When President Obama signed the American Recovery and Reinvestment Act (ARRA) on February 17, 2009, the headline framing was jobs and infrastructure — roads, bridges, unemployment insurance, state fiscal relief. Less visible in the initial coverage, but significant for the federal research enterprise, was a substantial one-time infusion of money into the science agencies: the National Science Foundation, the National Institutes of Health, and the Department of Energy's Office of Science each received Recovery Act funding on a scale that, for NSF and DOE Science in particular, represented a meaningful percentage addition to their annual budgets. Nearly two years on, with most of that money obligated and a good deal of it spent, it is a useful moment to look at what the science funding in ARRA actually did.
This is not a piece about whether the Recovery Act worked as macroeconomic stimulus — that question is contested and outside the scope of what a STEM-focused review can usefully adjudicate. It is a narrower and more tractable question: what did the science money specifically fund, how did it move through the federal research agencies, and what effects is it having on the research workforce, including the graduate students, postdocs, and early-career researchers who make up a disproportionate share of the people actually doing the funded work.
How Much Money, and Where It Went
The National Science Foundation received $3.0 billion in ARRA funding: $2.5 billion for Research and Related Activities, $400 million for its Major Research Equipment and Facilities Construction account, and $100 million for Education and Human Resources. For an agency with a regular annual budget in the neighborhood of $6–7 billion at the time, a $3 billion supplement was not a rounding error — it was closer to half a year's additional funding, compressed into a roughly two-year disbursement window.
The National Institutes of Health received the largest sum in absolute terms: approximately $10.4 billion in Recovery Act funding for use across fiscal years 2009 and 2010, a roughly one-third supplement to NIH's regular appropriation for that period. The Department of Energy's Office of Science received $1.6–2.0 billion, depending on how the accounting is drawn, on top of a regular FY2009 appropriation of approximately $4.77 billion — a substantial percentage increase for an office that funds much of the nation's basic physical-science research infrastructure, including the national laboratories.
The Government Accountability Office, tasked with tracking Recovery Act spending across agencies, reported in its ongoing reviews that the science agencies moved this money through existing grant-making mechanisms rather than creating new programs from scratch — a decision that had real consequences for how quickly the money reached researchers and what kind of research it could fund.
Why "Shovel-Ready" Research Meant Something Different at NSF Than at DOT
The Recovery Act's broader design emphasized speed: money was supposed to move quickly into the economy, which is straightforward for a repaving project with an engineering plan already on file and considerably less straightforward for basic scientific research, where the timeline from proposal to result routinely runs years rather than months. NSF and NIH addressed this tension in a similar way: rather than solicit large numbers of new proposals (which would have taken many months to review under normal peer-review timelines), both agencies leaned heavily on funding meritorious proposals that were already in the pipeline — grant applications that had been reviewed and scored highly but not funded due to normal budget constraints.
This approach had a specific and, in retrospect, important effect: it meant ARRA science funding disproportionately supported research that had already cleared a competitive peer-review bar, rather than newly conceived projects assembled quickly to capture stimulus dollars. Critics of stimulus spending in other sectors sometimes pointed to rushed, poorly vetted projects; this was a substantially smaller risk in the NSF and NIH portions of ARRA precisely because the "shovel-ready" research was, in most cases, research that had already been vetted through the normal competitive process and simply lacked funding.
The Graduate Student and Postdoc Effect
For the STEM workforce specifically, the most direct effect of ARRA science funding was on graduate research assistantships and postdoctoral positions. A large fraction of NSF and NIH research grants fund personnel — the graduate students and postdocs who do the actual experimental and computational work — rather than equipment or overhead alone. The sudden availability of previously-unfunded, already-reviewed grants meant a wave of new graduate assistantship and postdoctoral positions opening in fiscal years 2009 and 2010, concentrated in the fields and institutions where the underlying proposal pipeline was strongest.
This created a genuine, if temporary, expansion of opportunity for people already in or entering STEM graduate programs during this window — but it also created a well-documented structural concern that research-funding economists and science-policy researchers raised at the time: a funding bulge that is not sustained creates a workforce bulge that the system cannot later absorb. Graduate students who entered PhD programs on the strength of ARRA-funded assistantships in 2009 would be entering the postdoctoral and faculty job markets several years later, at a point when the temporary funding increase had already lapsed and NSF and NIH budgets had returned to their pre-ARRA growth trajectories (or, in the years immediately following, faced budget pressure of their own). Several science-policy analyses published in 2009 and 2010 flagged this mismatch explicitly, warning that the Recovery Act's science funding, however well-intentioned, risked amplifying an already-documented problem of PhD production outpacing available research career positions in several fields.
