Funding isn’t enough to fix science

28th May 2021

Funding is at the forefront of policy conversations on how to improve scientific research, but our institutions suffer from larger problems that call for structural reform.

PoliticsScience

The Senate just spent a fevered 72 hours debating the U.S. Innovation and Competition Act, a bipartisan piece of legislation that aims to spur American innovation and bolster our competitiveness in science and technology by authorizing over $100 billion for federal research and development (R&D). With bipartisan support and the backing of the White House — not to mention a widely shared concern about China’s aspirations of scientific and technological dominance — some version of this bill appears likely to make it into law, eventually.

There is an urgent need to shore up America’s research establishment. But inadequate funding is only the tip of the iceberg: More fundamental problems have dogged American science and innovation for years.

Today, our breakthrough scientific discoveries appear less fundamental and less frequent than in decades past; scientists are spending increasing amounts of time on paperwork, administration, and bureaucratic compliance rather than on research itself; funding has become concentrated in a handful of well-established institutions; the average age of primary researchers continues to climb; and the so-called replication crisis has raised concerns about the reliability of many published findings. All of these point to deeper, structural problems in America’s R&D system. 

The goodwill provided by the current R&D push has created an ideal opportunity to implement much-needed reforms to U.S. science funding. Yet, as it stands, the U.S. Innovation and Competition Act (USICA) does little, if anything, to address these problems.

Start with the good news. The USICA shows that lawmakers are finally doing something about the R&D funding gap — the decades-long decline of federal research spending relative to both the private sector and foreign competitors, notably China. The USICA — previously known as the Endless Frontier Act — would substantially increase federal spending in key areas of emerging technology, from artificial intelligence and quantum computing to biotechnology and advanced energy. The latest version of the bill appears considerably less ambitious than the one originally unveiled by Sen. Schumer in 2020, with new NSF spending declining from $100B to $38.5B. But the USICA would nevertheless provide a significant boost to federal R&D with an additional $54.5B being split between the Department of Energy, the Department of Commerce, DARPA, and NASA.

Moreover, while the original bill allocated $100 billion for a new technology directorate housed within NSF, the latest version apportions funds more evenly between the new directorate and the traditional NSF. Critics worried that the original bill would warp the culture of the agency — whose annual budget currently stands at about $8.5 billion — and undermine its core mission, which is to support basic scientific research, rather than mission-directed technology research. Outstanding questions remain, including how the directorate would interact with existing NSF programming and whether it would be better suited to another federal agency or a stand-alone initiative. But there is less risk that basic science will get left behind in the rush to spend more on R&D.

Yet, the USICA does little to address the many other systematic issues that have bedeviled the American research establishment for years. Most conspicuous, perhaps, is the so-called replication crisis. 

Independent confirmation of experimental findings — whether by reproducing experiments or replicating experimental results — is generally held to be an important, even essential, part of the scientific process. Yet, in recent years, it has emerged that an alarming number of published scientific findings cannot be reproduced or replicated. The problem is particularly acute in psychology, biomedical research, ecology, economics, and the social sciences.

The crisis is not, so to speak, merely academic. The types of research potentially implicated by it include areas of science that have direct bearing on medicine, public health, and government regulation. Lawmakers have taken note. In 2018, Congress directed NSF to commission the National Academies of Sciences (NAS) to study the crisis and provide a diagnosis as well as potential prescriptions.

There is deep and ongoing disagreement within the scientific community about the nature and extent of the crisis — including whether it is a genuine problem and what, if anything, can be done about it. Nevertheless, the NAS report points to a growing consensus that the replication crisis is symptomatic of deeper flaws with the way scientific research is currently organized and conducted. These include obstacles to sharing scientific data, inadequate statistical training among researchers, problems with the peer-review system and grant culture, misaligned financial and professional incentives, and the corrosive effects of publish-or-perish.

Yet, the Senate bill hardly acknowledges — much less addresses — any of these problems. 

The USICA does give a nod to the problem of geographic and institutional concentration. It is well known that government research dollars are highly localized in a few wealthy regions of the country and their prominent institutions. The bill addresses this through a $10 billion program of regional “tech hubs” across the country with special provisions for historically black and other minority-serving colleges. But the bill includes no serious attempts to modify or reform existing funding formulas within the NSF or other funding agencies, much less combat the various systematic problems in American R&D that inhibit such geographic and institutional diversity in the first place.

The latest version of the bill does include provisions intended to protect American research from undue foreign influence or theft, including bolstering intellectual property, cybersecurity, and supply-chain resiliency. These, too, are important goals. Yet, insufficient attention has been paid to whether and how these measures may help or hinder the other problems facing American research, especially the compliance costs associated with byzantine federal grant requirements.

To be sure, these problems cannot all be solved by a single funding bill. But some are easier to address than others. Principal researchers should not be spending 42 percent of their time on grant maintenance and compliance. Besides diverting scientific talent from research to paperwork, this creates barriers to entry for smaller institutions with fewer financial and personnel resources. Lawmakers could experiment with different ways to reduce the compliance burden, from streamlining the federal grant process to waiving unnecessary regulations and administrative requirements. Such simple reforms should be tied to funding increases, rather than left as vague guidance to agency staff.

Other problems — such as the replication crisis — are much harder to tackle. But there are a number of things lawmakers could do in the scientific spirit of evidence-gathering and experimentation. For starters, new R&D spending should include allocations for research to study the replication crisis and attendant problems and solutions (as the NAS report on reproducibility recommends). Next, lawmakers could introduce more oversight and accountability into the federal R&D system to see what kinds of funding models work well for which purposes. 

New Zealand, for example, has been experimenting since 2015 with a lottery system for scientific grants. Lottery systems set aside some portion of well-qualified government grant seekers to be selected at random, rather than through the traditional merit system. The underlying idea is that scientific discoveries often come from unexpected places and require breaking out of existing consensuses. A modified lottery system could be another way to introduce more diversity into a grant-selection process that can be overly rigid, cautious, and preferential to established institutions. 

Lottery systems are just one of a number of possible funding models that should be studied, tried, and compared. Others include long-term funding for individual researchers or research teams, rather than short-term funding for discrete projects; increasing the percentage of small grants, especially to less-established research institutions; prioritizing grants to early-career scientists; and integrating more viewpoints into the grant selection process, for instance by diversifying grant committee participants or encouraging researchers with adversarial views to work together on shared projects.

Because such reforms challenge the status quo, they are likely to face resistance, especially from well-established agencies and research institutions. Indeed, this kind of structural experimentation — requiring real-time oversight and dynamic feedbacks to gather evidence about what works and what doesn’t— cuts against the natural tendencies of the modern administrative state, which is to formalize and solidify processes into explicit regulations and procedures. Congress should not simply authorize more R&D funding without demanding more diversity and experimentation in return.

American science does need more funding. But it needs structural reform too. The USICA risks offering American research establishment a spoonful of sugar — billions of dollars’ worth —without any medicine. Reforming long-established institutions and programs is difficult, but the future of American innovation demands it.