Roughly 4.05 million Americans could not vote in 2024 because of a felony conviction. That is about 1.7 percent of the voting age population, or one in every 59 adults. The figure sounds like it should describe the prison population, and it does not. Around 70 percent of those people are living in their communities: on parole, on probation, or finished with their sentence entirely.
The sharpest number in the whole debate is that last group. More than 1.6 million people had completed every day of their sentence, including any parole and probation, and still could not vote. That is roughly 40 percent of everyone affected. They work, pay taxes, raise children and live under laws made by legislatures they have no say in choosing.
The trend is the part most people get wrong. The total peaked at about 5.88 million in 2016 and has fallen to 4.05 million, a drop of 31 percent in eight years. Since 1997, more than half the states have expanded restoration in some form, returning the vote to roughly two million people. Voting restrictions are usually assumed to be tightening. On this particular measure they have been loosening for a decade.
There is no national rule. In Maine, Vermont and the District of Columbia nobody loses the vote at all, including people currently serving a prison sentence. In roughly 23 states the vote comes back automatically on release. In about 15 more it waits until parole and probation are finished. In around 10 states it can be withheld indefinitely for certain offences, or require a governor's pardon.
The practical result is that the same conviction produces opposite outcomes depending on geography. The gap between states is enormous. Tennessee bars 7.68 percent of its adults, Florida 6.13 percent and Alabama 5.95 percent, while Maine and Vermont bar nobody. Florida alone accounts for nearly a quarter of the national total.
None of this is settled law in one direction. Virginia is putting its own restoration rule directly to voters, and several states have moved in both directions within the same decade. The map you would draw today is not the map from five years ago, and it will not be the map five years from now.
Florida is where this debate became concrete. In 2018 voters approved Amendment 4 with 64.55 percent support, restoring the vote to people who had completed their sentences, while keeping exclusions for murder and felony sexual offences. Estimates suggested about 1.4 million people would be affected.
The legislature then passed a law conditioning eligibility on payment of all outstanding fines, fees and restitution. The difficulty was practical as much as philosophical: Florida had no central database of what anyone owed, because the records sat across 67 county clerk offices. People could not find out whether they qualified. Some registered after being told they were eligible and were later arrested.
A federal appeals court sitting en banc upheld the requirement over several dissents, holding that court costs are penalties rather than poll taxes. Supporters of the law say voters approved restoration on completion of a sentence, and money owed to victims is part of a sentence. Opponents say a rule the state itself cannot calculate is not a rule anyone can follow.
Critics often describe disenfranchisement as a relic that the Constitution barely tolerates. The text is less accommodating to that view than it sounds. Section 2 of the Fourteenth Amendment reduces a state's representation if it denies the vote to adult citizens, but it carves out denial for participation in rebellion, or other crime. The Reconstruction Congress wrote that phrase deliberately.
In Richardson v. Ramirez in 1974 the Supreme Court read that carve-out as affirmative permission and upheld California's ban six votes to three. That decision still stands. It means the argument against these laws has to be made on policy and on history, not on a straightforward claim that the Constitution forbids them.
The historical argument is where it gets uncomfortable for the other side. Mississippi's list of disqualifying crimes was written into its 1890 constitution, and delegates at that convention said plainly that the purpose was to exclude Black voters. It is still in force. Nationally about one in 22 Black adults was barred in 2024, more than three times the rate for everyone else.
Both sides agree a sentence should end. They disagree about what the sentence includes.
One view is that the ballot is part of citizenship rather than part of the punishment, so it should return the moment the punishment is served, and arguably should never have been taken. Restoring it also supports reintegration, and research comparing states finds lower rearrest rates where restoration is automatic, though the researchers are careful that this is an association rather than a proven cause.
The other view is that voting is a share in writing the rules everyone else must obey, and asking for a demonstrated period of law-abiding life before handing that back is a reasonable condition rather than a punishment. Federal research following hundreds of thousands of releases found roughly two-thirds were rearrested within three years. A waiting period, on this view, is a test anyone can pass.
Public opinion has moved a long way in one direction. Once a sentence is complete, 89 percent of Democrats, 66 percent of independents and 54 percent of Republicans support restoring the vote. What remains genuinely contested is everything before that point.
Want the core arguments from both sides, side by side?
See both sides of the Felon Voting Rights debate →