China’s Best New Recruiter? The White House

20,000 scientists already fled. This new rule could make that look like a rounding error

Photo by Barbara Burgess on Unsplash

Here is a fact that should stop every American cold: Over the last fifteen years, so quietly that no major headline captured it, the United States became a net exporter of scientific talent to China. What does “net exporter” mean here?

It means more top scientists left the U.S. for China than came from China to the U.S. And these weren’t refugees or dissidents seeking safety. These were scientists. The very people who develop vaccines, design advanced computer chips, and map genomes voting with their passports and choosing to leave.

Between 2010 and 2021, nearly 20,000 scientists of Chinese descent who began their careers in the US left for other countries, including China, and by 2021 the proportion of those who left the US for China had increased to two-thirds. The trigger wasn’t subtle.

The migration accelerated 75% higher in three years, coinciding with the launch of the China Initiative in 2018 — a Trump-era Justice Department dragnet ostensibly built to stop espionage that instead convinced a generation of brilliant people that America no longer wanted them.

That was Act One. It was a mistake born of overreach. It was clumsy, but at least aimed at something real. Act Two, unfolding right now in the bloodless prose of a 412-page Office of Management and Budget filing, is not a mistake. It is a design.

The rule that says the quiet part in regulatory language

On May 29, the White House proposed a rewrite of federal grantmaking regulations touching roughly $1.2 trillion per year in federal grants and cooperative agreements, of which $179 billion goes to small entities.

Strip away the bureaucratic camouflage and the core provision is almost startling in its bluntness: every grant must “demonstrably advance the President’s policy priorities,” giving political appointees a veto over any science that conflicts with an administration’s ideology.

Peer review — the mechanism by which physicists judge physics and virologists judge virology — is not eliminated, exactly. It’s demoted. The rule explicitly downgrades peer review to an advisory role rather than a binding recommendation, while requiring senior political appointees to conduct a “pre-issuance review” of every discretionary grant, forbidden from deferring to peer reviewers or routinely ratifying their recommendations.

Sit with that for a second. A person whose qualification is loyalty, not expertise, will have final say over whether a glioblastoma trial, a CERN collaboration, or a hurricane-forecasting model gets funded, and is instructed, by regulation, not to trust the scientists who reviewed it.

The provision governing merit review is somewhat misleadingly titled, since it has key provisions that undermine merit review. Even the termination clause is engineered for maximum discretion: awards can be pulled mid-project if a political appointee decides they no longer serve the “national interest,” and agencies need only provide a brief written rationale; no finding of noncompliance or fraud is required.

The rule comes from OMB Director Russell Vought, lead architect of Heritage Foundation’s Project 2025 plan for the Trump administration. And when Rep. Rosa DeLauro confronted him directly, accusing the rule of subjecting research to “an attempt to subject all federal funding to a political litmus test,” Vought didn’t deny the mechanism. He insisted “OMB will not be making these determinations; we don’t have a formal role in reviewing every Notice of Funding Opportunity.” That is a distinction without comfort. The rule doesn’t require OMB to review every grant. It requires someone answerable only to the President to.

Why this Is the sequel, not a new movie

Give the administration this much: it learned from its own failures. The first-term and early-second-term attempts to control science funding — the proposed cap on indirect research costs, the mass grant freezes and terminations — ran headlong into federal courts, because they violated the statutes under which Congress had appropriated the money.

This rule is different. It is not an executive order or a one-off memo; it is a regulation, amending 91 parts across 256 different sections of the government-wide “Uniform Guidance” that has governed federal grants since 2014. Converting non-binding guidance into binding regulation is the tell.

Under the Uniform Guidance, which is not legally binding, agencies have discretion to interpret federal grant-making policies to meet agency-specific needs; by turning the guidance into regulations, which are legally binding, the OMB rules would remove that discretion permanently, government-wide, immune to the next administration’s opinion of it, unless it is unwound through the same laborious rulemaking process.

