Trump’s NIH And NSF Cuts Estimated To Cost The U.S. Economy $10 Billion Annually

America’s long-term growth is at stake. Congress is weighing sweeping FY2026 cuts to fundamental science, including $18 billion from the National Institutes of Health and $5 billion from the National Science Foundation.

The proposed dismantling of NSF raises an urgent question: do these cuts actually save money or merely delay spending until the bill gets larger? The answer is clear: these cuts will cost the economy billions.

How do we know this? New macro-empirical research finds that every dollar invested in non-defense public R&D yields $1.40–$2.10 in economic output, and since World War II, government funding has driven roughly 20% of U.S. productivity.

In this post, I explore why public investment in research matters, explain how the evidence was gathered and examine the political forces that keep America from investing more.

Bottom line: the proposed NIH and NSF cuts alone would eventually strip at least $10 billion per year from U.S. output.

Why Governments Fund Basic Research

Governments fund basic scientific research because, without them, it wouldn’t happen. The reason is market failure. Basic research generates knowledge that’s difficult to contain. Ideas diffuse, accumulate, and benefit competitors far beyond their original source. This diffusion, known as “knowledge spillover,” is great for the economy but bad for firms trying to capture a return on their investment. Unlike a product that can be patented or sold, basic science doesn’t stay put. It leaks into the broader ecosystem, lifting all boats but rewarding no one in particular. Hence, the private sector underinvests.

Economists have long recognized this. In 1959, RAND economist Dick Nelson formalized the logic: firms rationally underinvest in upstream science because they capture only a sliver of the returns. The rest accrues to society at large. These are the “social returns” that make public R&D so valuable — and so are under-provided in a pure market system.

To make sense of public science funding, it helps to understand how research is classified. The NSF divides R&D into three categories:

• Basic research seeks to advance knowledge without a specific practical application.
• Applied research aims to solve a defined problem or develop a product.
• Experimental development turns ideas into commercial tools or processes.

The pattern of investment is clear and exactly what we would expect. Private companies overwhelmingly focus on development, directing just 6 cents of every R&D dollar to basic research. In contrast, public non-defense R&D allocates 34 cents of every dollar to basic research that maintains the upstream innovation pipeline that the private sector depends on but won’t fund itself.

While the private sector has stepped up its role in basic research, market incentives still limit how much firms invest in ideas without immediate commercial payoff. Public funding has long anchored the early stages of scientific discovery. For decades, much of that funding for basic research has flowed through NIH and NSF.

The Macroeconomics of Scientific Research

Connecting government research spending to economic growth isn’t as simple as plotting the two on a graph. Funding for science tends to rise during recessions (as in 2009) or in response to global shocks (like the 1970s oil crisis). Productivity can lag for years behind innovation. And the unpredictability of politics adds noise to both. That’s why economists refer to this as an identification problem: it’s hard to tell what causes what, both in general and in a technical statistical sense.

Recently, two economists — Andrew Fieldhouse and Karel Mertens — cut through this identifiability problem. They built a new dataset of appropriations shocks — year-to-year funding changes to major science agencies like NIH, NSF, NASA and DOE that are unrelated to short-term economic conditions — by reviewing decades of congressional reports to identify moments when funding spiked or dropped for reasons that had nothing to do with the business cycle. Policy changes unrelated to the state of the economy were classified as exogenous and used as instrumental variables in an econometric analysis to estimate how much public R&D investment contributes to economic output, separating its impact from other forces like infrastructure spending or business cycles.

This approach allowed them to isolate the effect of non-defense research funding on long-term productivity growth.

One key result is an estimate of the elasticity of output to public R&D capital — in plain terms, how much the economy grows when government R&D spending increases. Fieldhouse and Mertens convert this elasticity into a rate of return by scaling it to the size of public R&D relative to GDP. The result is staggering: depending on assumptions, the implied gross return ranges from 140% to 210%. Crucially, they arrive at nearly the same result using a completely different method — a direct statistical estimate of how changes in R&D investment affect productivity growth over time. The convergence of these two approaches attests to the strength and reliability of their findings.

To ensure these results weren’t an artifact of their method, they tested the model dozens of ways: adjusting lags, varying assumptions and checking instrument strength. The results held. Government investment in basic science doesn’t just generate ideas. It produces real economic growth.

How Research Fuels Prosperity

What do taxpayers get when the federal government invests in science? According to Fieldhouse and Mertens, the gross return on non-defense public R&D ranges from 140% to 210%, depending on model assumptions. The estimate from their preferred model is 171%. For every dollar spent, the economy eventually produces an additional $1.71 in output. That return exceeds nearly every other form of public or private investment.

By comparison, the gross social return on private R&D investment is about 55%, according to a 2013 study by Stanford economics Professor Nick Bloom and coauthors. Public infrastructure investments yield much lower returns — about 9.2% after accounting for depreciation, according to the Congressional Budget Office. “The estimated returns,” conclude Fieldhouse and Mertens, “substantially exceed those for public infrastructure, implying significant misallocation of public capital.”

Productivity That Lasts

The effects aren’t immediate. Like planting an acorn, the benefits of R&D take time to grow. Fieldhouse and Mertens estimate that the economic impact of public science takes eight to 10 years to fully materialize. That’s the delay between Congressional appropriations and measurable increases in total factor productivity, which is the economy’s output not explained by labor or private-sector capital. Economists think that productivity growth reflects technological progress and greater know-how — which is what drives long-run growth.

