It started, like most things these days, with an online video. Late last year, conservative YouTuber Nick Shirley and his crew filmed themselves dropping by a handful of Somali-run daycares in the Minneapolis area that they suspected were fraudulent operations soaking up government money. Many of them did seem to lack the visible presence of children, and at least two have reportedly since closed — though state investigators said the centers were “operating as expected.” Three of the daycares later sued Minnesota, alleging that regulators had unfairly paused their operations in the wake of the viral exposé. The story, set in a city that could use a break from the national spotlight, offered a potent combination of murky details, immigration, and race.

It also drew attention to the very real issue of fraud plaguing government programs. Now, President Donald Trump has the executive branch on the case. The Justice Department has consolidated its anti-fraud efforts into a new National Fraud Enforcement Division, and the president has christened JD Vance the “fraud czar.” Trump contends that the nation’s losses to fraud, especially in Democratic-run states, are so big that if Vance’s new task force succeeds, “we would literally be able to balance our American Budget.”

Yes, that’s an overstatement. The federal budget is fast approaching a $2 trillion annual deficit — this year, we’ll spend roughly $5,500 more per U.S. resident than we take in — and anti-fraud efforts will not come anywhere near bridging that gap. Don’t blame Vance if, like Elon Musk before him, he leaves the nation’s fiscal problems mostly unsolved.

But let’s consider what we taxpaying chumps are losing to fraud — likely hundreds of billions every year — and how basic oversight, stepped-up enforcement, and structural reforms might improve things.

* * *

If a program spends government money, then someone, somewhere, has figured out a way to steal from it. Shocking amounts are lost because of mundane lies, and in some cases innocent mistakes, by those claiming benefits. Other schemes are larger in scale, more infuriating, and sometimes even kind of funny.

Small-time cheats simply claim benefits they aren’t eligible for, hoping no one notices. The Internal Revenue Service, for example, struggles to determine whether people claiming the child and earned-income tax credits meet the eligibility requirements — especially whether children live with them for at least half the year. As a result, it loses gobsmacking amounts in improper payments annually. Based on audits of random samples of filers, the agency estimates that improper payments of these credits total nearly $20 billion, including instances of both intentional fraud and genuine confusion about the rules.

Individual recipients might also “repurpose” their benefits. It’s long been common for food-stamp recipients to circumvent the program’s food-only restriction by selling their benefit cards to friends for cash at a fraction of the face value. One step up the fraud ladder, a store might swipe food-stamp cards and give part of the deducted value to customers in cash. Another step up, an organized fraud ring was recently caught processing millions of dollars in fraudulent food-stamp transactions through electronic benefits transfer (EBT) terminals that participating businesses, including smoke shops, weren’t even authorized to have. The terminals were obtained through fraudulent applications, with help from an accomplice inside the government.

Indeed, there are many ways that a business — or a purported business — might falsely claim to be doing something the government is willing to pay for. During the pandemic, one enterprising tax preparer helped numerous clients fraudulently claim credits intended to help small businesses stay afloat or keep employees on the payroll, collecting a fee for himself in the process. The fraudsters tried to claim $170 million in credits, of which the IRS paid out about one-third. To top it off, the guy running the scam failed to report the proceeds on his own taxes.

These scams can run into a simple problem, though: When you ask the government to pay you for something, it sometimes asks you to prove what you used the money for, such as by identifying the people you provided services to.

Getting around this, for a while at least, might be as simple as making up some names. In another example of pandemic-era fraud, Minnesota’s Feeding Our Future scandal involved fraudsters exploiting a federally funded nutrition program to the tune of about $250 million by falsely claiming to provide meals for children. They submitted fake documentation to the state agency overseeing the program, including attendance rosters with names and ages that were recycled at multiple sites.

Other times, better evidence is needed. If you need verifiable identities, you can give people kickbacks for signing up as clients but not receiving services — as happened in a recent case in California, where taxpayers paid for hospice care for people who were not dying. Another work-around, seen in both daycare and home-based health care, is to use family members as clients. A 2019 Minnesota audit described one case in which a daycare center offered jobs to mothers in exchange for enrolling their kids, and then “paid the mothers to essentially watch their own children.”

Health-care providers have special opportunities here, given the enormous amount of spending that flows through government health programs and the difficulty of verifying whether services were warranted or provided. Recent cases handled by the Justice Department have involved a medical-equipment company that bribed doctors to sign off on $11.4 million in unnecessary products billed to Medicare, and a pharmacy that billed Medicaid almost $270 million for expensive, unnecessary medications while paying kickbacks for patient information and prescriptions. Meanwhile, the Government Accountability Office (GAO), a federal watchdog, recently submitted fake applications for Affordable Care Act plans and found that they were rarely caught. It also detected “at least 160,000 applications in plan year 2024 that had likely unauthorized changes by agents or brokers,” which suggests that these actors were inflating enrollments to collect subsidies. State Medicaid programs have also been found to pay for fraudulent autism-therapy providers — as Adam Omary and Jeffrey Singer detail in this issue — and for the care of dead people and ineligible prisoners.

