The Great American Welfare Heist
Part IV — The Math No One Wanted to Run
The estimates that follow necessarily involve ranges rather than fixed points. Participation rates vary, benefit levels differ by household, enforcement outcomes lag real-time behavior, and fraud attribution exists on a continuum rather than as a binary. Precision is neither possible nor required here.
The purpose of these figures is not to establish an exact total, but to determine whether the order of magnitude is survivable. Even under materially more conservative assumptions, the structural imbalance remains.
Structures Reveal Themselves Through Scale
By this point, motive no longer matters. Systems do not require intent to fail. They require arithmetic.
Fraud, when isolated, is criminal. When persistent and protected, it becomes structural. Structures reveal themselves through scale, and numbers remove intent from the argument. They do not care who meant well. Instead, they show whether a system can survive the behavior it permits.
Minnesota did not experience a scandal. It experienced an accumulated imbalance.
Let’s start with inputs.
The Somali diaspora in the United States numbers roughly 260,000 people, with approximately 108,000 in Minnesota according to recent Census estimates. Workforce participation reaches around 70 percent in Minnesota—higher than some national averages—yet average household incomes remain in the $30,000–$45,000 range. Applying a combined effective tax rate of 12–20 percent (reflecting lower brackets and state burdens), annual tax contributions from Somali households nationally remain in the hundreds of millions.
That is the optimistic ceiling.
Now consider outputs.
Participation in means-tested programs among Somali households in Minnesota stands unusually high. Analyses from Census-derived data place usage of at least one major program at 81 percent overall for Somali immigrant households (89 percent with children), including 73 percent with Medicaid access and 54 percent receiving SNAP. Medicaid, SNAP, housing assistance, childcare subsidies, SSI, and refundable credits dominate expenditures.
Conservative estimates place average annual public assistance costs near $15,000 per recipient when healthcare, housing, and tax benefits or returns are factored in. Applied to high participation rates across the population, baseline assistance exceeds $1 billion annually in Minnesota alone—before scaling nationally or adding enforcement and incarceration overhead.
Minnesota is not representative. It is illustrative.
The balance is inverted.
This framing still understates the failure, because it treats fraud as marginal rather than multiplicative.
Federal prosecutors, led by First Assistant U.S. Attorney Joe Thompson in late 2025, indicated roughly $18 billion flowed through fourteen high-risk Minnesota programs since 2018, with half or more—approximately $9 billion—suspected fraudulent. This excludes standalone cases like Feeding Our Future and reflects coordinated billing for nonexistent services, kickbacks, and diversions.
Fraud does not merely extract money. It reshapes the system.
Each stolen dollar reduces available aid, forces program freezes, increases borrowing, expands enforcement costs, and erodes political tolerance for the safety net itself. Minnesota froze child nutrition reimbursements statewide after oversight collapsed under abuse. Legitimate recipients absorbed the damage.
Then there are secondary losses.
Illicit funds routed abroad—often through informal remittance systems—represent permanent capital flight. That money does not circulate locally, does not generate taxable activity, and does not support domestic infrastructure. When such funds transit through regions with security risks, downstream costs emerge: intelligence operations, security spending, diplomatic pressure, and aid overhead.
Conservative estimates place these indirect costs at multiples of the direct fraud loss. A multi-billion-dollar fraud problem becomes a far larger long-term burden.
Duration compounds magnitude.
Welfare systems are designed as transitional mechanisms. Most refugee populations stabilize within a decade. Within twenty years, they are typically net contributors. This pattern holds across nearly every major resettlement cohort—except where insularity disrupts integration.
In Minnesota, data indicate persistent high dependency among Somali households well beyond ten or twenty years.
This is where the refugee designation becomes operationally relevant.
A refugee, by definition, cannot safely return. Yet travel back to Somalia occurs among Somali-Americans once citizenship allows—visiting family, investing, and conducting business.
If return is safe, the classification no longer holds. What remains is economic migration operating inside systems designed for temporary refuge. Because when policy is built on definitions and those definitions stretch, costs escalate.
No malice is required.
Only three conditions are necessary:
High-trust systems.
Weak or politicized enforcement.
Internal loyalty structures that discourage reporting.
Under those constraints, abuse is rational behavior.
The system behaved as designed. Oversight failed not because warnings were absent, but because acknowledging them carried institutional risk. And the damage extends well beyond money.
It has corroded legitimacy. And legitimacy, once broken, is more expensive to restore than any program line item. Because now the taxpayer doesn’t trust the stewardship of their remittances.
In the next section, we examine how institutional silence—not individual wrongdoing—allowed this arithmetic failure to compound year after year, until the denial became unsustainable.