Using impulse response functions (a standard technique in time-series analysis) they show that after an initial lag, productivity begins to rise gradually, eventually peaking more than a decade after the initial investment.

Where the Productivity Comes From

The rise in productivity isn’t a mystery. In the years following a public R&D shock, researchers get to work.

“We show in the paper that government R&D kicks up a flurry of activity for both the inputs and outputs to the knowledge production function,” Fieldhouse told me. “You’re employing more scientific researchers. They’re publishing more technology books. At a slightly later lag, you’re seeing an increase in an index of economically valuable patents. We’re seeing an increase in the number of new STEM Ph.D. students — both an input to knowledge production, but also an output of sorts with some plausible lag. At a longer period of time, you see all sorts of innovative activity being spurred.”

The data bear this out. Fieldhouse and Mertens document clear spikes in research activity following funding shocks: rising STEM Ph.D. enrollments, more scientific publications and an increase in patents that are widely cited and cross-disciplinary — hallmarks of high economic value.

Importantly, public science doesn’t displace private R&D. It complements it. Fieldhouse and Mertens find that for every dollar the government spends on non-defense R&D, the private sector invests an additional $0.20 of its own. This “crowding-in” effect flies in the face of a common argument against government investment: that it steals market share or talent from the private sector. If that were true, we’d expect to see public spending depress private R&D, not amplify it.

The effects don’t stop with patents and private investment. Public R&D spending also prompts a wave of new investment in physical infrastructure — buildings, equipment and facilities — much of it at universities. For every $1 of federal non-defense R&D capital, the stock of non-defense structures rises by $1.46 over the next eight years. This is at least partly due to indirect costs: overhead on federal research grants that institutions use to build or maintain labs and equipment.

And the fiscal picture is just as strong. According to Fieldhouse and Mertens, the rate of return is so high that public science effectively pays for itself. Assuming a 171% return, the federal budget breaks even so long as it recovers just 9 cents in taxes per additional dollar of GDP — well below the historical average. That means science funding isn’t just good policy. It’s self-financing.

Why Politics Undercuts Public Science

Although economically wise, public investment is politically difficult. Timing plays a central role. As Fieldhouse puts it, “The benefits of public R&D arrive nearly twice an electoral cycle later — hardly an incentive to expand funding.” The economic gains take time to materialize. Fieldhouse and Mertens estimate that approximately 14 years passes between an appropriation shock and its peak effect on productivity growth. That’s too long for any single administration to take credit. It might be more politically expedient to cut funding and pocket the short-term savings.

Political and fiscal pressures have steadily eroded America’s commitment to public research. In real dollars, R&D spending has generally grown. But as a share of the U.S. economy, it has fallen sharply since its Cold War peak. Non-defense R&D, in particular, has failed to reclaim its historical role as a cornerstone of innovation-led growth. And once budgets are cut, they tend to stagnate relative to GDP — even after the crisis that triggered the reduction has passed.

In the short term, cutting research may seem like a fiscally responsible way to balance the budget. But that’s an illusion. Economists Klaus Prettner and Katharina Werner developed a macroeconomic model that incorporates education, fertility and the time it takes for research benefits to materialize to better understand the effects of public investment in basic research.

As they put it, the problem is that “raising public investments in basic research toward the optimal level reduces the growth rate of GDP and welfare in the short run because taxes have to increase and resources have to be drawn away from other productive sectors of the economy.” Yet these same investments yield substantial long-run gains. According to their analysis, the welfare-maximizing level of public R&D is much higher than what OECD countries typically spend. The short-term drag helps explain, they argue, why governments — and their currently living voters — underinvest in science despite its long-run payoff.

Could the private sector pick up the slack? Unlikely. Firms underinvest in basic research because they can’t capture all the benefits. Spillovers are a public good — good for society, but hard to monetize. As a result, economically productive ideas will go unexplored unless the government steps in. Empirically speaking, multiple lines of research suggest that dollar-for-dollar public investment in R&D is about three times as impactful as private R&D.

The result is a recurring mismatch between what science is worth and what politics will fund.

Science Is a Public Trust, Not a Political Football

Public investment in science is among the most productive uses of taxpayer dollars. The latest macroeconomic research finds that every dollar spent on non-defense public R&D generates between $1.40 and $2.10 in economic output. It spurs private-sector innovation, draws investment into infrastructure, and accounts for roughly a fifth of all U.S. productivity growth since World War II.

Cutting that investment, as proposed in the FY2025 budget, isn’t fiscally prudent. Indeed, it’s economically self-defeating. The Trump administration plans to cut $23 billion from NSF and NIH. Even using the most conservative end of the return estimates from Fieldhouse and Mertens (a 140% gross return), that reduction would eventually cost the economy at least $32.2 billion in lost output. Subtract the original investment, and the net economic loss is roughly $9.2 billion. At the midpoint return estimate of 171%, the loss grows to nearly $16 billion. Either way, cutting basic science doesn’t save money.

“If you had a single silver bullet for our long-run fiscal challenges — which are real — it would be faster productivity growth,” economist Andrew Fieldhouse told me. Public science, he argues, is the best way to get that growth.