Most people who lose Medicaid to the new work requirement are already working

Start with the two people this argument keeps invoking. One is heavily disabled; there is no honest version of “just get a job” for that person. The other is working, drawing a real paycheck, and still carries the card. Each side keeps its favorite in view and looks past the eligibility group that person is actually in. That is why the fight stays hot, and why it so rarely lands on what the rule says.

Federal floor, state hands

For years, this was optional. If a state wanted to make Medicaid recipients work for their coverage, it had to go to Washington and ask for special permission, and most never bothered. Ohio did get that permission back in 2019, then never turned the rule on before the next administration cancelled it in 2021. So it stayed a political argument, not something that pulled anyone’s coverage. The big 2025 federal budget law ended the choice. Work, or what the rule calls “community engagement,” is now baked into federal law as a condition of keeping Medicaid, in every state at once. Ohio and Pennsylvania do not get to decide whether to do this anymore. They only get to decide how.

Who is even on this kind of Medicaid

So who does this reach? Not everyone on Medicaid, which is the first thing the shouting gets wrong. It reaches one specific slice, the one Ohio files under the bland label “Group VIII,” the expansion group. Here is what that means in plain English. When the Affordable Care Act let states widen Medicaid back in the 2010s, Ohio and Pennsylvania both said yes, and saying yes opened the program to a set of adults who had never qualified before: working-age people, roughly 19 to 64, earning under about $22,025 a year for a single person, not sick enough to be on the old disability track and not old enough for Medicare. Before the expansion, a childless adult could be flat broke and still get turned away at the door. Those are the people the expansion covers, and they are the only ones this rule touches. Children, seniors, pregnant women, and people on the disability track are not in it.

What the states still control is how hard the rule bites: how they check the hours, how they decide who counts as too sick to work, whether they take the optional hardship exemptions, how fast they mail the letter, how workable they make the website. Ohio puts about 700,000 people in that expansion group; Pennsylvania about 750,000. Parents and caretakers of a child 13 or under, or of a disabled family member, are meant to be exempt, and people already meeting the work rules for food stamps or cash assistance can carry those across.

What counts as working

The qualifying activity, from CMS and Ohio’s own FAQ:

• 80 hours a month of work, community service, or a work program, or
• school at least half time, or
• a combination that adds up to 80 hours, or
• monthly earnings of at least 80 times the federal minimum wage, which works out to $580 in 2026.

Ohio’s minimum wage runs higher than the federal floor, so a person paid Ohio’s minimum clears the income test in fewer hours; Signal Cleveland walked the state’s math out to about 53. That single wrinkle is why one person hears “80 hours” and another hears “I already work enough, as long as the paycheck shows it.” Both can be reading the same rule correctly.

One state over: Pennsylvania

Ohio is not doing this alone, and its neighbor is a useful comparison. Pennsylvania runs the same expansion program under the same federal floor, with roughly 750,000 adults in its expansion group, a shade more than Ohio’s 700,000. The January 1, 2027 deadline is identical, and Pennsylvania’s Department of Human Services says outreach to enrollees begins by September 2026. What differs is how far along each state is on the parts it controls. Ohio has published its Group VIII target and signaled it does not plan to ask for the long delay. Pennsylvania, as of this writing, is still working out the details of its plan, which means its verification rules, the exemptions it does or does not adopt, and any delay request are not yet settled.

Both states argue about the size of the risk the same way. The Pennsylvania Health Law Project estimates about 310,000 Pennsylvanians could lose coverage, though that figure combines the new work requirement with the law’s move to twice-a-year renewals, not the hours rule alone. The Urban Institute, modeling the work requirement by itself, puts both Ohio and Pennsylvania in the group of states that would each shed more than 100,000 expansion enrollees. Read the two side by side and the pattern is the one the statute keeps proving: the federal law sets the minimum, and the state you live in decides how hard it is to meet.

How many lose the card

The honest answer to “how many get removed” is a range, because it depends on how cleanly a state runs the verification. The Congressional Budget Office estimates the work requirement will push about 5.2 million adults off federal Medicaid nationally by 2034, out of roughly 18.5 million subject to it each year, and leave about 4.8 million more people uninsured. Ohio’s slice, in the RAND model state analysts cite, runs on the order of 172,600 over the next decade. Pennsylvania’s advocates put its figure near 310,000, though that one folds the new twice-a-year renewals in with the hours rule. Every number here is a projection, not a turnstile count, and the gap between them is mostly a gap in guesses about how many eligible people the paperwork knocks off by mistake.

