I'll examine current evidence on how common voter fraud is in the U.S.
Is voter fraud a problem in the United States? Empirical research and official reviews consistently find that it exists but is extremely rare and has not been shown to affect outcomes of federal or statewide elections.
What the evidence shows
Multiple independent analyses put the rate of proven voter fraud at effectively zero for practical election outcomes:
•A widely cited Brennan Center for Justice analysis found that voter impersonation occurs at a rate between 0.0003% and 0.0025%. Other summaries of the Brennan Center's work cite incident rates between roughly 0.0003% and 0.0025% — orders of magnitude below what would be needed to "rig" statewide outcomes.
•A review cited by the Brennan Center of 84 million votes in 22 states yielded 14 instances referred for prosecution — an implied fraud rate on the order of 0.00000017% in that sample. Another description of that same review notes it identified just 14 prosecutions in that sample, producing an estimated fraud rate on the order of 0.00000017 percent.
•Legal scholar Justin Levitt documented just 31 credible cases of in-person voter impersonation out of roughly a billion ballots in a long-term study.
•A Massachusetts Institute of Technology study of mail ballots over two decades called mail-in fraud "exceedingly rare" and reported rates on the order of 0.00006 percent of votes nationally. Other reviews put mail ballot fraud at about 0.00006% over two decades.
•Brookings' review of Heritage Foundation data found, for example, that in Arizona across 36 elections and 42.6 million ballots there were 36 documented cases — about 0.0000845% of ballots — and that no federal outcome was changed. Pennsylvania's 30-year sample produced roughly 39 cases out of more than 100 million ballots — about 0.000039 percent.
•The Associated Press in 2021 identified fewer than 475 instances across six battleground states out of about 25 million votes.
The largest database of proven cases
The most-cited public tally comes from the Heritage Foundation's Election Fraud Database, which catalogs proven instances that led to prosecutions or findings of wrongdoing:
•That database lists approximately 1,465 proven cases over roughly 44 years in one summary and around 1,561 instances (with 1,325 convictions) in other reporting, indicating on the order of 30–40 proven cases per year on average.
•Other summaries note the Heritage Foundation's Election Fraud Database catalogs roughly 1,561 instances that produced about 1,325 convictions across many decades, and that the Heritage Foundation's database lists roughly 1,285–1,561 proven instances across multiple decades.
•The Heritage database is explicit that it is a sampling and not exhaustive. When measured against the number of ballots cast, proven instances are minuscule — for example, a database search spanning decades found roughly 1,000 examples amid billions of ballots.
Federal election security agencies have reached similar conclusions. The Cybersecurity and Infrastructure Security Agency (CISA), which debuted in 2018, called the 2020 election "the most secure in American history," and added: "There is no evidence that any voting system deleted or lost votes, changed votes or was in any way compromised."
What kinds of fraud do occur?
When scholars tally proven cases they typically find fraud that is individual or small-scale: falsified voter registration forms, people casting ballots in more than one jurisdiction, absentee-ballot mishandling or forgery, and isolated instances of vote buying or coercion.
Allegations can be broadly categorized into several typesincluding:
•Voter Impersonation
•Double Voting
•Registration Fraud
•Ballot Harvesting
Federal law defines voter/ballot fraud to include a voter intentionally giving false information when registering to vote, such as false citizenship claims, voting when ineligible, voting more than once or using someone else's name, and vote buying schemes where the voter receives money or something of value in exchange for voting.
High-profile isolated cases do occur, such as the 2019 North Carolina 9th Congressional District ballot-trafficking scheme where an operative collected and altered absentee ballots, leading officials to overturn the result. But such cases are treated as criminal exceptions and are detected through audits and prosecution.
Why the debate continues
While scholarly and official consensus is that voter fraud is exceedingly rare, concerns about election integrity persist for other reasons:
•Some analysts argue voter rolls are chronically bloated with dead registrants, people registered in multiple states, and non-citizens added during motor-voter pushes, and that mail-in and absentee balloting removes chain-of-custody safeguards and that signature verification is inconsistent.
•Public confidence is also affected by the sheer scale of elections - with more than 150 million votes cast in a presidential election, even a few dozen proven cases can generate significant news coverage, while the denominator (billions of ballots over decades) makes the rate appear minuscule in statistical reviews.
Long-running empirical reviews find incident rates that are extremely small compared with total ballots cast, and broad scholarly consensus summarized for the Supreme Court is that voter fraud is exceedingly rare.
In short, researchers, courts, and election security agencies agree: voter fraud is real and prosecuted when found, but measured as a percentage of votes cast it is infinitesimal and has not been shown to have changed the outcome of a modern federal election.