Does Science Policy Matter?
It would if we had a real science policy, but what we have now is science politics.
It is not only axiomatic but also true that federal science policy is largely played out as federal science budget policy. Science advocacy organizations such as the American Association for the Advancement of Science (AAAS), the National Academies, and various disciplinary professional societies carefully monitor the budget process and publish periodic assessments, while issue-focused interest groups such as disease lobbies and environmental organizations focus on agencies and programs of specific relevance to their constituencies. Overall, it is fair to say that marginal budgetary changes are treated by the science and technology (S&T) community as surrogates for the well-being of the science enterprise, while the interested public considers such changes to be surrogates for progress toward particular societal goals (for example, budget increases for cancer research mean more rapid progress toward cures). In this dominant science policy worldview, yearly budget increases mean that science is doing well, and doing good. When budgets are flat or declining—or even when rates of budget increases are slowing—then science must suffer and so, by extension, must the prospects for humanity.
In recent years, this worldview was perhaps most starkly on display in discussions about the National Institutes of Health (NIH), whose budget doubled between 1998 and 2003 as a result of a highly effective lobbying effort, a sympathetic Congress, and a brief period of overall budgetary surplus. During this period, the NIH budget went from $13.6 billion to $27.0 billion, and the NIH share of all civilian federally funded research rose from its already dominant 37% to 48%. Nevertheless, when the fiscal year (FY) 2004 budget debates began, NIH and its advocates in the research community portrayed the situation as one of crisis arising from a sudden decline in the rate of budget increase. Said a representative of the Association of American Medical Colleges: “Two or three years of 2 or 3% increases, and you’ve pretty much lost what you’ve gained . . . And you’ve certainly lost the morale of investigators who can’t help but be demoralized by trying to compete for funding under those circumstances.” In another notable example, the president-elect of the AAAS in 1990 solicited letters from 250 scientists and discovered that many were unhappy because they felt that they did not have enough funding. From this information he inferred an “impoverishment” of basic research even though, as he acknowledged, science funding had been growing steadily.
In this article I will argue that the annual obsession with marginal changes in the R&D budget tells us something important about the internal politics of science, but little, if anything, that’s useful about the health of the science enterprise as a whole. In particular, marginal budget changes give almost no information about the capacity of the science enterprise to contribute to the wide array of social goals that justifies society’s investment in science. I will return to this point later; first I will focus on the more parochial reality that the annual federal budget numbers for science cannot be understood unless they are placed in a broader political and historical context. A given year’s marginal budget increase says as much about the health of the science enterprise as the nutritional value of a single meal says about the health of one’s body.
One of the most astonishing aspects of science policy over the past 30 or so years is the consistency of R&D funding levels as a proportion of the discretionary budget (Figure 1). (Discretionary spending is the part of the budget that is subject to annual congressional decisions about spending levels.) Since the mid-1970s, nondefense R&D budgets have constituted between 10 and 12% of total nondefense discretionary spending. Total R&D (defense and non-defense) shows a similar stability at 13 to 14% of the total discretionary budget. This consistency tells us that marginal changes in the R&D budget are tightly coupled to trends in discretionary spending as a whole.
Given the Balkanized manner in which science budgets are determined, such stability at first blush may seem incomprehensible. After all, no capacity exists in the U.S. government to undertake centralized, strategic science policy planning across the gamut of federal R&D agencies and activities. The seat of U.S. science policy in the executive branch is the Office of Science and Technology Policy, whose director is the president’s science advisor. The influence of this position has waxed and waned (mostly waned) with time, but it has never been sufficient to exercise significant control over budgetary planning. That control sits with the Office of Management and Budget, which solicits budgetary needs from the many executive agencies that conduct R&D, negotiates with and among the agencies to reach a final number that is consistent with the president’s budgetary goals, and then combines the individual agency budgets for reporting purposes into categories that create the illusion of a coherent R&D budget. But this budget is an artificial construct that conceals the internal history, politics, and culture of each individual agency.
The situation in Congress is even more Byzantine, with 20 or more authorization and appropriations committees (and innumerably more subcommittees) in the Senate and House each exercising jurisdiction over various pieces of the publicly funded R&D enterprise. Moreover, the jurisdiction of the authorizing committees does not match that of the appropriations committees; nor do the allocations of jurisdiction among Senate committees match those of the House. Finally, the appropriations process puts S&T agencies such as the National Science Foundation (NSF) and the National Aeronautics and Space Administration (NASA) in direct competition with other agencies such as the Department of Justice and the Office of the U.S. Trade Representative for particular slices of the budgetary pie.
