The Trump administration has settled on a strategy: frame efforts to withhold congressionally appropriated funds as fraud prevention, and hope that hard questions about evidence, process, and legal authority disappear. They don’t. Federal law has well-developed procedures for how agencies must actually combat fraud. The administration is conspicuously bypassing them. That bypass is not incidental: it is both the tell of pretext and the source of the administration’s growing legal vulnerability. And courts are beginning to say so.
While evidence shows some amount of fraud in federal programs—a problem that administrations across time have taken steps to address—there are a lot of reasons to be skeptical that the Trump administration is earnest in its recent efforts. For one thing, President Trump has pardoned major fraudsters who bilked taxpayers and consumers out of billions. For another, the administration has dismantled the parts of the federal government best-equipped to root out fraud; for example, the twenty inspectors general that President Trump has fired or demoted identified more than $50 billion in waste and abuse in the 2024 fiscal year alone.
Most importantly, the president has wielded the power of the federal government to punish his perceived political opponents. But when the government has been as candid as the president about targeting political opponents, courts have intervened. So the administration has turned to a new approach: invoke “fraud” as grounds for freezing an entire state’s Medicaid funds, redesigning SNAP recertification in the middle of a benefit cycle, and withholding cash assistance for low-income families with children only in Democrat-led states, all while claiming responsible stewardship rather than political retaliation. Hundreds of billions in federal funding for Medicaid, SNAP, and family assistance programs flow annually to state and local budgets and, most importantly, to the individuals and families that rely on the programs for food, healthcare, and shelter. The administration’s effort to use these funds as a political weapon should fail.
Fraud is a genuine problem in federal programs, but invoking it does not suspend the statutory procedures that govern how agencies must investigate, document, and remedy it. Federal law has well-developed procedures for how agencies must address fraud. They require evidence, process, and proportionate remedies. These include requirements that agencies identify a credible basis for suspecting fraud before pausing funds, provide notice and an opportunity to respond, and impose penalties proportionate to findings, not blanket freezes. They are tools that administrations of both parties have used effectively. For example: the Medicare Fraud Strike Force, launched under President George W. Bush, has charged thousands of defendants and recovered tens of billions of dollars doing it the right way: working across government agencies and with state counterparts to mine claims data, name bad actors, and identify and prosecute fraudsters. The Trump administration, too, has produced record-breaking results using that model. And Acting Attorney General Todd Blanche also recently announced the creation of DOJ’s new Fraud Division, and called it a necessary effort to “zealously investigate and prosecute” fraud and better coordinate enforcement across agencies. But the vice president made plain that the White House will exert unprecedented political control over the division and recent administration actions reflect a different pattern: limited evidentiary grounds for actions, unequal application, and broad funding freezes as a go-to bludgeon.
Fraud as Purported Justification
In 2026 the Trump administration has accelerated its strategy of claiming that fraud justifies administrative action. On March 16, 2026, President Trump issued an executive order creating the “Task Force to Eliminate Fraud” and appointed Vice President Vance to lead it. As relevant here, the order’s key provision encourages federal agencies to “develop appropriate controls that operate before funds are obligated or disbursed to prevent improper payments in Federal benefits programs, including by coordinating agency action to determine when ongoing fraud or potential fraud require proactively pausing certain types of funding until such controls can be established.”
The order directs agencies to develop fraud-detection controls before disbursing funds — a facially legitimate objective that prior administrations have pursued through the Inspector General system and interagency coordination mechanisms. But the order’s authorization for agencies to “proactively pause” funding based on “potential fraud,” with no evidentiary threshold specified, creates the same procedural opening the administration has already exploited in targeted actions against Minnesota and Colorado. The order feints at process, but it does not require it where a funding pause would have the administration’s intended retributive effect.
Although tangible results from the order remain to be seen (the Task Force’s plan is due in June), the policy directives in the order echo actions that the administration has already taken to target its political adversaries and cut benefits to eligible recipients in the name of fraud prevention. For example, after months of public disputes between federal and state officials about immigration enforcement, in January 2026 the Centers for Medicaid and Medicare Services notified Minnesota that it intended to withhold roughly $2 billion in future Medicaid funding due to alleged fraud in the program. In February, the agency paused an additional $259.5 million in 2025 reimbursements.
Reportedly in an effort to punish Colorado’s governor for not pardoning a prominent election denier, in December 2025, the Trump administration enrolled the state in a “pilot project” related to the Supplemental Nutrition Assistance Program (SNAP), without the state’s consent. Citing unevidenced concerns about nationwide benefits fraud, the administration’s brief letter to the state directed it to recertify, through an extensive process that could include in-person interviews of more than 100,000 households, the eligibility of SNAP recipients in its five most populous counties. The deadline for compliance was 30 days and the administration threatened “Colorado’s continued participation in SNAP” if the state failed to comply. And in January 2026, the administration attempted a similar freeze with respect to childcare funding programs in five Democratic-led states to address “potential” fraud in the programs. A federal district court subsequently blocked the pilot project, finding the administration had violated SNAP’s substantive statutory requirements and that the retaliatory context “gives the game away.”
How Litigants are Pushing Back
As the Trump administration invokes fraud to justify administrative actions, litigants — especially state attorneys general — are challenging those actions in court and finding early success on several legal theories. The fundamental theme connecting them all is that the administration has bypassed established procedures precisely because they require producing evidence the administration appears to lack. (It’s no wonder why the administration is attracted to fraud as a rationale: sometimes the threat or initiation of a fraud investigation can have its intended retributive effect without requiring the administration to provide substantiating evidence, especially if funds are paused in the meantime.)