Infrastructure and Equipment: The Longer-Lived Effect
Separate from the personnel funding, NSF's $400 million Major Research Equipment and Facilities Construction allocation and a substantial share of the DOE Science funding went toward research infrastructure — instrumentation, facility upgrades, and construction projects at national laboratories and university research centers. This category of spending has a different time profile than personnel funding: equipment and facilities, once built, continue to provide research capacity for years or decades afterward, independent of whether the operating budget that funds their use returns to pre-ARRA levels.
Several DOE national laboratories used Recovery Act funds to accelerate previously-planned upgrades to major scientific facilities — synchrotron light sources, computing clusters, and similar shared infrastructure used by large numbers of researchers across many institutions rather than by a single funded project. This is, in a fairly direct sense, the most durable legacy of the science portion of ARRA: physical research capacity that continues to be used well after the one-time funding bulge has been fully spent.
What This Means for the Women-in-STEM and Broader Workforce Question
For WIGSAT's audience, the relevant question is not simply how much money moved, but who it reached. The available agency-level reporting on ARRA science funding did not include comprehensive, publicly disaggregated data on the gender or demographic composition of the graduate students and postdocs who benefited from the funding bulge — a genuine gap in the public accountability record, and one that science-policy researchers have specifically flagged as a lost opportunity, since demographic tracking at this scale is exactly the kind of data that would let policymakers evaluate whether a large one-time funding intervention narrowed or widened existing STEM participation gaps.
What can be said with more confidence is structural: because the funding flowed through existing peer-review pipelines rather than new targeted programs, it likely reproduced, roughly proportionally, the existing demographic composition of who was already submitting fundable NSF and NIH proposals in each field — meaning it neither specifically corrected nor specifically worsened existing representation gaps, but rather funded more of what was already in the pipeline, at whatever demographic composition that pipeline already had.
Accountability Questions the GAO Has Flagged
Because the Recovery Act as a whole carried an unusually high public-accountability mandate — a dedicated Recovery Accountability and Transparency Board, mandatory public reporting of funded projects, and sustained GAO oversight — the science agencies' portion of the spending has been subject to more granular external review than a typical annual appropriation. The GAO's periodic status reports on science-related Recovery Act funding have flagged several recurring themes worth noting for anyone trying to assess how well the money was managed, not simply how much of it there was.
One recurring GAO observation was that the compressed obligation timelines the Recovery Act required — money generally needed to be committed within a defined window rather than sitting available indefinitely — created real administrative pressure on agencies used to a slower, more deliberative grant cycle. NSF and NIH program officers, by most accounts, managed this pressure by leaning on the already-reviewed proposal backlog described above rather than rushing new solicitations, which mitigated but did not eliminate the tension between speed and normal diligence. A second recurring theme was reporting burden on grant recipients: universities and individual investigators receiving Recovery Act funds faced additional quarterly reporting requirements — job counts, spending detail, project-progress narratives — beyond what a standard NSF or NIH grant normally requires, and several university research-administration offices publicly noted the compliance cost this added, particularly for smaller institutions without large dedicated grants-administration staff.
None of this amounts to a finding of mismanagement. It is closer to a documented set of trade-offs inherent in trying to move a large amount of money through a research-funding system built for a slower, steadier cadence, under a political mandate that prioritized speed and visible job-creation reporting. Readers evaluating how well the science portion of ARRA performed should weigh both the genuine research and infrastructure output it funded and the real administrative friction the accelerated timeline created.
Frequently Asked Questions
How much money did the 2009 Recovery Act provide to science agencies?
The National Science Foundation received $3.0 billion, the National Institutes of Health received approximately $10.4 billion for use across FY2009–2010, and the Department of Energy's Office of Science received roughly $2.0 billion — each a significant one-time supplement to their regular annual appropriations.
Did Recovery Act science funding go to new research or existing projects?
Primarily to already-reviewed research. NSF and NIH largely funded meritorious grant proposals that had already cleared competitive peer review but lacked funding under normal budget constraints, rather than soliciting new proposals — a strategy suited to the Act's speed requirements while limiting the risk of funding poorly vetted research.
Did the funding bulge create long-term problems for the STEM workforce?
Science-policy analysts raised concerns at the time that a temporary funding-driven expansion in graduate assistantships and postdoctoral positions could produce a workforce bulge that the research job market would struggle to absorb once funding returned to pre-Recovery Act levels — amplifying an already-documented mismatch between PhD production and available research career positions.
What is the longest-lasting effect of the Recovery Act's science funding?
Research infrastructure investments — instrumentation, facility upgrades, and construction at national laboratories and university research centers — represent the most durable legacy, since built research capacity continues to serve large numbers of researchers for years after the one-time personnel-funding bulge has been fully spent.
← GASAT 12 Recap (2010) | ← Women in ICTs: The African Access Gap (2009)