The Thirteenth Amendment abolished slavery in 1865, with one qualification. Slavery and involuntary servitude were prohibited except as a punishment for crime, and that phrase remains in the federal constitution today. It was not a drafting accident. The exception was debated at the time, and its consequences appeared quickly in the convict leasing systems that spread across the South in the decades that followed.
What it means in practice is that a state may require a convicted person to work, and may punish refusal. Most do. The question that has moved to state ballots over the past eight years is not whether prisons should offer work, but whether they should be able to compel it.
The Supreme Court has never squarely held that compelled prison labour violates the Constitution, and lower courts have consistently read the exception as permitting it. That leaves the question where it now sits: with legislatures and with voters, rather than with judges.
The most comprehensive count comes from a 2022 study by the American Civil Liberties Union and the University of Chicago Law School. Of roughly 1.2 million people held in state and federal prisons, about two thirds work. Their labour was valued at approximately 11 billion dollars a year in goods and services.
Pay ranges from 13 to 52 cents an hour on average. Seven states, including Alabama, Arkansas, Florida, Georgia, Mississippi, South Carolina and Texas, pay nothing at all for most assignments. More than three quarters of workers surveyed said they were not permitted to refuse an assignment, and that refusal could bring solitary confinement, loss of family contact, or a mark against them at parole.
The figure that draws most attention is the comparison with the federal minimum wage of 7.25 dollars an hour, itself unchanged since 2009. An incarcerated worker earning the average of 33 cents makes roughly one twenty second of that floor. Defenders of the system note that room, board and medical care are provided, and that deductions for restitution and court fees would apply to any wage. Critics respond that those deductions are precisely why the base rate matters.
Colorado removed the exception from its state constitution in 2018. Alabama, Nebraska, Nevada, Oregon, Tennessee, Utah and Vermont followed. In November 2024 California voters rejected Proposition 6, which would have done the same, and a Louisiana measure also failed.
The states that removed the clause did not close their work programmes. Assignments continued, generally reframed as voluntary, and participation stayed broadly stable. Wages did not rise significantly. Vermont's corrections department confirmed it would continue employing incarcerated workers, because the amendment concerned compulsion rather than employment.
The measures were not uniformly worded, which matters more than the headline suggests. Analysts at Columbia Law School found that the Alabama and Vermont amendments closed the loophole unambiguously, while the Tennessee and Oregon versions left interpretation to implementation. That variation is part of why the practical effects have been uneven.
For reform advocates, the outcome complicates the case that the exception clause is doing the work. If removing it changes little, the practical problem may lie in wage law, grievance procedures and labour protections rather than in constitutional language.
For defenders of the current system, it removes the strongest practical objection to reform. The warning was that ending compulsion would empty the kitchens and laundries and make facilities unmanageable. Nine states have now tested that prediction and it did not happen.
Both sides have therefore lost an argument they were relying on, which is unusual. It also suggests the debate has been conducted at the wrong level of abstraction for some time.
Strip away the constitutional framing and the disagreement is narrower than it looks. Almost nobody argues that prisons should have no work at all. Employment is among the few things that measurably improves outcomes after release, and idleness in a facility is a genuine safety problem that corrections officials on all sides take seriously.
The dispute is about two specific things: whether a person can be punished for declining, and whether cents an hour is a defensible rate for work that would otherwise be paid for. Those are answerable questions, and they do not require settling anything about the Thirteenth Amendment.
The state constitutional amendments were mostly symbolic, which is not the same as pointless. They established that voters in nine states, several of them conservative, will remove the language when asked directly. What they did not do is change the conditions the language permits.
There is a version of this reform that both sides could plausibly accept: work stays, compulsion goes, wages rise enough to matter on release, and grievance procedures apply. Nothing in the nine state experience suggests that combination is unworkable. Whether it is politically achievable is a different question, and the failure of California's Proposition 6 in a state with a large Democratic majority suggests the answer is not obvious.
Want the core arguments from both sides, side by side?
See both sides of the Prison Labor debate →