That last part is not a hunch. Arkansas is the one state that has run a Medicaid work requirement, back in 2018, and it is the closest thing to a real-world test we have. In about nine months, more than 18,000 people lost coverage, roughly a quarter of everyone subject to the rule, and a Harvard study found no measurable gain in employment. The reason is the uncomfortable one: nearly everyone targeted was already working or already qualified for an exemption. They did not lose the card because they refused a shift. They lost it because they never heard about the rule, or could not clear the reporting, and follow-up research tied that to skipped medications, delayed care, and new medical debt.

Who ends up losing the card

Take the person who genuinely cannot work: a body or mind that will not deliver 80 reliable hours of anything. If they already get federal disability checks, what is called SSI, and their Medicaid comes through that door, this rule is not aimed at them. If they landed in the expansion group because they never won a disability determination, they now have to prove frailty or lose the card, and winning that box is a form, a doctor, a timeline, and a state that may or may not be generous. The Health Policy Institute of Ohio notes that Ohio is passing up some optional hardship exemptions, including a clean out for high-unemployment counties. A RAND estimate cited in that analysis puts Ohio’s coverage loss on the order of 172,600 people over the next decade, with billions in federal funds moving with them. Treat that as a model, not a headcount; its direction is why disability advocates are loud.

For someone with a genuine disability who is not on SSI, the exemption is exactly where the difficulty hides. The statute exempts the “medically frail,” which covers disabling physical and mental conditions. But the rule leans on state data-matching to confirm exemptions automatically where it can, and medical frailty is one of the categories a database usually cannot confirm on its own. So the burden falls back on the person: a form, a doctor’s documentation, and a ruling from a state that may not be generous, filed not once but re-cleared every six months at the state’s regular check-in, with a cutoff notice waiting if a letter goes unanswered. That is the machinery Arkansas exposed. The people most likely to fall through it are not the able adult the speeches picture. They are the sick, the disabled who never won a formal determination, and the caregiver who assumed the exemption was automatic.

Now take the person who is working and still on Ohio Medicaid. Most often that is simply allowed, because expansion Medicaid is an income test and a full-time job at a low wage can sit under the line. Ohio also runs a separate door, the Medicaid Buy-In for Workers with Disabilities (MBIWD), which lets workers with disabilities ages 16 to 64 earn more and keep coverage by paying a premium. “They are taking advantage” is the charge that gets leveled here, and it is worth splitting in two. Sometimes the person is poor and insured, which is exactly the design. Sometimes a file is stale and an income change should have closed the case already, which is a verification failure rather than a character flaw. Work requirements do not catch stale files; verification does. And when verification is run clumsily, the red tape catches the tired and the sick first, which was the recurring result the last time a state actually ran this, in Arkansas. And the gaming is real, but the numbers keep it small: nationally, more than 90 percent of the Medicaid adults who could face this rule are already working or have a solid reason they cannot, according to KFF. The people taking advantage exist. They are a sliver, not the story.

Both speeches skip a sentence. The first skips how many people who should qualify for an exemption will fail the form anyway. The second skips that an income-tested program was always going to collide with a cultural argument about who deserves the card. That collision is why this stays contentious. It is not really an argument about 80 hours of math. It is an argument about who deserves the card, with the 80 hours standing in as the test everyone can point to.

Both pictures are real, and neither one describes 700,000 people, let alone the 750,000 next door. The person who cannot stand for a shift deserves the exemption the statute actually wrote, and the harder task is making sure the paperwork does not fail them anyway. The person already earning above the line was never the target of the hours rule; ordinary income eligibility handles that case, and always did. The heat comes from using either face to stand in for everyone, and that is the move worth dropping before the first letters land in Ohio and Pennsylvania alike.

Reported and edited to HeadLines Decoded newsroom standards. E.G. v4.17 · I.R.G. v1.14 · L.R.G. v1.9 · P.L.G. v1.10.