Thirty-one Democratic senators have called the maneuver what it is: an attempt to let the executive branch do by regulation what it could not do by decree. Their letter to Vought argued the proposal would create limits on funding based on undefined terms, making it impossible for grant recipients to comply, since Congress “has historically exercised” its spending power through statute, and the rule instead substitutes the President’s preferences for it.

Whether that argument prevails in court is a real question. This may well get enjoined, in whole or in part, the way earlier maneuvers were. But note the asymmetry the administration is exploiting: the legal fight will take years. The chilling effect is instant. A cancer researcher deciding tomorrow whether to spend three years designing a clinical trial doesn’t wait for the D.C. Circuit. She just doesn’t apply.

Why the accountability argument collapses on contact

Let’s steelman this, because the honest version of this column has to. Is peer review perfectly neutral? No. It has its own guild biases, its own fads, its own risk-aversion that punishes genuinely novel ideas in favor of safe incrementalism.

Political appointees overseeing how taxpayer trillions get spent is not, in the abstract, an outrageous idea. It’s arguably part of what accountable government is supposed to look like. And OMB’s own framing leans on this: it says the overhaul is needed to improve transparency, accountability, and oversight and to reduce “waste” and “abuse” of American tax money.

Here is where the argument dies. Accountability means checking whether the money was spent as promised, competently, without fraud. It does not mean substituting a political appointee’s judgment about which scientific questions are worth asking for the judgment of the people who’ve spent their careers learning to ask them.

The rule doesn’t add an inspector general. It adds an ideologue with veto power and immunity from the very expertise the inspector-general model would require him to defer to. That’s not oversight. It’s censorship wearing an oversight badge.

It’s the distinction Vought’s own defenders keep trying to blur when they cite “waste” as the justification for a rule that spends most of its 412 pages on content, not accounting: banning “anti-American values,” restricting international collaboration, capping travel to conferences, making publication costs presumptively unallowable.

The preamble relies heavily on Heritage Foundation reports, partisan Senate committee documents, and White House fact sheets, rather than independent scientific or administrative assessments. That is not the evidentiary record of a rule designed to fight waste. It’s the evidentiary record of a rule designed to fight ideas.

And should you doubt the “gold standard” language is really about scientific rigor, consider that Vought, pressed in committee for an example of where he’d have used this termination power, reached not for a wasteful cronyism case but for a claim that the NIH should have been allowed to rescind grants to Wuhan when its gain-of-function research “caused” the COVID-19 pandemic. A theory that has been debunked. The tell isn’t subtle. When asked to justify the power, the architect reached for grievance, not evidence.

The death by a thousand line items

The grant rule doesn’t operate in isolation. It’s the capstone on a budget strategy that has been remarkably consistent in its direction, if inconsistent in its success. The FY2027 request — the administration’s second consecutive attempt — again proposes gutting the two agencies most responsible for America’s basic-research base: a cut of nearly 55 percent to NSF, reducing the agency’s budget from about $8.75 billion to $4 billion, and a $6 billion, or 13 percent, cut to NIH.

NSF itself estimates the practical consequence: success rates falling from 19 to 8 percent and the number of awards dropping from 7,400 in FY25 to 2,900 in FY27. Translate that out of percentages: it means thousands of promising research proposals — vetted, scored, ready to go — simply die on the vine, not because they failed on the merits, but because the merits no longer determine the outcome.

Last year, Congress rejected an even harsher version of these cuts. It may do so again. But the point isn’t whether any single year’s budget survives contact with appropriators. The troubling part is that it has become a pattern. A researcher building a lab, training graduate students, and committing to a five-year study cannot plan around a coin flip repeated annually.

And it’s not just the money. Under the OMB rule, agencies aren’t required to give scientists a real path to contest a mid-stream funding cancellation, and advocacy groups warn this could restrict certain international collaborations and tighten requirements for disclosing foreign funding in ways that function as workarounds for court rulings that had limited politically motivated grant cancellations.

Uncertainty isn’t a side effect of this policy. Uncertainty is the policy. The fog machine that achieves through ambient dread what explicit censorship cannot achieve through direct order, because ambient dread is much harder to sue.

