I believe work requirements attached to public benefits are a legitimate policy position, and I’m not going to pretend otherwise to score points with readers who disagree. The idea that able-bodied adults receiving government assistance should, where reasonably possible, be working, training, or contributing has real support across the political spectrum, including from plenty of people who aren’t conservative by any other measure. That’s not what this piece is about.
This piece is about what happens when a policy idea meets actual government execution, because the Government Accountability Office just handed us a case study, and the numbers are not close.
Georgia’s Pathways to Coverage program — the nation’s only active Medicaid work-requirement experiment — has, from 2021 through the first half of 2025, spent $54.2 million on administrative costs. In that same period, it spent $26.1 million on actual medical assistance for enrollees. Read that again. The state spent more than double what it spent covering people’s health care on the paperwork, IT systems, and bureaucratic apparatus required to verify whether those people were working enough hours to deserve coverage. About 88 percent of that administrative spending came from federal taxpayers — meaning this isn’t a Georgia problem alone, it’s a national accountability problem wearing a state’s name.
The GAO report is polite about it, in the way federal watchdog reports usually are: it notes that the federal approval process for state Medicaid experiments “does not take into account the extent to which demonstrations will increase administrative costs,” and that Georgia was never required to detail those costs before implementation. Translation: nobody built in a check on whether the mechanism for enforcing the work requirement would end up costing more than the benefit it was designed to gatekeep. Georgia officials attribute part of the overrun to a two-year legal fight with the Biden administration that forced duplicated spending on IT and staff training — a real factor, but one that should have triggered a serious re-evaluation, not a rubber-stamp extension.
Instead, in September 2025, the Centers for Medicare and Medicaid Services extended Pathways through the end of 2026.
I want to be precise about where the accountability failure sits, because it’s not where either party’s talking points want it to sit. This isn’t a story about whether poor people should have to work for benefits — that’s a legitimate debate to have on its own terms. This is a story about a government program that, regardless of your position on the underlying policy, is demonstrably failing at basic cost-effectiveness, and getting extended anyway with no apparent consequence for that failure. If this were a private contractor blowing through a budget two-to-one on overhead versus actual service delivery, it would be a scandal and a canceled contract. Because it’s a state agency implementing a politically popular policy framework, it got another year.
That’s the pattern I keep coming back to in government accountability stories, across administrations and across issues: the politics of a policy and the execution of a policy get evaluated on completely different timelines, and usually only the politics gets graded. Supporters of work requirements should be the first ones angry about this report, not the last — a badly executed version of your policy is the best gift you can hand to people who want to kill the idea altogether. If Pathways can’t demonstrate it’s getting more Georgians covered per dollar than an unconditional expansion would, defenders of the work-requirement concept need better evidence, not just another extension.
Wisconsin doesn’t currently run a program like this, but it’s exactly the kind of policy that gets proposed here on a two-year cycle, and it will keep getting proposed nationally as long as it polls well. The lesson from Georgia isn’t “work requirements are bad.” The lesson is “if you’re going to build a gatekeeping mechanism, you’d better be able to prove the gate costs less than what it’s keeping out” — and right now, nobody in Georgia or Washington can prove that.
— Marcus Hart