The Courage Was Free

When courts ordered a batch of illegally terminated NIH grants restored, the relief didn't land where the damage had. A STAT analysis found that $2.1 billion in reinstated funding flowed back to institutions in Democratic congressional districts. Republican districts got back $62 million. Not because the administration had targeted blue-state science with surgical precision, and not because red-state researchers had somehow avoided the original axe. The gap runs roughly thirty-four to one, and it has almost nothing to do with where the grants were. It has everything to do with who sued.

This is the piece the first installment didn't have room for. The Vought impoundment project keeps working partly because the Impoundment Control Act carries no real penalty for violating it. But the other half of why it keeps working is that the officials best positioned to use the tools that do exist have been choosing, quietly and consistently, not to use them against their own party's president.

That deserves more examination than it has gotten.

The Defense Contractor Theory

It has a solid track record, and it's not wrong about what it's measuring.

The standard model for how Congress checks executive overreach in the absence of partisan will is pork: spread enough federal money across enough districts, and members will protect the spending to protect themselves, regardless of which party holds the White House. The F-35 program achieved something close to political immortality by sourcing components across something like forty-five states. No member of Congress votes to kill the jobs in their own district, whatever they privately think of the plane.

NIH funding looks, at a glance, like it should work the same way. Every state has land-grant universities and public medical schools. Research money flows broadly enough that every congressional district has some stake in the system. And sure enough, when the administration proposed its sharpest cuts, Republican senators from states with serious research footprints said the right things. Susan Collins, the Republican chair of Senate Appropriations, cited Maine's research base and went on record calling rescinding appropriated funds without congressional approval a clear violation of the law. Katie Britt flagged the University of Alabama at Birmingham's roughly $90 million in annual indirect-cost reimbursements and broke with the administration on the indirect cost cap. Both were right. Both made the point publicly. Both helped deliver the floor votes that kept the deepest proposed cuts from landing in FY2026.

That's the mechanism working. The question is where it stops working, and why.

The Distribution Problem

NIH isn't the F-35, and it was never built to be.

The DOD component-sourcing strategy is deliberate political insurance, engineered from the start to create a broad enough coalition of stakeholders that no single Congress can kill the program. NIH funding follows peer review, not political geometry, and the actual distribution looks nothing like a defense procurement map. A handful of states with elite medical schools absorb a disproportionate share of the total, which is precisely why Congress had to create the IDeA program as a dedicated set-aside: roughly half the states get comparatively little NIH funding and would otherwise have next to no institutional stake in the system. The spread exists, but it's a byproduct of university geography rather than a lobbying strategy anyone engineered.

This changes the baseline in a specific way. If your state's NIH footprint is modest, the self-interest calculus is weaker going in, and the floor-speech calculation is easier precisely because there's less at stake.

The administration's own termination pattern made the red-state exposure concrete anyway. The first month of grant cancellations hit institutions across twenty-five red states, seventeen blue, and eight purple, with ninety-six institutions flagged in Florida, Ohio, and Indiana alone. Whatever logic was driving the targeting, it wasn't protecting the Republican coalition from collateral damage. The red-state research base took the hit alongside everyone else.

The Fork

Here is where the model breaks, and the gap explains itself.

Here's the number that deserves more attention than it has gotten: after the restoration orders, $2.1 billion went back to institutions in Democratic congressional districts. Sixty-two million went back to Republican ones. That is not a rounding error. It is a ratio, and the ratio has a cause.

Democratic state attorneys general went to court. They had standing, they filed claims, and they won injunctions that forced reinstatement of terminated grants within their jurisdictions. Most Republican state attorneys general, facing the identical harm at institutions in their own states, did not bring comparable suits. One legal scholar's assessment, cited in coverage of the disparity, was direct: red-state officials were reluctant to directly challenge the administration, even on behalf of their own institutions.

So here is what the self-interest model actually produces when you run it at the state AG level: the material incentive to protect your state's research institutions is real, and it holds right up until acting on it requires public confrontation with a president whose base is your base too. At that point, the incentive loses. Not always, not everywhere, but consistently enough to produce a thirty-four to one funding gap between states that sued and states that didn't.

The floor speech costs nothing. The lawsuit costs something. The record shows which one happened.

Why This Is the Other Half of Piece One

The mechanism only runs if nobody pulls the emergency brake.

The Vought impoundment project, documented in the first installment, depends on a specific precondition: the officials with real tools to stop it have to be unwilling to use them. Courts have shown they'll order reinstatement when asked. State AGs with standing have a proven track record of winning. The Congressional Power of the Purse Act sits in the House, drafted and ready, built specifically to close the pocket-rescission loophole and attach actual penalties to Impoundment Control Act violations. None of these tools are theoretical.

What the evidence shows is that each check works in precise proportion to how cheap it is to exercise. Collins and Britt can object on appropriations and take a floor vote without that vote showing up in a primary attack ad. A Republican AG from Florida suing the Trump administration over NIH grants faces a different calculation, and so far that calculation has almost universally produced the same answer.

Vought called the GAO violation findings rearview mirror stuff. That read as bluster at the time. It was bluster. It was also accurate, because he had already done the math on who would use the leverage they had and who wouldn't. The scoreboard is not close.

The Honest Accounting

Pork protects things. Principle protects things. The question is which one you are actually depending on at the expensive end.

The theory that distributed federal funding reliably checks executive overreach assumes two things: that the money is spread deliberately enough to create genuine cross-partisan stakes, and that self-interest is a stronger force than tribal loyalty when the two come into conflict. NIH funding is spread, but not deliberately. And the current evidence is pretty direct about what wins when the test is expensive: being on the wrong side of your own party's president, on the record, in a filing with your name on it.

The appropriators who went to the floor and held on FY2026 should get credit for that. It mattered. It kept the worst of the proposed topline cuts from landing. But the same appropriators would have to vote for a law that permanently limits a Republican president's spending authority, backed by real penalties, to close the door on this entirely. That's a different order of ask, and nothing about the current arithmetic suggests it gets there.

The checks that exist are working as hard as they're free to exercise, and no harder. The costly ones are sitting on the table, and everyone who could pick them up has found somewhere else to look.

This is the second installment in a Long Arc series on the structural mechanics of executive power and the tools that exist to check it. Part One: Traumatically Affected laid out the Vought impoundment project from its 2019 origins through its current form. Part Three is forthcoming.

Like what you've read, but not able to switch to paid? Why not buy me a coffee instead! All money goes to help defray the expense of running the site.