What does it all add up to? Since an unknown amount of fraud is never detected, it’s impossible to say for sure. But the GAO estimates that the federal government alone lost $233–$521 billion annually between 2018 and 2022. That may not be enough to balance the budget, and the pandemic years were particularly fraud-heavy. But the midpoint of that range still amounts to roughly $1,000 per American every year.

Other estimates are even higher. Haywood Talcove, CEO of LexisNexis Special Services and LexisNexis Risk Solutions–Government, noted at a recent Senate hearing that the GAO numbers left out several key agencies. He suggested that the true figure could be closer to $1 trillion.

* * *

We can reduce fraud, but it won’t be easy or free. Anti-fraud efforts pose real trade-offs, including paperwork burdens, law enforcement costs, and more.

The fact that so many scams flourished during the pandemic is an excellent illustration. During that time, the government shoved a lot of money out the door quickly — and helped a lot of people at a very rough time economically. But the very speed and efficiency of this process inevitably served as a flashing neon sign advertising opportunities for fraud. It’s taken years to root out many of these schemes, and there’s little doubt that a good number of thieves simply got away with it.

In an ideal world, we’d leave fraudsters fewer programs to exploit by shrinking and simplifying the government. But what are some more realistic options? Just as fraud flourished during the pandemic when controls were loosened, it can be reduced if we slow things down and keep an eye on who’s getting money.

First, government programs should verify eligibility and screen out likely fraud by collecting rigorous documentation and checking it against other databases before handing out funds. It’s far easier to decline a bad application than to chase down money that’s already been paid out. In his testimony to the House Oversight Committee in April, OJ Oleka, CEO of the State Financial Officers Foundation, observed that “federal benefit programs such as Medicaid, [food stamps], and unemployment insurance still rely heavily on self-attestation or periodic checks rather than real-time verification of income, residency, and citizenship status before payments are issued.”

Fraudsters extracted tens of billions of dollars from California’s unemployment system during the pandemic, in part, the state’s auditor found, because the system “did not block addresses used to file unusually high numbers of claims, and it removed a safeguard preventing payment to individuals who had unconfirmed identities.” Recipients included an applicant named “Poopy Britches.”

Similarly, in a recent paper for the Niskanen Center, Josh McCabe highlighted some options for requiring parents to document their eligibility for child-related tax benefits before receiving them, instead of claiming tax credits and providing documentation only upon being audited. The current Canadian system and an IRS pilot project from the 2000s prove that such a system can be practical and effective. The simplest approach is to have parents document with whom the child is living at birth and then require new documentation for any changes (rather than, for example, collecting documentation of living arrangements every year).

Second, those being paid for services they claim to provide should always be subject to random inspections or audits to confirm that they are keeping up their end of the bargain. In some cases, a level of verification can even happen automatically. Oklahoma, for example, tracks and compensates attendance by having parents swipe a card when they bring their kids to subsidized daycare.

Meanwhile, other states don’t seem to want to identify certain types of fraud. As my Manhattan Institute colleagues Christopher Rufo and Kenneth Schrupp recently noted in City Journal, this is evident in California. Providers of subsidized in-home care — often family members of the individuals being cared for — “self-report their timecards and check-in records,” while “state regulations prevent [regulators] from making unannounced visits unless a whistleblower files a ‘specific’ complaint or regulators identify clear red flags.”

Third, governments at all levels should use the best technology and databases available to detect potential fraud and should investigate cases aggressively. In a 2023 report from the GAO featuring a survey of federal agencies, the top issues — described as great or moderate challenges by the vast majority of agencies — centered on having sufficient staff, expertise, and data to find likely fraud. Artificial intelligence may provide new and efficient ways of reviewing large amounts of data for signs of fraud.

The Justice Department’s new focus on fraud and its establishment of a data-driven National Fraud Detection Center are steps in the right direction. Beefing up the resources of auditors and watchdogs such as the GAO, offices of inspectors general, and state financial officers could also pay dividends, as could better data-sharing across programs, agencies, and levels of government.

And fourth, when states administer federally funded programs, such as Medicaid, food stamps, and unemployment insurance, the federal government should insist on or directly enforce much higher standards to counteract the warped incentives states face. At their best, joint federal-state efforts can combine the strengths of both levels of government; more frequently, they encourage states to hoover up as much money as they can with little concern for how it’s spent.

Beyond ramping up enforcement and oversight, federal policymakers might rethink such programs on a fundamental level, if not by shrinking them, then at least by requiring high-fraud states to pay more of their own costs. Last year’s One Big Beautiful Bill Act jump-started this conversation by requiring states with high payment-error rates in the food-stamp program to pay for more of their spending starting later this decade. Time will tell whether this reform will prove effective.

Fraud is everywhere, and it will always be thus. Criminals go where the money is flowing, whether that’s private businesses or government programs, blue states or red. But they also go where no one’s watching — and they shouldn’t be anywhere near as successful as they are when it comes to bilking American taxpayers.