The federal government built a system designed to stop fraudulent payments before they go out the door. It cross-references applicant data against death records, incarceration databases, and other federal watchlists. It is free for state agencies to use. And yet, most states have simply ignored it.
The tool is called Do Not Pay, and as City Journal detailed in a recent analysis, it represents one of the most straightforward, lowest-cost fraud prevention measures available to government at every level. The problem is not technology. The problem is political will.
Do Not Pay is a federal data-screening platform maintained by the Bureau of the Fiscal Service within the U.S. Department of the Treasury. It works by running payment recipients against a set of databases before money leaves the government's hands. Those databases include the Social Security Administration's Death Master File, state and federal incarceration records, the System for Award Management exclusion list, and other sources that flag individuals or entities ineligible for government funds.
The concept is simple: check whether someone is dead, in prison, or otherwise disqualified before cutting them a check. Federal agencies have been required to use the system since 2013 under the Improper Payments Elimination and Recovery Improvement Act. But state participation remains voluntary, and most states have declined to opt in.
That voluntary status has created an enormous gap. State governments collectively administer hundreds of billions of dollars in federal funds each year through Medicaid, unemployment insurance, food assistance, and other programs. Without pre-payment screening, those programs remain vulnerable to the most basic forms of fraud: payments to dead people, payments to incarcerated individuals who do not qualify, and payments to entities already flagged on federal exclusion lists.
Federal improper payments have been a persistent scandal for decades. The Government Accountability Office has repeatedly flagged the problem, and annual estimates of improper payments across federal programs have regularly exceeded $100 billion. A meaningful share of those payments flow through state-administered programs.
The fraud is not abstract. During the Covid-19 pandemic, unemployment insurance fraud alone cost taxpayers an estimated $191 billion, according to Department of Labor Inspector General figures. Much of that fraud exploited state-level systems that lacked basic identity verification and cross-referencing. States that had adopted stronger screening tools caught more fraud. States that had not adopted them became targets.
The pattern is familiar to readers who followed the Minnesota meal fraud scheme that resulted in a nearly 42-year federal sentence for its ringleader. That $250 million fraud exploited a federally funded nutrition program administered at the state level. Basic cross-referencing of claims data could have flagged the scheme far earlier.
The reasons states give for not adopting Do Not Pay tend to fall into a few categories. Some cite data-sharing concerns, arguing that connecting state systems to federal databases raises privacy issues. Others point to bureaucratic inertia: the tool exists, but no one in state government has been tasked with implementing it. A third group objects on ideological grounds, arguing that pre-payment screening creates barriers for legitimate recipients.
That last argument deserves scrutiny. Screening a payment against a death record does not create a barrier for a living person. Checking whether a benefit recipient is currently incarcerated does not harm someone who is not in prison. These are binary checks against objective databases, not subjective eligibility reviews.
The privacy objection is similarly thin. The data in Do Not Pay already exists in federal systems. State agencies that administer federal programs already handle sensitive personal information. The tool does not collect new data; it cross-references existing records to flag payments that should not go out.
What remains, stripped of the policy language, is a reluctance to prioritize fraud prevention. In some cases, that reluctance reflects a political calculation: aggressive fraud screening can reduce enrollment numbers, and reduced enrollment can become a talking point for critics who claim a state is cutting benefits.
The Department of Government Efficiency, known as DOGE, has made improper payments a central target. Elon Musk's team has pushed for broader adoption of existing fraud-prevention tools across federal agencies and has pressured states to participate in screening programs they have historically ignored.
That pressure has drawn predictable opposition. Critics of DOGE argue that its approach is too aggressive, that it risks cutting off legitimate recipients, and that fraud prevention should not come at the expense of program access. But the Do Not Pay tool is not a blunt instrument. It flags payments for review; it does not automatically deny them. The distinction matters.
Federal watchdog agencies have reinforced the case for broader screening. As one financial watchdog report found, fraudsters exploited Biden-era policies to steal billions in taxpayer funds, often through programs that lacked the most basic verification steps. The Do Not Pay system addresses exactly that gap.
For a state to adopt Do Not Pay, the process is not complicated. The Bureau of the Fiscal Service provides the platform at no cost to state agencies. Integration requires connecting state payment systems to the federal screening tool, a technical lift that varies by state but does not require new legislation in most cases. The Treasury Department has offered technical assistance to states willing to participate.
States that have adopted the tool or similar screening measures have reported measurable results. Pre-payment checks catch fraudulent claims before money moves, which is vastly more efficient than trying to recover funds after they have been disbursed. Recovery rates on improper payments are notoriously low. Prevention is cheaper than pursuit.
The law enforcement side of fraud recovery illustrates the difficulty. When the FBI tracked down a $1.2 billion Medicare fraud fugitive hiding in the Philippines, the arrest was a victory, but the recovery of stolen funds remained uncertain. Catching fraudsters after the fact is expensive, slow, and often incomplete. Stopping the payment before it goes out costs almost nothing by comparison.
There is a reason fraud prevention has become a conservative priority and a progressive discomfort. The left has spent years framing benefit programs primarily through an access lens: the goal is to get money to as many people as possible, as quickly as possible. Screening tools, by definition, slow that process, even if only slightly.
Conservatives have argued that protecting taxpayer dollars from fraud is not in tension with serving legitimate recipients. It is a prerequisite. Every dollar stolen by a fraudster is a dollar unavailable to someone who actually qualifies. Fraud does not just cost taxpayers; it undermines the programs themselves.
That argument gains force when the fraud is brazen. Cases like the Green Bank co-founder sentenced to 14 years for defrauding lenders and investors show what happens when institutions fail to screen and verify before disbursing funds. The pattern repeats across sectors: lax front-end controls, massive losses, expensive back-end prosecutions.
The Do Not Pay tool does not solve every fraud problem. It will not catch sophisticated identity theft or collusive billing schemes. But it catches the easy fraud, the kind that should never succeed in the first place: payments to dead people, payments to prisoners, payments to entities the federal government has already flagged as bad actors.
The barrier is not technical and it is not financial. The tool is built. It is free. It works. The barrier is institutional. State agencies that have operated without pre-payment screening for years have no internal champion pushing for adoption. Budget offices do not count fraud they never detected as a cost. And political leaders who might mandate adoption face no immediate penalty for inaction.
That calculus changes only when taxpayers and legislators demand it. Federal mandates could force the issue, and some in Congress have proposed requiring states to use Do Not Pay as a condition of receiving federal program funds. DOGE has signaled support for that approach. Whether Congress acts remains an open question.
In the meantime, every state that declines to run a free check against a death database before mailing a benefit payment is making a choice. It is choosing to tolerate fraud it could prevent at zero cost.
When the tool exists, when it is free, and when the fraud is measured in billions, the refusal to use it stops looking like caution and starts looking like complicity.
