Who regulates the physical and informational risks of scientific research? Who makes sure that researchers keep intensely personal data secret, or lock their dangerous chemicals away safely, or don’t haphazardly create superviruses that could kill us all? You could be forgiven for assuming it was, well, a regulator—OSHA or the CDC or some similar entity that can promulgate binding rules and enforce them. But as Doni Bloomfield shows in the fascinating How Science Grantmakers Govern, you’d be (mostly) wrong. Instead, a key substantive and procedural regulator of the risky business of scientific research is, in fact, the funder of that research. Bloomfield’s case study considers the leviathan of basic research funding, the National Institutes of Health (NIH). In the first part of this article, Bloomfield pulls apart how NIH acts as a regulator. The second part of this piece draws out implications in a clearly written and fascinating read.
One framing note: Bloomfield purposefully focuses on how grantmakers governed up until the beginning of the second Trump administration. It’s a reasonable choice; there’s much to say about how NIH became a significant regulator of scientific research, and what that means for innovation and administrative law theory. Accordingly, the paper doesn’t center how the second Trump administration has used the grant system as a cudgel to beat down universities, or how it’s tearing apart the system of scientific funding. Bloomfield does address the recent changes, but they’re not his focus. There is a lot in this paper; it’s full of insight and replete with interesting historical details (Did you know that the term “peer review” was first used by NIH? I did not!). It’s worth a read, and then another.
The first half of the paper makes a convincing descriptive case that NIH does, in fact, regulate the scientific research process, using a rich set of historical examples to show both what NIH regulates and how it does so. For example, do you know why genomic data is hugely concentrated in a big database with pretty strict access controls? It’s because NIH requires it!
Bloomfield groups NIH regulation of research into three buckets: how research is conducted, how the research workplace functions, and how results are shared. He illustrates each with examples like biosafety rules and Institutional Review Boards (IRBs) for conduct, sexual harassment policies for the workplace, and genomic-data sharing and dual-use-technology publications for results sharing. In each context, NIH uses some combination of funding mandates (if you want our money, follow our rules), expertise (NIH is full of scientist researchers), and prestige to create and enforce a substantive regulatory regime.
To dig a little deeper, take IRBs. Approximately all federally funded research on human subjects—and a substantial penumbra beyond that—has to be approved by an IRB, which is typically a local body comprising experts and nonexperts that evaluate research to make sure it’s ethical and adequately protects the humans involved. Today, the IRB system is mandated by federal law and linked to over a dozen different agencies. But it started as an internal NIH initiative, then was imposed on grantees as a condition of funding, and later spread (through NIH’s prestige) as a model across a wide swath of the research enterprise.
The IRB system also exemplifies what Bloomfield calls “federally mediated self-governance” or “embedded regulation.” The regulated parties (researchers) do much of the governing themselves, but through NIH-created structures, with NIH-created rules, and with NIH looking over their shoulders and enforcing compliance through its funding mandates. It’s a fascinating blend of federal oversight and self-regulation, and it applies to many of the regimes that Bloomfield lays out.
This is really interesting descriptive work, and it’s convincing and novel. I’ve spent oodles of time thinking about NIH, and I’d never thought of it as a regulator. I do now. I suspect you will too. And Bloomfield writes clearly and includes plenty of signposting and recaps to guide the reader.
But wait, there’s more! In the second half of the paper, Bloomfield lays out implications: What does this mean for innovation theory, for administrative law, and for scientific grant systems going forward (i.e., after the Trump administration)? The interaction between the first two is particularly fun. There’s recently been a rise in literature stepping beyond classic innovation tools like patents or grants to explore how other regimes, including regulation, shift innovation incentives. This paper highlights the opposite: a classic innovation incentive that also does significant regulatory work. The way NIH administers grants influences how it can regulate. Scientists need to keep getting grants, which gives NIH effective enforcement power; and NIH needs scientists to conduct peer review, which embeds NIH in the scientific community and locates its expertise there. It all fits together.
There’s lots left to ponder, of course. How much is this a story about grantmakers governing in general, and how much is it a story about NIH in particular? Bloomfield acknowledges the limitation, and hints at regulation by other funders, such as the Department of Defense and the Department of Energy, or even non-science-funders, like the Department of Housing and Urban Development. Each sets codes that must be followed to access their funds. My hunch is that the more a funder is either necessary for work to happen or so prestigious that everyone follows their lead, the more effectively it can regulate. NIH is both overwhelmingly large and highly prestigious. Does the National Science Foundation have the same regulatory oomph? HUD? We’ll have to wait for further work to see.
Having grantmakers govern scientific research also leaves a lacuna: if their oversight becomes the dominant mode of research regulation, who steps in when research isn’t funded? The era of big corporate labs has returned, most notably in the AI space, and it’s become obvious that regulation of that research is…shall we say, limited? How much did a default of grantmaker governance, in lieu of administrative governance, contribute to that flat-footedness?
There remain many questions and vistas for exploration. But Bloomfield provides a terrific entrée into a new area for scholarship, one which will only grow in importance as the science funding system keeps transforming at an astonishing pace.