In total, the decentralization of influence over S&T budgeting in the federal government precludes any strategic approach to priority setting and funding allocations. Although an “R&D” budget can be—and is—constructed and analyzed each year, this budget is an after-the-fact summation of numerous independent actions taken by congressional committees and executive-branch bodies, each of which is in turn influenced by its own set of constituents and shifting priorities. From this perspective, if science policy is mostly science budget policy, then one can reasonably assert that there is no such thing as a national science policy in the United States.
If central science policy planning in the United States is impossible, how is one to make sense of the remarkable stability of R&D spending as a proportion of the total discretionary budget? Several related factors come into play. First, the political dynamics of budget-making result in a highly buffered system where every major program is protected by an array of advocates and entrenched interests fighting for more resources and thus, on the whole, offsetting the efforts of other advocates and interests trying to advance other programs. Second, annual marginal changes in any program or agency budget are generally small: This year’s budget is almost always the strongest predictor of next year’s budget. Large changes mean that a particular priority has gained precedence over other, competing ones, and such situations are not only uncommon but usually related to a galvanizing political crisis, such as 9/11, or the launching of Sputnik. Third, in light of the previous considerations, annual changes in expenditure levels, whether for R&D programs or judges’ salaries, are on average going to be in line with overall trends in the federal discretionary budget as a whole. Thus, long-term stability in R&D spending as a percentage of the whole budget is what we should expect to see exactly because of the decentralized essence of science policy.
Of course this reality means that federal support for S&T is subject to the same political processes and indignities as other federal discretionary programs. Although such a notion may offend the common claims of privilege made on behalf of publicly funded science, it also offers evidence of a durable embeddedness of S&T in the political process as a whole, an embeddedness that has offered and will probably continue to offer significant protection against major disturbances in overall funding commitments for R&D activities. For this reason, predictions of impending catastrophe for research budgets (for example, in the mid-1990s, many science policy leaders believed that cuts of up to 20% in R&D were all but inevitable) have not come true. On the other hand, in periods of particular stress on the discretionary budget (and we are now in such a period) R&D faces the same budgetary pressures as other crucial areas of government budgetary responsibility, from managing national parks and supporting diplomatic missions to providing nutritional programs for poor infants and mothers or monitoring the safety of the nation’s food supply.
Thus, any time of famine (or feast) for public civilian R&D funding as a whole will be a time of famine (or feast) for most other nonmilitary government programs subject to the annual budgeting process. Any argument that R&D deserves special protection from budgetary pressures is implicitly an argument that other programs are less deserving of protection. One sure way for R&D advocates to threaten the considerable stability of research funding in the budget would be to begin to target other, non-R&D programs as somehow less deserving of support. However compelling such arguments might seem to those who recognize the importance of a robust national investment in R&D, they will also be a provocation to those who are similarly compelled by competing priorities.
Stability means growth
It will not have escaped the alert reader’s notice that the 1960s do not fit into the story I have been telling so far. As Figure 1 shows, civilian R&D funding relative to discretionary spending increased markedly in the early 1960s, peaked in 1965, and then declined to the levels that were to characterize the next 30 years. This excursion can be explained in one word: Apollo.
In the wake of Sputnik and at the height of the Cold War, President Kennedy’s decision to send people to the Moon represents by far the most notable exception to the highly stable, buffered system that characterizes recent public funding for R&D: NASA’s budget increased about 15-fold between 1960 and 1966. At the apogee of Apollo spending in 1966, nondefense R&D accounted for 25% of nondefense discretionary expenditures, but if you subtract the NASA component, the R&D investment falls to only about 6%. The perturbation was driven by external geopolitical forces; this was not the internal logic of scientific opportunity making itself felt.
Apollo aside, the stability of the government’s commitment to R&D as a proportion of its entire portfolio of discretionary activities also represents a commitment to growth. In 1960, before the Apollo ramp-up, nondefense R&D made up about 10% of nondefense discretionary spending, a level to which it returned after Apollo. Meanwhile, from 1962 (the first year for which reliable comparable date are available) to 2006, total nondefense inflation-adjusted R&D expenditures (in FY 2000 constant dollars) rose 335%, a rate of increase that closely mirrors general budgetary growth as a whole, rather than some natural rate of expansion of the knowledge-producing enterprise.
