sense.
Examples of confusion include a July 15, 2004 statement by Shirley Ann Jackson, President of Rensselaer Polytechnic Institute and President of the American Association for the Advancement of Science (AAAS) that “The Federal investment in research, measured as a share of the Gross Domestic Product (GDP), has declined by almost two-thirds since the 1980s.” (3) And presidential candidate John Kerry included in a June, 2004 white paper on science and technology an observation that “[G]overnment support for many key disciplines of science and engineering, particularly the physical sciences and engineering, has been declining.” (4) The problem with both statements is that they are factually incorrect. Research as a fraction of GDP has not declined by two thirds and most areas of science and engineering are currently receiving more funds than ever before. If science policies are to be informed by on a comprehensive picture of trends in support for R&D, then it makes sense that those trends are well understood and interpreted.
This perspective describes the sources of data on federal funding for R&D, what is measured, what the data show, and some implications for science policy and politics.
Because the United States does not allocate a lump sum for R&D activities of the federal government, Congress mandated that the National Science Foundation (NSF) “provide a central clearinghouse for the collection, interpretation, and analysis of data on scientific and engineering resources and to provide a source of information for policy formulation by other agencies of the Federal Government.” In addition, the AAAS (publisher of Science) and the Organization for Economic Cooperation and Development (OECD) also provide related estimates of United States R&D. Understanding federal funding for science and technology requires that budgets be compiled from the various budgets of agencies that support science and technology.
There are currently four different compilations of data: by OECD, AAAS, and two by the NSF Division of Science Resources Statistics. OECD’s collection focuses on R&D in the public and private sectors and depends upon official government sources for its U.S. government estimates (5).
One set of NSF data focuses on tallying budget requests from agencies to the Office of Management and Budget (OMB) in the Executive Office of the President, which has responsibility each year for compiling the President’s Budget. According to NSF the OMB compilation focuses on 23 agencies that are responsible for 99% of all federal research and development (6). NSF contracts to the AAAS to collect this data. The funding requested by agencies may or may not all be spent in a single year, but instead over several years. Hence, a measure of budget requests will lead to a different total than actual expenditures.
NSF oversees the collection of a
second set of research and development funding data by contracting with
a consulting firm to collect, via a survey, the research and development
expenditures by 29 federal agencies and 73 of their subdivisions. (7)
The NSF survey data focuses on the actual expenditure of funds, which
may represent Congressional appropriations over more than a single year.
What to measure?
Of course, any taxonomy implies an underlying structure and the compilation of R&D data is no different. NSF describes research as “basic” or “applied” as follows: “In basic research, the objective of the sponsoring agency is to gain fuller knowledge or understanding of the fundamental aspects of phenomena and of observable facts without specific applications toward processes or products in mind. In applied research, the objective of the sponsoring agency is to gain knowledge or understanding necessary for determining means by which a recognized and specific need may be met.” This taxonomy institutionalizes the “linear model” of science where basic research leads into applied research, to development, and then to societal benefits. Even though NSF admits and analysts have frequently demonstrated (e.g., 8) that the basic-applied distinction has little meaning in reality and fails to accurately characterize how research is actually related to societal benefits, it nonetheless continues to use the framework to describe R&D activities.
The basic-applied taxonomy leads to an additional complication in that Congress has granted only NSF and the National Aeronautics and Space Administration (NASA) a legislative mandate that includes conducting research for research sake. For all other agencies, research is an instrumental means for fulfilling the agency mission. Thus, “basic research” in an agency like Interior may be very different than “basic research” in NSF or NASA (9, 10).
Other taxonomies have been
proposed, but none has replaced the basic-applied taxonomy. For
example, in 1995 a committee of the National Research Council proposed a
metric call “Federal Science and Technology” (FS&T) funding which to
"generally favor academic institutions because of their flexibility and
inherent quality control, and because they directly link research to
education and training in science and engineering" (11). Perhaps
predictably, scientists in government labs objected to the measure as
creating a funding bias against non-university-based research (12).
First are the aggregate numbers.
Figure 1 shows aggregate funding for research and development in current
and constant FY 2004 dollars from 1982-2005. The inflation-adjusted
data show periods of increases and stasis, but no systematic,
significant decreases. The most recent decade saw rapidly accelerating
increases to record levels of funding.
Figure 1. The second golden age: Federal funding for R&D 1982-2005, Source: AAAS, 2005 is an estimate.
Figure 2. Trends in federal funding in comparison to GDP (right axis) and as portion of discretionary spending (left axis), Source: AAAS. Also shown is discretionary spending in comparison to GDP (left axis), Source: CBO.
The data show that over the past
decade science and technology have been in a second golden age.ii
While the first golden age was led by increases in physics and
engineering, the second golden age has been led by increases in health
and security research. Federal government expenditures for research and
development have increased dramatically over this period to record
levels, and have represented an ever-growing share of discretionary
government expenditures. The second golden age reflects both society’s
faith in science and technology as a source of societal benefits, but
also the prowess of the science and technology community in the annual
fray over finite government resources.
