The era of constrained federal science budgets is over. With Congress poised to boost public spending on research and development (R&D) by anywhere from tens of billions to hundreds of billions of dollars, science agencies may be preparing for an infusion of federal money on a grand scale that has the potential to transform the institutions of science and technology.
Unfortunately, however, these proposals to bolster American science and technology — spurred by a global pandemic, mounting concerns about climate change, and a rising China — focus almost exclusively on the need for more federal money. Yet there are other problems facing the U.S. R&D system that, if unresolved, could undermine these goals. Foremost among these problems is the increasing bureaucratization of the scientific enterprise.
Scientists have complained for years about a growing number of federal rules and regulations that hamper research productivity. Today, researchers spend nearly half their time on paperwork and administrative tasks, rather than research. Increasing R&D spending will not solve this problem — and may even make it worse.
The bureaucratization of science is not a new problem. In 1961, barely a decade after the creation of the National Science Foundation (NSF), the American physicist Alvin Weinberg was asking whether Big Science — big teams of scientists working on big scientific projects with big government grants and contracts — was “ruining science” by transforming scientists into bureaucrats. Weinberg, then the director of Oak Ridge National Laboratory, identified three causes for concern.
First, large budgets do not necessarily correlate with high-quality science, and Weinberg worried that big grants would become ends in themselves. Second, large budgets inevitably produce a need for large numbers of administrative staff to help secure and manage funds. Third, Weinberg observed that pressure to win large federal grants and contracts was transforming science itself by “converting university professors into administrators, housekeepers, and publicists.”
The bureaucratization of science has only gotten worse since Weinberg’s time. For instance, a 2014 NSF report noted that concerns about the “overregulation” of science and increasing “administrative workload” associated with federal rules and requirements had been documented in past surveys and reports for more than a decade. Among the most cited sources of administrative burden by researchers? “Financial management; the grant proposals process . . . time and effort reporting.” Weinberg appears to have been prophetic.
Lawmakers have taken note. At Congress’s request, the National Academies issued a report in 2016 that assessed the problem. The report estimated that the number of “new or substantially changed” regulations on scientific research issued by the federal government increased from an average of 1.5 per year during the 1990s to 5.8 per year from 2003 to 2012. These findings are consistent with other assessments, including the Faculty Workload Survey, which in 2005 and 2012 estimated that researchers spend a whopping 42.3 percent of their time on “tasks related to research requirements (rather than actively conducting research).”
In 2016, Congress passed the 21st Century Cures Act, which directed the Office of Management and Budget (OMB) to set up a Research Policy Board to “provide the federal government with information on the effects of regulations related to federal research requirements.” In 2017, Congress passed the American Innovation and Competitiveness Act, which included provisions to create a working group to “reduce administrative burdens on federally funded researchers.” The executive branch, too, has made efforts to streamline administrative requirements and has convened working groups and task forces to reduce bureaucratization.
But the problem persists. A 2016 Government Accountability Office (GAO) report found that “continuing efforts” to reduce the universities’ administrative burden on researchers had shown “limited results.” Now, half a decade later, it appears the situation is unchanged and may even be worsening. Indeed, according to the latest Faculty Workload Survey from 2018, the average amount of time researchers lost to bureaucratic compliance has increased — to 44.3 percent.
Why is bureaucratization such a danger to science?
The most obvious danger is opportunity cost. How many first-rate papers have been left unfinished or scientific discoveries left unmade because researchers are spending nearly half their time on administrative tasks? Of course, administrative tasks are essential to sustain science and its institutions. And the boundary between “scientific” and “administrative” tasks may sometimes be murky, especially in large scientific institutions or on interdisciplinary research projects. Some scientists can and must be effective administrators. But in general, the most effective allocation of scientific talent is on scientific rather than administrative tasks.
Another danger is that the increasing number of non-scientist administrators needed to secure and manage large grants can become a barrier to entry for smaller or less well-established institutions. As a result, those institutions — typically prestigious and wealthy ones — capable of growing and sustaining such a bureaucratic apparatus may reap a disproportionate share of federal funding. The geographic concentration of federal science funding around a relatively small number of elite research universities offers indirect evidence of this.
Finally, there is a risk that scientists will come to be evaluated on their capacity to attract grants, rather than on the quality of their work. Under such conditions, one might expect to see a decline in the quality of scientific research, for, as Weinberg pointed out, “science dominated by administrators is science understood by administrators” and “such science quickly becomes attenuated if not meaningless.” Today’s concerns about the quality of scientific research and the inability to “replicate” many published research findings suggest that Weinberg might have been prophetic about this effect of bureaucratization as well.
These problems are well known in scientific and science-policy circles. What is striking is how little attention has been paid to any of them in the current debates about boosting American science and innovation, especially given this long history of concern. Yet without appropriate reforms, it is hard to see how large funding increases will do anything but make such problems worse. Moreover, if the rise of federally funded research is itself implicated in the current crisis, there is a real danger that attempted solutions at the federal level could do more harm than good.
So what can lawmakers do?
Despite these challenges, there are important steps that lawmakers could take. Here are four recommendations that should be considered in conjunction with any proposals to significantly increase funding for federal science agencies.
First, the OMB should establish the Research Policy Board (RPB). Although required to do so by the 2016 Cures Act, the OMB never established the RPB. The RPB is by statute set to expire this month, so Congress should reauthorize it, as was recently recommended by the GAO, and expand its scope to include a multi-stakeholder assessment of best practices for R&D funding generally.
Second, Congress should direct science agencies to establish a “second look” program for science grants to experiment with alternative funding mechanisms. This could be done through new pilot programs. One would be a modified lottery system, which awards funding based on a lottery among a subset of “most meritorious” proposals that have not yet received federal funding. Another would waive additional federal rules and requirements for a subset of well-qualified scientific researchers.
Third, Congress should direct the NIH to expand the scope of its Director’s Pioneer Award to include researchers from basic scientific fields in the natural sciences. Additionally, Congress should direct the NSF to create similar programs that fund people rather than projects, tailored to its own areas of focus and excellence.
Fourth, Congress should establish appropriate feedback mechanisms to gather evidence about the effectiveness of these policies in coordination with relevant scientific agencies, the GAO, and the newly reauthorized and expanded RPB.
By themselves, these proposals would not resolve all the problems plaguing the U.S. R&D system. But they would be a step in the right direction, recognizing that the way we fund science — not just how much we fund it — can affect the organization and practice of scientific research, for better or worse.
There are strong arguments for increasing federal investment in scientific research today, especially in basic science. And the rapid development of multiple COVID-19 vaccines — to take just one particularly striking example — offers a recent and powerful reminder of how astonishingly successful the U.S. R&D system can be at its best. But if lawmakers are serious about preserving America’s scientific and technological prowess, they must find ways to fix and not simply fund science. The emerging political consensus about the need to bolster federal R&D offers the perfect opportunity to do so.