Of course these macroscale trends conceal internal variations. NASA’s rapid budgetary ascension was followed by a more gradual decay curve, but even at its 1974 post-Apollo perigee, NASA’s budget of $3.2 billion (in current dollars) exceeded that of any other civilian R&D agency. NIH did not catch up to NASA until 1983, and did not leave it in the dust until the doubling began in 1998 (today, NIH’s budget is almost 2.5 times NASA’s). In 1977, in the wake of the Arab oil embargos, President Carter consolidated several agencies and programs to create the Department of Energy (DOE), with budgets of NASA-like magnitude. DOE’s fortunes declined by almost 30% in terms of spending power under President Reagan, and today, after accounting for inflation, its budget is less than it was at its inception. NSF, whose importance for supporting university research belies its relatively modest share of overall nondefense R&D, has experienced budget increases in all but 2 of the past 42 years. By remaining more or less aloof from focused political attention, NSF has avoided the volatility of DOE and NASA and, like the old blue-chip stocks, has yielded persistent if unspectacular growth. Its 2006 level of $5.5 billion ($4.2 billion for research) is still considerably less than that of NASA ($11.3 billion), DOE ($8.6 billion), or NIH ($28.4 billion). Overall, while civilian R&D as a whole is under the grip of a sort of budgetary lock-in due to larger political forces, the internal texture of the R&D budget is continually rewoven by dynamic political processes.
The stability of the federal commitment to S&T is matched, indeed exceeded, by a particular commitment to basic research. One of the most persistent myths of science policy is that government support for basic research is soft and has eroded over time. Indeed, the vulnerability of basic research to the vulgarities of politics has been an article of faith among science advocates at least since World War II, when Vannevar Bush, chief architect of NSF, explained in the classic report Science, the Endless Frontier that both government and industry were naturally inclined to favor investments in applied research over basic. Yet federal basic research investments (nondefense and defense) over the past 40 or so years have risen more quickly than R&D budgets as a whole. In 1962, the government investment in basic research was about 60% that of applied research; basic and applied reached parity in the late 1970s; and in recent years, driven especially by the NIH doubling, basic has exceeded applied by as much as 40%.
Despite the warnings of Vannevar Bush and subsequent science advocates, the political case for basic research is both strong and ideologically ecumenical. Unlike applied R&D, basic research appeals to the political left as an exemplar of the free expression of the human intellect, to the political right as an unambiguously appropriate area of government intervention because of the failure of the market to provide adequate incentives for private-sector investment, and to centrists as an important component of the government’s role in stimulating high-technology innovation. In 1994, when Democrats lost their grip on Congress to a Republican majority bent on budget cutting, I recall that many of my scientist friends went into a panic, certain that academic basic research would be on the chopping block. But the value of federal investments in basic research was one thing that President Bill Clinton and House Speaker Newt Gingrich could agree on, and basic science fared well—better than it had under the Democrats in the first two years of the Clinton regime. Moreover, the general public, so often characterized as scientifically illiterate by politically illiterate scientists, has for decades shown very strong support for basic research in public opinion surveys.
Of course research that is “basic” is not necessarily irrelevant. Not only may scientists be curious about problems of abiding practical interest, but scientists may pursue questions that to them are of purely intellectual interest but to research administrators and policymakers are part of a more strategic effort to advance a particular mission. Economists such as Nathan Rosenberg and Richard Nelson and historians such as Stuart Leslie have demonstrated that basic research agendas have throughout the post–World War II era been strongly tied to the priorities of the private sector and national defense. Even arcane fields such as subatomic particle physics were justified during the Cold War in part because they were the training grounds for the nation’s next generation of weapons scientists. The great majority of federal research categorized as basic is funded as part of larger agency missions, NIH being the most obvious example today, and the Department of Defense in earlier decades. NSF is the most conspicuous exception, although this, too, has been changing as NSF priorities increasingly focus on real-world priorities ranging from climate change to nanotechnology. Although such realities take the gloss off notions of scientific purity, they help to explain why, year after year, 535 members of Congress, most of whom have little if any deep knowledge of science, continue to treat basic science with a level of consideration equal to that of new post offices, interstate cloverleafs, and agricultural price subsidies.
The budgetary picture for R&D is not just about public investments, of course. When public and private support for R&D are considered together, evidence of consistent growth is even more pronounced, with inflation-adjusted expenditures rising from $71 billion in 1962 to $270 billion in 2004 (in FY 2000 dollars). Most of this growth has come in the private sector; indeed, this period shows a progressive decline in the ratio of federal to private funding for R&D. In 1962, the government funded twice as much R&D as did private industry. The continued growth of the high-technology economy led to increasing private-sector investment in R&D, and by 1980 the share of R&D funded by industry slightly exceeded the public share. By 2004, industrial R&D funding was more than twice that of government. From a simple market failure perspective, this trend represents a tremendous success: The private sector is taking on an increasing share of the knowledge-creation burden of society as the government investment brings increasing long-term economic returns.
