The White House Drafted a Veto Committee for NIH Grants. It Died in Four Days — and Section 157 Is Why.

The fastest-moving federal grant policy story of September 2026 lasted four days and produced no document anyone can cite. That is precisely why it matters.

On Friday, September 18, an Oval Office meeting turned to a complaint Office of Management and Budget Director Russell Vought has been making all year: that the National Institutes of Health continues to fund research misaligned with the administration's stated priorities. The remedy under discussion was an executive order creating a committee of political appointees — Vought among them, alongside NIH Director Jay Bhattacharya — that would hold final authority over NIH awards after those awards cleared statutory scientific peer review. Reporting indicated the committee would have been required to reach unanimous decisions. The same draft would have expanded agencies' ability to terminate existing grants "for convenience."

By Tuesday, September 23, Senate Appropriations Chair Susan Collins had sent a letter to Vought and Bhattacharya opposing it. By that evening the order was being described as shelved. Collins told reporters it was "looking dead." The White House has not formally withdrawn anything, which is the part applicants should sit with.

The Argument That Worked Was Not About Science

Plenty of scientific societies argued the order would corrupt merit review. That argument has been made continuously since early 2025 and has not previously stopped anything.

Collins made a different argument, and it landed. Her letter said the proposal "risks undermining the Congressionally-mandated peer review process that all NIH grants must undergo before they are approved for federal funding" — and then made the point with teeth: the draft order would replicate a provision Congress had already blocked three weeks earlier in Section 157 of the continuing resolution.

That is the whole story. Section 157 of the Continuing Appropriations Act, 2027 — carried on H.R. 6500, signed September 2, 2026 — bars OMB from issuing or finalizing its May 29, 2026 proposed Regulation for Federal Financial Assistance, or a substantially similar rule, through December 11, 2026. We covered the statutory text and its anti-workaround language when it passed. The May 29 rewrite of 2 CFR Part 200 contained two provisions that overlapped almost exactly with the NIH draft: political appointee pre-issuance review of discretionary awards, which would have made peer review advisory only, and a termination-for-convenience standard keyed to "Federal agency priorities, or the national interest as they exist at the time of the termination."

An executive order is not a rule under the Administrative Procedure Act. On a literal reading, Section 157 does not reach one. The draft order was, functionally, a route around a statutory roadblock — and the chair of the committee that writes NIH's appropriation said so in writing. Appropriators do not enjoy being routed around, and an appropriator with a markup pending has leverage that no scientific society has.

What Would Have Changed If It Had Issued

Understanding the counterfactual tells you what to prepare for, because the underlying policy goal has not moved an inch.

Peer review would have become a qualifying round. Study section scores would still determine which applications reach the eligible pool, but a political committee would decide which of those get funded. For applicants, this converts an application from a scientific argument into a two-audience document: one audience scoring significance, approach, and rigor; a second audience scanning for alignment with presidential priorities. Those two audiences reward different prose.

Award timelines would have stretched unpredictably. A unanimity requirement across a committee that includes the OMB Director is not a throughput mechanism. Every award becomes subject to a veto that requires no stated reason. NIH's obligation pace was already the slowest in modern memory; the Interior Department's political review threshold offers the template — a $50,000 trigger and two appointees produced a 1,400-grant, $362 million backlog.

Existing awards would have gotten less durable. The expansion of termination for convenience is the provision with the widest blast radius, because it touches money already obligated. Executive Order 14332, signed August 7, 2025, already directs agencies to insert termination-for-convenience clauses into new and amended discretionary awards "to the maximum extent permitted by law." A further expansion aimed specifically at NIH would have put multi-year R01s and program project grants on a footing closer to a federal contract than a grant.

The Real Date Is December 11

Here is the calendar that should be on every research administrator's wall.

Section 157's protection is coterminous with the continuing resolution. The CR funds the government through December 11, 2026. When it expires, so does the statutory bar on OMB issuing the Uniform Grants Regulation. Nothing in Section 157 restricts OMB's ability to finalize the rule after that date, and twelve appropriations bills remain unstaged in the Senate.

That produces three scenarios worth planning against, and only one of them is comfortable:

1. A full-year appropriation or new CR carries Section 157 forward. The pre-issuance-review and termination-for-convenience provisions stay dormant into 2027. This is the outcome the 105 organizations that lobbied for Section 157 are working toward, and it requires the provision to be re-enacted — it does not survive on its own.
2. A CR passes without Section 157. OMB can finalize the May 29 rule on its own schedule. Its original target was an October 1, 2026 effective date so that a single rule set would govern FY2027 awards; a December or January finalization simply moves that to mid-year, which is operationally worse for recipients than a clean fiscal-year transition.
3. The executive-order route returns in a narrower form. The NIH draft failed because it was broad, sudden, and traceable to a blocked rule. A version scoped to a single institute, or folded into an existing "gold standard science" implementation memo, draws less fire. Bhattacharya reportedly resisted this draft; that resistance is a personnel fact, not a structural protection.

What Applicants Should Actually Do

None of this argues for pausing submissions. It argues for building applications and budgets that survive either outcome.

Write the alignment paragraph, and write it honestly. Whether or not a political committee ever gets a vote, a pre-issuance reviewer at some agency is already reading your specific aims against a priorities list. A clear, non-euphemistic statement of what the project delivers to American patients, workers, or manufacturers belongs in the first half-page of the Specific Aims and in the Project Narrative field. This is not capitulation; it is answering a question that is being asked.

Assume termination-for-convenience language is in your next award, and price it. If your institution's subaward templates flow federal terms down verbatim, a convenience-termination clause reaching a subrecipient can strand a partner mid-year. Negotiate close-out cost recovery and notice periods in the subaward now, while you still control the paper.

Do not build a program on a single federal obligation. The practical lesson of 2026 is that a multi-year federal commitment is a schedule of intentions, not a contract. Bridge reserves, a second funder on the same aim, and staged hiring are no longer conservative — they are baseline.

Track the score-reporting change separately. NIH's proposal to replace overall impact scores with a three-tier reporting system is a distinct thread from this one, with an RFI comment deadline of October 13. If scores become coarser at the same time that post-review discretion widens, the information an unsuccessful applicant receives about how to improve gets materially thinner. That combination, not either change alone, is what should worry resubmitters.

Watch obligation pace, not announcements. NIH, NSF, AHRQ, and the Institute of Education Sciences all closed FY2026 with unobligated balances that tell you more about award probability than any press release. We laid out the numbers in the September 30 obligation endgame.

The Pattern Worth Naming

Three times in 2026 a sweeping grant-policy change has been announced, drafted, or proposed — and then blunted by a procedural objection rather than a substantive one. The May 29 rule drew more than 500,000 comments and was stopped by a rider on an Africa trade bill. The NIH order was stopped by a letter citing that rider. In both cases the policy goal survived the defeat intact.

For anyone who writes or administers federal grant applications, the operational conclusion is unglamorous: the rules governing your next award are being decided on a two-to-three-month cycle, and the deciding venue is appropriations language, not the Federal Register. Build the assumption of change into your budget narrative, your subaward templates, and your cash-flow model. Then keep submitting — because the one thing every scenario above shares is that agencies still have money they are statutorily obligated to spend, and applications still have to exist for that to happen.

Keeping a live read on which programs are actually obligating money, and which are quietly stalled, is the part Granted was built to shorten.