Stagnant: A Decade of Small, Fast Grants at the NSF

I should start with my admiration: the National Science Foundation (NSF) is the best federal science funding agency for both small grants and fast funding. Their EAGER and RAPID mechanisms have been catalytic for starting new research directions and responding quickly to urgent issues, like the COVID-19 pandemic. They’re the gold standard among federal research agencies. But they could be better.

I recently wrote a proposal for the Day One Project (I’ll link once it’s published), suggesting a tweak to the current arrangement: creating Micro-ARPA program managers whose only job is allocating small, fast grants. Below is an addendum with the data behind the argument. More importantly, I added suggestions for what this data means for science funders everywhere—not just the federal agencies.

In my original “Science Angels” writing, I leaned heavily on a 2013 paper by Caroline Wagner and Jeffrey Alexander called “Evaluating transformative research programmes: A case study of the NSF Small Grants for Exploratory Research programme”. The paper analyzed the Small Grants for Exploratory Research (SGER) program at the NSF—the small, fast grants program that predated and inspired EAGER and RAPID—and found the mechanism to be wildly effective and efficient: one in ten of the small grants turned into transformative results, which they defined rigorously. One in ten! As far as research funding goes, those are good stats.

The other major point of the paper: for some unknown reason, the NSF program officers were not using the SGER mechanism as much as they could have been. It was effective and underutilized.

This was supposed to change with EAGER and RAPID. In 2009, when Arden Bement, NSF Director at the time, launched the EAGER and RAPID program to replace SGER and “urged NSF program officers to spend up to 5% of their budgets on them.”

That’s never happened. Not even close. Even in the RAPID-heavy year of COVID response in 2020, the combined EAGER and RAPID budget didn’t crack 3%. The Wagner and Alexander paper, with their glowing report card, didn’t change the trajectory much, either.

Here’s the total amount of EAGER and RAPID funding in the decade following the report:

There was a hopeful bump, then the COVID spike, but now tailing off. Here’s how that relates to the rest of the NSF budget:

Still far from Bement’s hope of 5%, and trending in the wrong direction.

Another interesting way to view the data is by the volume of grants given—not just the total dollar amounts. After all, these are small grants, given with the expectation that a small amount could go a long way to proving a new idea. The numbers tell a similar story there:

Again, the data shows blips of enthusiasm, but overall it’s a steady decline.

For fun and perspective, I added the grant volume of our small non-profit, the Experiment Foundation. Our numbers will increase by ~50% again this year and—if trends hold on both sides—we could eclipse the volume of NSF EAGER grants in 2026 or 2027, which is wild to imagine. We have one half-time staff member (me) and a few dozen Science Angels doing this work. To be fair, our grants are much smaller—mostly less than $10k. But it’s worth making the point: more is definitely possible. I’ve lived it.

A commitment to early, exploratory grantmaking matters. An experienced ARPA funder told me that she couldn’t find a correlation between the success of a project and the size of the grant. Any financial investor would tell a similar story. Given the uncertainty of outcomes, small grants should be favored simply because we can do more with less.

Every EAGER grant that goes out represents a new and hopeful research direction. The more funding we allocate to EAGER and RAPID, the higher the volume of grants the agency can deploy, and the more scientists are on the playing field.

So what? What to do about this data? I have three suggestions.

One is for the federal funding agencies: try something different. Don’t just admonish Program Officers to use these mechanisms more. That approach has flatlined. It’s time to test a new arrangement. My Day One Proposal to create Micro-ARPA Program Managers to disperse the funding is one such tweak, but there may be others.

Also, the NSF has been the leader so far, but there’s no reason other agencies couldn’t implement the mechanism and improve the process. NOAA, BOEM, the Forest Service, etc—I’m looking at you! Your research budgets aren’t nearly the size of the NSF or NIH, but you could turn that into your advantage. There is a niche to fill here.

The second suggestion is for the philanthropists, foundations, and donors funding scientific research: don’t mimic the NSF. I’ve noticed a tendency for philanthropists and foundations—when faced with the prospect of setting up a new research funding program—to replicate the example set by the NSF and NIH: put famous scientists in charge, run an extensive peer-review process, and try to correlate their dollar amounts to the grant sizes of the federal agencies. That’s all fine and good, but there’s a missed opportunity. The federal agencies have limitations, and small, fast grants are currently one of them. Philanthropic funders could fare better, and their funding going further.

Lastly, to the scientists: become the science funder you want to see in the world. I’ve now worked with more than a dozen science angels who’ve started their own micr0-grant program with the Experiment Foundation—finding and filling niches in the scientific funding landscape, and helping to kickstart hundreds of new projects and careers. One lessons is clear: the skill of being an early-stage science funder is unique and distinct from being a great scientist. We need more people to try their hand at this role—taking on the persona of a science angel—to maximize the opportunity for small, fast grants to turn into more transformative results. A better science funding ecosystem depends on it.