There are, however, many good reasons for investing in science beyond just compensating for market failure. I want to emphasize that the continual attention paid to the amount of investment in R&D, whether public or private, military or civilian, tends to come at the expense of attention to these other reasons. One can easily imagine a variety of very different R&D portfolios for given levels of investment. Presumably each of those portfolios would contribute to very different sets of social outcomes. One could even imagine a large R&D investment portfolio organized in a way that contributes less to public well-being than a different, smaller portfolio.
A real-world experiment
The real world provides a way to begin thinking about such issues. R&D policies, and the resulting structures of national R&D enterprises, do vary significantly from nation to nation. For example, among nations that invest substantially in R&D, industrial funding ranges from 75% of total R&D in Japan, to 63% in France, to 51% in Canada. Within the public investment sphere there are enormous differences in priorities among nations. The most obvious indicator of this diversity is biomedical science, which in the United States commands almost 50% of the total federal nondefense R&D budget, compared to 4% in Japan and Germany, 6% in France, and 20% in the United Kingdom. Similarly, Japan devotes about 20% of its civilian R&D to energy, whereas the United States spends about 3%, Germany 4%, France around 7%, and the United Kingdom 1%. The United States spends about 20% of its civilian R&D on space, France 10%, Germany 5%, and so on.
Such numbers don’t tell the whole story, because European nations and Japan distribute large chunks of their federal R&D dollars in the form of block grants to universities, which then have discretion in allocating among various fields. Both within and between nations there is an enormous diversity of policy models used to determine R&D priorities, to translate those priorities into actual expenditures, and to apply those expenditures to science. In some ways the United States, with its Balkanized budgetary authority, is more decentralized and more diverse than most other affluent nations. At the same time, in the United States there is a tighter linkage between specific agency missions and funding allocation to research performers than in many other R&D-intensive nations. And of course the decision processes for the disbursement of funds in universities and national laboratories (as well as the relative roles of different types of R&D-performing institutions) vary greatly from nation to nation and within nations. The role of peer review, smart managers, earmarks, block grants, and equity policies (for example, NSF’s Experimental Program to Stimulate Competitive Research and German efforts to support institutions in the east) are all highly variable and reflect different institutional and national histories, politics, and cultures. Human resources are also variable, with the proportion of scientists and engineers at over 9 per 1,000 in Japan, 8 in the United States, 6 in France and Germany, and 5 in the United Kingdom. And of course the role of public R&D in “industrial policies” has varied greatly over time within the United States alone and varies greatly between nations and between sectors. The 1991 Office of Technology Assessment report Federally Funded Research: Decisions for a Decade summed up the situation: “While there may be certain universality in science, this does not carry over to science policy.”
Of course it is these very details that make up the nuts and bolts of what science and technology policy is supposed to be all about, and much emotion and energy are invested in promoting policies that favor one approach, priority, or program over another. But given the great diversity in science policies both within and among affluent nations, and given the relative similarity of the macroeconomic and socioeconomic profiles of these same nations, I can see no reason to believe that there is a strong linkage between specific national science policies and general national-scale socioeconomic characteristics. Of course, the United States has a strong aerospace industry and a strong pharmaceutical industry in part because of R&D policy priorities during the past 50 years, but it still has a 20% pretax poverty rate, average life expectancy in the mid-70s, gross domestic product per capita above $30,000, a Human Development Index above 0.9, and so on, just like other affluent nations with very different approaches to investing in R&D. (The United States also has famously mediocre public health indicators despite its gargantuan investment in biomedical research.) So, although federally sponsored S&T are obviously causal contributors to public welfare and although S&T policies of some sort are necessary to ensure such contributions in the future, there is little reason to imagine that, at the macro policy level, particular policy models and choices make much of a difference to broad socioeconomic outcomes. There seems to be a diverse range of options that work more or less equally well, and this diversity may itself be a component of success.
From this perspective, the machinations of science policy—the constant stream of conferences, reports, and opeds; the dozens of committees and working groups; the lobbying and legislation; the hyperbole and anxiety—are best viewed as metabolic byproducts of a struggle for influence and funding among various political actors such as members of Congress, executive-branch administrators, corporate lobbyists, college presidents, and practicing scientists. The significance of this struggle is largely political and internal to the R&D enterprise; it is not a debate over the future of the nation, despite continual grandiose claims to the contrary. We are mostly engaged in science politics, not science policy. Or, to adopt the perspective of Thomas Kuhn, this is normal science policy, science policy that reinforces the status quo.