Contrary to procedural statutes. Congress has specified procedures for combatting fraud that the Trump administration is ignoring. Whatever deference courts might extend to an agency’s factual findings concerning fraud, a threshold question is whether the agency followed statutory procedures for making those findings in the first place.
When the Trump administration bypasses those mechanisms it is vulnerable to a claim that its actions are contrary to law. For example, the statute creating the Temporary Assistance for Needy Families (TANF) program outlines an extensive set of procedures for investigating fraud and, if it is found, imposing proportionate punishments on states. But when the Trump administration sought to target “Democrat-led” states for punishment, it ignored those statutory procedures and instead announced that it would immediately freeze $10 billion in funding, including for TANF, across five states. After receiving boilerplate letters from the administration, the states quickly filed suit, alleging among other things that the administration did “not even purport … to be attempting to follow” the “complex statutory and regulatory scheme that provides specific processes for” seeking remedies. A federal district court in New York agreed and granted a preliminary injunction against the administration’s freeze, explaining that “[i]n imposing what was essentially a punishment” against the states, the administration “took none of the required steps mandated under statutes and regulations.”
Contrary to substantive statutes. Beyond whatever procedural statutes an agency might violate when manipulating federal programs to punish its opponents, it may also violate the substantive statutes authorizing the programs in the first place. For example, the statute establishing SNAP tasks the Department of Agriculture with “providing food assistance to raise levels of nutrition among low-income individuals” and “provid[ing] timely, accurate, and fair service” to SNAP participants. It also allows the Department to create voluntary pilot projects for states to participate in so long as any waivers of SNAP’s requirements in the program do not “den[y] assistance to an otherwise eligible household or individual.” But when the administration enrolled Colorado in the mandatory “pilot project,” a district court found that the administration had violated each of the substantive statutory provisions described above. The “pilot” would have made it impossible for Colorado to comply with the administration’s demands while maintaining access to the SNAP program for eligible residents.
Arbitrary-and-capricious decisionmaking: pretextual reasoning. Courts have concluded that the administration’s claims of fraud in state programs or other noncompliance are pretext for political and personal retribution campaigns. As the federal district judge in the Colorado SNAP “pilot project” case explained, “[t]he larger context,” including “taunting social media posts” and threats toward a state’s leaders “gives the game away.” A federal judge in Illinois suspected pretext where the administration “retroactively applied” a particular regulation “only to entities related to the City of Chicago and the City of New York” without “a justification for targeted enforcement of what should otherwise be a policy applicable to all … grants nationwide.”
As Governing for Impact has written elsewhere, the Administrative Procedure Act’s ban on “arbitrary” or “capricious” agency action prohibits pretext. The administration might claim it is combating fraud, but judges are “not required to exhibit a naiveté from which ordinary citizens are free,” so litigants might continue to highlight instances where the government offers contrived reasons to distract from its actual agenda—penalizing its opponents. Official statements about the administration’s actual motives can be useful evidence in pretext litigation, and the president has been unusually forthcoming. Just recently, President Trump announced Vice President JD Vance as the “‘FRAUD CZAR’” and charged him with tackling “‘FRAUD’ in the United States” (the scare quotes are the President’s), asserting that it is prevalent “primarily in those Blue States” led by “CROOKED DEMOCRAT POLITICIANS.” Courts have also been especially skeptical of the government’s official fraud rationale where it lacks substantial evidence for its claim.
Of course, alleging and ultimately proving pretext can be difficult. A district court in Minnesota recently declined to grant that state’s request for a preliminary injunction against the Trump administration’s deferral of Medicaid funds during a fraud investigation, rejecting—at least at this preliminary stage—each of the state’s arguments why the administration’s fraud rationale is contrived.
Other APA claims. Litigants have also successfully invoked the APA’s arbitrary-and-capricious standard when the administration has failed to offer reasons for suspecting fraud, evidence for its claims, justification for imposing a draconian punishment, or consideration of serious reliance interests engendered by the agency’s prior approach. With respect to that last requirement, states and program beneficiaries have made decisions in reliance on funding they had every reason to expect would continue, and the administration must account for the disruption its freezes would cause. Separately, the administration’s announcement of new policies and severe sanctions may be the kind of substantive legal changes that require notice-and-comment rulemaking under the APA.
Constitutional claims. Litigants have brought constitutional claims, some of which might be available especially where litigants can demonstrate that the administration’s fraud rationale is pretextual. The Spending Clause, for example, prohibits the federal government from “surprising [s]tates with postacceptance or retroactive conditions” on their participation in federal spending programs. In the Colorado SNAP case, the judge found that the state was not on notice that it would be forced to participate in pilot projects when it accepted SNAP funds. In addition, equal protection principles might apply where the federal government targets some program participants and not others, and the only plausible explanation for that line-drawing is political animus. Finally, the First Amendment’s protections could ground a challenge to the administration’s efforts to retaliate against protected speech or political association.
* * *
The administration’s fraud task force is set to deliver its action plan in June, and the pattern of justifying actions with claims of fraud is likely to continue. Litigants should not be deterred. The administration’s legal exposure is not incidental: it flows directly from its decision to bypass the statutory and administrative procedures that fraud claims require. Those procedures exist precisely to distinguish genuine enforcement from pretext and invoking fraud does not obviate these bedrock requirements. Courts have already identified the gap between the administration’s stated rationale and its chosen methods, and that gap is not closing.
Litigation against insincere claims of fraud and corruption also matters beyond individual cases. The administration’s actions reveal an approach that treats fraud not as a governance problem to be solved through evidence and process, but as a political instrument invoked against adversaries and abandoned when it would implicate allies. The stakes extend beyond any individual program or state. Americans who rely on federal programs to survive and thrive are left in the crossfire.
