Is voter fraud a problem in the US?

No, voter fraud is not a significant or systemic problem in the United States. It occurs in isolated instances and is illegal (with criminal penalties), but extensive research, government investigations, audits, court records, and databases consistently show it is extremely rare—far too infrequent to affect the outcomes of major elections like presidential or most statewide races.[1]

Key Evidence on Rates and Scale

•Documented rates are tiny: Studies of carefully examined elections find fraud rates on the order of 0.0003% to 0.0025% of votes cast (or lower). A U.S. Department of Justice review of 2002–2004 federal elections found a rate of about 0.00000013%. A 2014 analysis identified just 31 credible cases of in-person voter impersonation out of more than 1 billion ballots cast from 2000–2014.[1]

•Heritage Foundation database (conservative source): This tracks a sampling of proven cases (convictions, judicial/official findings, etc.) and lists roughly 1,500–1,600 instances spanning decades (roughly 1982–2025). Even using this data, analyses (e.g., by Brookings) calculate rates like 0.0000845% in Arizona (36 cases over 25 years and >42 million ballots) or similarly minuscule figures in Pennsylvania (~39 cases over 30 years and >100 million votes). Across hundreds of millions to billions of ballots nationwide, this is a tiny fraction. Heritage itself notes it is not comprehensive, but the known cases remain negligible relative to total votes.[2]

•Specific types:

•In-person impersonation: The rarest form; more likely to be struck by lightning than to successfully impersonate a voter at the polls.

•Noncitizen voting: Vanishingly rare (dozens of proven cases over decades in major databases; recent enforcement efforts and state audits yield tiny numbers, e.g., fractions of a percent or less of total votes/registrations). It is a federal and state crime.[3]

•Mail-in/absentee fraud: Also rare (e.g., Brookings estimates ~4 cases per 10 million mail ballots on average in recent cycles). Studies find no evidence that mail voting systematically increases fraud risk.[3]

•Impact on outcomes: Researchers and analyses (including of Heritage data) find no evidence that voter fraud has altered the outcome of a U.S. presidential election or most major races in modern times. Isolated local elections have occasionally been affected or required re-runs (e.g., a 2018 North Carolina congressional race due to an absentee ballot scheme). Post-2020 reviews, audits, recounts, and dozens of failed lawsuits confirmed no widespread fraud capable of changing results; an AP review of battleground states found fewer than 475 potential cases amid tens of millions of votes—nowhere near victory margins.[4]

Sources spanning the political spectrum (Brennan Center, Brookings, academic studies, MIT Election Lab, state/federal investigations, and even critical analyses of Heritage data) reach the same conclusion: fraud is not widespread or systemic. Many alleged cases turn out to be administrative errors, misunderstandings, or unsubstantiated claims rather than intentional fraud.[5]

Why It Remains Rare

U.S. elections are highly decentralized (run primarily by states and localities) with layered safeguards: voter ID or affidavit requirements in many places, signature verification, ballot tracking, bipartisan poll watchers, audits/risk-limiting audits, cross-checks of rolls (e.g., via ERIC in participating states), prosecution of violations, and severe penalties. The personal risk/reward calculation deters most would-be fraudsters—one illegal vote rarely tips anything meaningful, while detection risks jail time and fines. Dead people or noncitizens on rolls (more common as administrative issues) do not automatically mean fraudulent votes are cast and counted.[6]

Claims of massive fraud (e.g., “millions of illegal votes”) have repeatedly been investigated and rejected by courts, election officials (of both parties), and independent reviews. Perception of a big problem often exceeds the evidence, driven by rhetoric, which can erode trust without corresponding data.[7]

Bottom Line

Voter fraud is real in the sense that individual cases happen and should be prevented/prosecuted—no serious observer claims a perfect zero rate. However, it is not a problem at a scale that undermines the integrity or results of U.S. elections. Resources are better focused on accessible, secure voting, accurate rolls, and clear processes rather than measures addressing a largely nonexistent widespread threat. Elections officials continuously improve safeguards, and the system has proven resilient.

Citations (7)

1legalclarity.orgCommercial content
https://legalclarity.org/voter-fraud-statistics-types-rates-and-prosecutions/
Backs: “It occurs in isolated instances and is illegal (with criminal penalties), but extensive research, government investigations, audits, court records, and databases consistently show it is extremely rare—far too infrequent to affect the outcomes of major elections like presidential or most statewide races.”
2brookings.eduNonprofits and think tanks
https://www.brookings.edu/articles/how-widespread-is-election-fraud-in-the-united-states-not-very/
Backs: “Heritage itself notes it is not comprehensive, but the known cases remain negligible relative to total votes.”
3nytimes.comNews and mediaquality 0.86
https://www.nytimes.com/2026/04/07/us/politics/trump-mail-voting-elections-fact-check.html
Backs: “Vanishingly rare (dozens of proven cases over decades in major databases; recent enforcement efforts and state audits yield tiny numbers, e.g., fractions of a percent or less of total votes/registrations). It is a federal and state crime.”
4legalclarity.orgCommercial content
https://legalclarity.org/instances-of-voter-fraud-proven-cases-types-and-prevention/
Backs: “Post-2020 reviews, audits, recounts, and dozens of failed lawsuits confirmed no widespread fraud capable of changing results; an AP review of battleground states found fewer than 475 potential cases amid tens of millions of votes—nowhere near victory margins.”
5brennancenter.orgNonprofits and think tanks
https://www.brennancenter.org/topics/voting-elections/vote-suppression/myth-voter-fraud
Backs: “Many alleged cases turn out to be administrative errors, misunderstandings, or unsubstantiated claims rather than intentional fraud.”
6thehill.comNews and mediaquality 0.90
https://thehill.com/regulation/court-battles/4962270-us-election-fraud-rare/amp/
Backs: “Dead people or noncitizens on rolls (more common as administrative issues) do not automatically mean fraudulent votes are cast and counted.”
7brookings.eduNonprofits and think tanks
https://www.brookings.edu/articles/widespread-election-fraud-claims-by-republicans-dont-match-the-evidence/
Backs: “Perception of a big problem often exceeds the evidence, driven by rhetoric, which can erode trust without corresponding data.”