Publicly funded research is justified on the basis of promised contributions to desired social outcomes: to “increase quality and years of healthy life [and] eliminate health disparities” (U.S. Department of Health and Human Services), to “conserve and manage wisely the Nation’s coastal and marine resources” (National Oceanic and Atmospheric Administration), or to ensure “a safe and affordable food supply” (U.S. Department of Agriculture). What is the capacity of a particular science policy decision to advance a given desirable outcome? This should be the most fundamental science policy question, because if one cannot answer it, then one cannot know whether any particular policy is likely to be more or less effective than any alternative policy. And if one cannot choose among alternative policies in terms of what they may achieve, then policy preferences are revealed as nothing more than expressions of parochial values and interests. This, as I’ve discussed, turns out to be a perfectly good approach to ensuring that R&D investments are treated as well as other public investments. But it does not tell us whether different investment choices would yield better (or worse) outcomes.
At the highly aggregated level of “national” R&D policies, it appears that different approaches yield more or less similar outcomes. But when trying to connect R&D to particular desired outcomes, policy choices obviously can matter greatly. The doubling of the NIH budget is a case in point. This doubling occurred without any national dialogue about what it might achieve or about alternative paths toward better national and global health. In part because of the close coupling between NIH research agendas and biomedical industry priorities, high-technology intervention (often at high cost, as well) has been adopted as the national strategy for improved health, and no serious consideration was given to alternative health investment strategies that might be equally effective in contributing to public well-being but less likely to contribute to significant corporate profitability. Even further from the debate was the question of whether biomedical research was the area of science that could yield the most public value from a rapid increase in investment, rather than, say, energy R&D. One might compellingly have made the case that the nation’s health challenges are far less of an immediate threat to well-being than its dependence on fossil fuels imported in large part from other nations. That there are no mechanisms or forums to explore these types of tensions between alternative approaches to a particular goal, such as better health, or between competing goals, such as better health or better energy, is precisely the problem.
The internal political dynamics of science budgeting help to explain why R&D policy discussions are dominated by concerns about “how much” and avoid like the plague serious questions about “what for.” But I’m suggesting here that the “how much” obsession may paradoxically reduce the potential contribution of R&D to well-being, because “more” and “better” are simply not the same things. For example, “how much” carries with it a key, but unstated, assumption: that everyone is made better off by investments in science. If the benefits of science are broadly and equitably distributed, then “how much science can we afford?” is a reasonable central question for science policy, especially given the decentralization of the policy process. Whatever the priorities may be, we can expect that all will benefit. But if the positive and negative effects of science are unevenly distributed, the primacy of “how much” becomes more difficult to justify from a perspective of good governance and good government. And of course the positive and negative effects of science are indeed unevenly distributed. In fact, given what we know about such problems as unequal access to health care and the disproportionate exposure of poor people to environmental hazards, my colleague Edward Woodhouse and I have recently suggested a hypothesis that seems to us both reasonable and worthy of careful testing: New scientific and technological capacities introduced into a highly stratified society will tend disproportionately to benefit the affluent and powerful.
It is not very difficult to imagine the types of questions that might help to inform a transition from science policy based on “how much” to science policy based on “what for,” though it is certainly the case that such questions may be rather unwelcome in national R&D policy discussions and that even partial answers will not always be available, at least at first. But here are 10 questions that, if made explicit in science policy discussions, could help with the transition.
- What are the values that motivate a particular science policy?
- Who holds those values?
- What are the actual goals that the policy is trying to achieve?
- What are the social and institutional settings in which the R&D information or products will be used?
- What are the reasons to expect that those are settings for effectively translating the results of R&D into the goals that justify the policy?
- Who is most likely to benefit from the translation of the research results into social outcomes?
- Who is unlikely to benefit?
- What alternative approaches (through either other lines of research or nonresearch activities) are available for pursuing such goals?
- Who might be more likely to benefit from choosing alternative approaches?
- Who might be less likely to benefit?
Addressing these questions does not require impossible predictions of either the direction of scientific advance or the complex interactions between science and society. It does require that unstated agendas and assumptions, diverse perspectives, and the lessons of past experiences be part of the discussion. The questions are as appropriate for academic scholarship as they are for congressional hearings or media inquiries. Taking them seriously would be a step toward a science policy that mattered.