Why Treasury Needs Tax Data to Protect Taxpayer Dollars

The Case for Limited, De-Identified Tax Data Use in Government-Wide Payment Integrity

The quiet debate now unfolding around the federal government’s Do Not Pay system is a referendum on whether the United States is willing to modernize how it protects taxpayer dollars in an era of industrial-scale fraud.

Photo by Sean Lee on Unsplash

In recent testimony and briefings to Congress, officials at the Department of the Treasury have been unusually candid. The government’s primary pre-payment screening tool—Do Not Pay— is underpowered, underutilized, and constrained by outdated statutory limits that prevent it from accessing some of the most reliable data the federal government already collects. Among the most consequential gaps is the inability to use tax data in any meaningful way to validate identity, income, or eligibility before payments go out the door.

This is not because Treasury lacks ambition or technical sophistication. It’s because the legal framework governing tax data use was built for an era in which fraud moved slowly, analytics were primitive, and data sharing itself was viewed as an inherent threat rather than a tool to be governed.

That world no longer exists.

Today’s fraud schemes are fast, coordinated, and digitally native. They rely on synthetic identities, fabricated income, and the exploitation of fragmented verification systems. Criminal networks understand that most federal programs still depend heavily on self-attestation and incomplete third-party checks. They also understand that the most authoritative source of income information in the federal government remains largely sealed off from the systems designed to stop fraud and improper payments.

The result is that the government pays first and asks questions later, while fraudsters move quickly through the gaps.

The Do Not Pay system was supposed to change that dynamic. In concept, it is a government-wide capability that allows agencies to screen payments against authoritative data sources before money is disbursed. In practice, its effectiveness is limited by what it is legally allowed to see. Treasury officials have acknowledged that without stronger authority to verify key identifiers and financial signals, Do Not Pay cannot fulfill its preventive mission. It can flag some obvious issues, but it cannot consistently detect the income misrepresentation and identity inconsistencies that drive some of the largest losses.

This is where tax data matters.

Tax records represent the most comprehensive and independently verified picture of income and economic activity the federal government possesses. When used appropriately, they offer precisely the kind of grounding signal that fraud prevention depends on. Yet under current law, particularly the strict confidentiality regime embedded in the tax code, even de-identified or analytical use of this information for payment integrity purposes is largely off limits.

That restriction is often defended as a necessary protection of taxpayer privacy. Privacy concerns are real, and they must be treated seriously. But the debate has become trapped in a false binary, as though the only options are unrestricted access or total isolation. That framing is outdated and unhelpful.

Modern fraud prevention does not require Treasury to inspect individual tax returns or to identify specific taxpayers. It requires the ability to test whether claims made to federal programs are plausible, consistent, and evidence based. There are well-established, privacy-preserving ways to do this without exposing sensitive personal information or repurposing tax data for enforcement.

De-identified and aggregated signals can confirm whether reported income falls within credible ranges without revealing exact amounts. Privacy-preserving record linkage techniques allow datasets to be compared without ever revealing identities in the clear. Federated query models allow Treasury to ask narrowly defined questions of tax systems and receive simple validation responses, without any raw data leaving its secure environment. These approaches are already used in other sensitive domains, including health data, census operations, and financial crime analysis.

Crucially, all of this can be governed by strict purpose limitation. Access can be legally confined to payment integrity and fraud prevention. Use for tax enforcement can be explicitly prohibited. Robust auditing and penalties for misuse can be built into statute.

This is not a call to weaken privacy protections; it is a call to modernize them so they reflect both technological reality and public risk.

Opponents of expanded access often warn of mission creep, and the concern is warranted. But by refusing to adapt, the government effectively outsources innovation to fraud networks while forcing agencies to operate with incomplete information. The status quo does not preserve trust. It erodes it, as repeated fraud failures undermine confidence that public funds are being managed responsibly (see, for example, Minnesota).

Treasury is already investing in better analytics and exploring the use of advanced tools to improve Do Not Pay. But those efforts will only go so far if the most relevant data remains inaccessible by law. Artificial intelligence can’t compensate for structural blind spots. Data science can’t conjure signals that the law forbids agencies from using.

The question before Congress is not whether tax data should be opened indiscriminately. It is whether carefully governed, privacy-preserving analytical access can materially improve the government’s ability to prevent fraud before money is lost. That question deserves a serious, evidence-based answer, not reflexive resistance rooted in decades-old assumptions.

A sensible first step is transparency. Requiring Treasury to formally assess whether de-identified and aggregated tax data would enhance government-wide payment integrity, to identify the precise statutory barriers that prevent such use, and to recommend legislative options that preserve taxpayer privacy while enabling fraud prevention is not radical. It’s a long-overdue first step in reforming privacy laws that have not kept pace with technology advancements and criminal sophistication.

Fraud prevention in the modern era must be about denying bad actors the opportunity to steal taxpayer money in the first place, not trying to claw it back after the fact. Treasury cannot do that while blindfolded. Expanded, carefully constrained access to tax data is not a threat to civil liberties. It is a necessary evolution in how the government protects both taxpayers and the public purse.

Doing nothing is a policy choice, and it is one fraudsters are happy to exploit.