Federal Judge Strikes Down Trump Administration's Retroactive Grant-Cutting Loophole
In a major blow to the Trump administration's sweeping campaign to claw back and defund federal programs, U.S. District Judge Indira Talwani ruled on July 17, 2026, that federal agencies cannot retroactively terminate previously awarded grants simply because they conflict with the administration's new policy priorities1. The ruling marks a major victory for a coalition of 21 state attorneys general and three governors—led by New Jersey, Massachusetts, and New York—who challenged the administration's aggressive use of a 2020 Office of Management and Budget (OMB) regulation to bypass congressional spending mandates.
The Ruling: Checking the "Slash-and-Burn" Campaign
Since taking office, the Trump administration has executed what states described as a "nationwide slash-and-burn campaign," terminating billions of dollars in federal grants for programs ranging from school lunches and mental health supports to scientific research and clean water. To justify these cuts, federal agencies relied on a revised OMB regulation (the "Termination Clause") stating that a grant may be terminated if the award "no longer effectuates the program goals or agency priorities."
However, Judge Talwani of the U.S. District Court for the District of Massachusetts rejected this interpretation, granting summary judgment to the states and denying the federal government's motion to dismiss. She ruled that the clause only permits termination if an awardee fails to meet the specific goals established when the grant was originally awarded, not because a new administration has changed its political preferences. Judge Talwani wrote:
"Defendants' interpretation of the Termination Clause is not clearly supported by the text of the provision, runs counter to the regulatory scheme, receives no support in the rulemaking history, and would violate the Spending Clause's requirement that conditions be imposed unambiguously."
New Jersey Attorney General Jennifer Davenport, who co-led the lawsuit, celebrated the decision as a critical check on executive overreach:
"The President and his allies cannot hold critical programs hostage to their personal whims and political ideologies, destabilizing the country by yanking essential federal funding that was already awarded to the states."
Broader Context and Parallel Education Funding Battles
This ruling establishes a powerful judicial barrier against the administration's efforts to exert political control over federal spending. It directly impacts several parallel legal battles across the country:
- School Mental Health Grants: A coalition of 15 states led by Washington State filed a "protective" lawsuit on July 10, 2026, to block the Department of Education from terminating $1 billion in school-based mental health grants under the Bipartisan Safer Communities Act. The Department had attempted to cancel these awards using the same shifting-priority justification.
- Education Research Funds: In the same Massachusetts federal court, a separate lawsuit filed on June 30, 2026, by the National Center for Learning Disabilities (NCLD), the Knowledge Alliance, and the Massachusetts Teachers Association challenges the withholding of $1.9 billion in education research funds. That lawsuit alleges that the OMB's freeze on Institute of Education Sciences (IES) funding violates the Administrative Procedure Act and the separation of powers.
By limiting the executive branch's ability to retroactively defund programs authorized and funded by Congress, Judge Talwani's decision reinforces the constitutional separation of powers and shields billions of dollars in state-level safety net, educational, and scientific programs from unilateral executive cancellation.
-
An instance of Unilateral executive efforts to withhold, redirect, or bypass legislatively mandated program funding face swift judicial blocks. — A federal judge blocked the administration from retroactively canceling active grants based on moving partisan priorities rather than program failures. ↩︎