Of course, science policy should not be about simply “How much?” but “Why?” (14). However, the S&T community typically focuses narrowly on “how much?” using a three-part strategy to argue for more public sector resources.iii It claims crisis, even in times of plenty (15). It calls for balance, to limit intra-disciplinary, intra-agency debates over priorities (10). And it claims that societal benefits are proportional to funds invested; more funds are equated with more benefit (16). Current debate over federal funding for R&D remains far from James Sensennbrener’s (R-PA, former Chair of House Science Committee) desire that “[Science and technology] funding should be driven by policy and not the other way around.” (17)
A focus on aggregate funding, rather than the marginal benefits of adding or cutting funding for particular programs, may prove problematic as R&D funding all but certainly cannot continue to grow at the pace that it has over the past decade, regardless of who occupies the White House, making tough choices within the scientific community inevitable (2).
Consider the following, perhaps representative situation. In July, 2004 NASA decided to steer a research satellite from earth orbit into the ocean upon completion of its mission. Scientists and some weather forecasters appealed to Congress to overturn NASA’s plans for a controlled reentry of the Tropical Rainfall Measurement Mission (TRMM) arguing that the benefits of the satellite’s data to weather forecasters far exceed the risks of an uncontrolled reentry resulting from using the mission’s remaining fuel to extend TRMM’s on-orbit research mission. The decision was important not only because of the risks involved but because the decision has financial consequences for NASA, its TRMM follow on mission, and scientific research related to TRMM.
Ideally, decision makers in NASA and Congress would have had a clear understanding of the costs and benefits associated with its available decision options in order to inform their actions. But as it turned out, information is lacking on costs of TRMM, the benefits of TRMM data, and the risks of reentry (18). The lack of information means that recent decisions about the future of TRMM were based almost completely upon anecdotal information and political sway among participants. TRMM’s extension through 2004 was determined via science politics, not science policy.
The TRMM situation
reflects the consequences of the S&T community’s historical predilection
for failing to systematically evaluate the relative costs and benefits
of marginal investments in a particular project or area—focusing instead
on aggregate measures of federal R&D support (1). However, after a
decade of record increases it seems unlikely that claims of crisis,
balance, or proportional benefit will avoid intra-S&T conflicts
resulting from stagnant budgets. If the extended period of increases in
overall funding for S&T is indeed ending (2), then a continued focus on
the forests, rather than the trees, does not appear sustainable.
iData from the Congressional Budget Offfice, http://www.cbo.gov
many scientists the recent decade may not
like a golden age, simply because the number of researchers competing
for federal funds far exceeds that ratio in the 1960s (14).
(1) Office of Technology Assessment (OTA). 1991. Federally Funded Research: Decisions for a Decade. OTA-SET-490. Washington, DC: GPO.
(2) AAAS Analysis of the Outyear Projections for R&D in the FY 2005 Budget
April 22, 2004 (revised May 7), http://www.aaas.org/spp/rd/proj05p.htm AAAS R&D Budget and Policy Program, Directorate for Science & Policy Programs, American Association for the Advancement of Science, Washington, DC.
(5) OECD Main Science and Technology Indicators 2004, http://www.oecd.org/document/26/0,2340,en_2649_34269_1901082_1_1_1_1,00.html
(6) National Science Foundation, Division of Science Resources Statistics, Federal Funds for Research and Development: Fiscal Years 2001, 2002, and 2003, NSF 04-310, Project Officer, Ronald L. Meeks (Arlington, VA 2004).
(7) National Science Foundation, Division of Science Resources Statistics, Federal R&D Funding by Budget Function: Fiscal Years 2001-03, NSF 02-330, Project Officer, Ronald L. Meeks (Arlington, VA 2002).
(8) Stokes, D. E.. 1997. Pasteur's Quadrant: Basic Science and Technological Innovation, Brookings Institution Press, Washington, DC.
(9) E. Eiseman, K. Koizumi, and D. Fossum, 2002. Federal Investment in R&D, MR-1639.0-OSTP, Science and Technology Policy Institute, RAND Corporation, Washington, DC. http://www.rand.org/publications/MR/MR1639.0/
(10) S. Merrill and M. McGeary, 1999, Science, 285:1679-1680.
(11) National Research Council, 1995. Allocating Federal Funds for Science and Technology, Commission on Physical Sciences, Mathematics, and Applications, National Academy of Sciences, Washington, DC, http://www.nap.edu/books/0309053471/html/index.html.
(13) AAAS R&D Budget and Policy Program http://www.aaas.org/spp/rd/usg03.pdf
(14) Sarewitz, D. 2003. Does Science Policy Exist, and If So, Does it Matter?: Some Observations on the U.S. R&D Budget, Discussion Paper for Earth Institute Science, Technology, and Global Development Seminar, Center for Science, Policy, and Outcomes, http://www.cspo.org/products/papers/budget_seminar.pdf
(15) D. Greenberg, 2001. Science, Money, and Politics, University of Chicago Press.
(16) D. Sarewitz, 1996. Frontiers of Illusion, Temple University Press, Philadelphia, PA.
(18) Pielke, Jr., R. A., 2001: Report of a Workshop on Risk-Benefit Assessment of Observing System Decision Alternatives, National Aeronautics and Space Administration, June 2000, Boulder, CO. http://sciencepolicy.colorado.edu/homepages/roger_pielke/workshops/trmm/index.html