The China question nobody in Washington wants to answer

Here is the argument almost no one is making, and it should be the headline, not the footnote: this administration’s stated obsession is defeating China in the technology race. Its actual policy is a subsidy program for Chinese science.

We’ve run this experiment before, inadvertently. When the DOJ launched the China Initiative in 2018 — ostensibly to stop espionage — it instead functioned as a recruitment poster for Beijing. Scientists didn’t need China to make itself more attractive; they needed America to make itself repellent, and it obliged.

A Princeton-Harvard-MIT research team found that departures increased by 75% following the implementation of the China Initiative, and surveys of Chinese American scientists in this country found 35% feel unwelcome, 72% do not feel safe, 42% are fearful of conducting research in the US, and 86% perceive it is harder to recruit top international students now compared with five years ago.

These weren’t marginal figures at the periphery of American science, they were Fields Medal contenders, Harvard nanochemists, Berkeley neuroscientists, people whose names anchor entire subfields, walking across the Pacific and taking their labs, their grants, and their graduate students with them.

Now overlay the OMB rule and the FY2027 budget onto that same dynami. But this time apply it to everyone, not just researchers of Chinese descent. Restrict international collaboration. Make publication costs presumptively unallowable, throttling the very act of putting American discoveries into the global scientific record. Require political loyalty tests before a grant can be issued. Give a political appointee the power to cancel a five-year clinical trial on a whim, with no appeal.

Ask yourself what a 28-year-old postdoc — American, Chinese, Indian, it no longer matters — does when she’s choosing between a five-year fellowship in a country where her funding can vanish because an appointee doesn’t like her university’s mission statement, and a lab in Shenzhen offering stability, resources, and a government that treats her expertise as an asset rather than a security risk. China isn’t winning the twenty‑first century by out-thinking American scientists; it’s winning by watching America lose its capacity for rational thought.

What made America the superpower was the thing being dismantled

There is a founding myth of American science that is actually true, which is rarer than it sounds: the reason the United States became the world’s dominant scientific power after 1945 was not that Washington picked winners. It was that Washington built an architecture — Vannevar Bush’s National Science Foundation blueprint, the NIH’s peer-review study sections, decades of bipartisan deference to expert judgment — whose entire design premise was that politicians should not pick winners.

Merit review was not incidental to American scientific supremacy. It was the mechanism of it. Discoveries with no obvious political constituency — recombinant DNA, the internet’s TCP/IP protocols, mRNA platforms two decades before anyone needed a vaccine overnight — got funded precisely because no appointee had to certify that they advanced anyone’s policy priorities. They only had to be true, and worth finding out.

What the OMB rule proposes is not a correction to that system. It’s an inversion of it. And the inversion reveals something the administration would rather not say plainly, so it’s saying it in Section 200.205(b) instead: that it does not believe it can win an argument about which research deserves funding on the merits, so it is deleting the concept of merits from the equation. That’s not confidence. That’s a confession.

The ending isn’t written yet

None of this is inevitable. More than 80,000 public comments have already poured in opposing the rule; the legal challenges are coming and may well succeed, at least in part, just as the indirect-cost cap and the mass grant terminations were struck down before. Congress rejected the harshest version of these budget cuts last year and may do it again. The system has more resilience than its architects are counting on, and American science has survived stupider administrations than this one.

But resilience isn’t the same as immunity, and the damage that matters most here isn’t the damage a court injunction can undo. It’s the graduate student who chooses a different country. It’s the tenured professor who takes the Tsinghua offer instead of waiting out the uncertainty. It’s the five years of a research program that never happens because nobody could plan around the coin flip. Injunctions can restore a grant. They cannot restore a career decision already made, a lab already relocated, a decade of momentum already handed, freely and without a fight, to the one strategic rival that has spent years building the infrastructure to receive it.

America did not become the world’s scientific superpower because its politicians chose the experiments. It became the superpower because, for eighty years, it chose not to. That choice is now up for a 45-day public comment period. Whether the country understands what it’s actually voting on may be the only variable left that matters.