Is voter fraud a problem in the US?

NoSays it is contested

No, voter fraud is not a significant or widespread problem in U.S. elections. It occurs occasionally (as with any large-scale human system), but extensive research, government investigations, court records, audits, and analyses from across the political spectrum show it is extremely rare—typically measured in tiny fractions of a percent of ballots cast—and almost never occurs at a scale that could affect the outcome of federal or major statewide elections.[1]

Key Data and Findings

Multiple independent studies and reviews consistently reach this conclusion:

•The Brennan Center for Justice (reviewing meticulously studied elections) found incident rates of voter fraud between 0.0003% and 0.0025%. In-person voter impersonation is so rare that an American is more likely to be struck by lightning.[2]

•A 2014 analysis identified only 31 credible instances of voter impersonation out of more than 1 billion ballots cast from 2000–2014.[3]

•A Department of Justice unit examining the 2002 and 2004 federal elections found a fraud rate of roughly 0.00000013% of ballots, with no evidence of coordinated efforts or in-person impersonation fraud.[3]

•Post-2020, an Associated Press review of every potential fraud case in six contested swing states found fewer than 475 potential cases out of more than 25 million ballots—far too few to affect results (Biden’s combined margin in those states exceeded 300,000 votes). Most involved isolated individuals, not schemes.[4]

•Hearst Television’s investigation found 513 proven cases (convictions or court actions) since 2016, out of an estimated 685 million ballots cast. Only a tiny handful led to overturned local results.[5]

Even the Heritage Foundation’s Election Fraud Database (a conservative source that tracks proven cases and describes itself as a non-exhaustive “sampling”) lists roughly 1,500–1,600 instances spanning decades (back to the 1980s or so), including convictions, civil penalties, and official findings. This covers all types (absentee fraud, false registration, double voting, felon voting, noncitizen voting, etc.). Analyses of their own data show minuscule rates: e.g., Arizona had 36 cases over 25 years across ~42.6 million ballots (0.0000845%); Pennsylvania had 39 cases over 30 years across >100 million votes. No major election outcomes were altered.[6][7]

Noncitizen voting is similarly rare. Heritage has identified dozens of convictions over decades (e.g., around 70–100 depending on the timeframe and definition). Brennan Center surveys and other reviews find rates like 0.0001% or lower in large samples; it has never been shown to tip a significant election.[8][9]

Fraud is more plausible (and occasionally impactful) in very small local races with tiny margins, where a handful of ballots matter, or in isolated schemes like vote-buying. National and statewide races involve massive volumes of votes, decentralized administration (thousands of jurisdictions), bipartisan oversight, paper trails/audits in many places, voter roll maintenance, and other safeguards that make large-scale undetected fraud extremely difficult.[4]

Context and Caveats

•Allegations vs. proven cases: Many high-profile claims turn out to be clerical errors, data mismatches, misunderstandings (e.g., someone voting while thinking they were eligible), or unsubstantiated. Prosecutions occur when evidence exists, and officials of both parties routinely note that fraud is “exceedingly rare.”

•Undercount possibility: Not every instance is detected or prosecuted, and databases like Heritage’s are samples. However, the consistency across academic studies, journalistic investigations, state audits, and even aggressive enforcement efforts (including recent ones) indicates the true rate remains vanishingly small relative to total votes (hundreds of millions per cycle; billions over decades).

•Public perception: Surveys show many Americans (especially along partisan lines) believe fraud is common, often due to repeated claims. This erodes trust even when data does not support widespread problems.[10]

•Safeguards and responses: Elections use layered security (ID or affidavit requirements in many states, signature verification for mail ballots, poll books, audits, chain-of-custody rules, etc.). States clean rolls, and prosecutors pursue cases. Recent federal efforts have charged individuals for illegal voting/registration, but the numbers remain small.[8]

In short, while any fraud is unacceptable and worth preventing/prosecuting, the evidence shows it is not a systemic problem that undermines the integrity or outcomes of U.S. elections at scale. Claims of “widespread” fraud sufficient to “rig” results have been repeatedly investigated and rejected by courts, officials, and researchers. Focusing on actual vulnerabilities (e.g., improving roll maintenance or cybersecurity) is more productive than unsubstantiated narratives.

Citations (10)

1npr.orgNews and mediaquality 0.93
https://www.npr.org/2024/10/11/nx-s1-5147732/voter-fraud-explainer
Backs: “It occurs occasionally (as with any large-scale human system), but extensive research, government investigations, court records, audits, and analyses from across the political spectrum show it is extremely rare—typically measured in tiny fractions of a percent of ballots cast—and almost never occurs at a scale that…”
evidence, article, points to no. Claim: Voter fraud is extremely rare, a tiny fraction of ballots, and almost never at a scale that affects election outcomes.
2legalclarity.orgCommercial content
https://legalclarity.org/voter-fraud-statistics-types-rates-and-prosecutions/
Backs: “In-person voter impersonation is so rare that an American is more likely to be struck by lightning.”
statistic, article, points to no. Claim: Brennan Center found fraud incident rates between 0.0003% and 0.0025%; impersonation rarer than being struck by lightning.
3brennancenter.orgNonprofits and think tanks
https://www.brennancenter.org/our-work/research-reports/debunking-voter-fraud-myth
Backs: “A 2014 analysis identified only 31 credible instances of voter impersonation out of more than 1 billion ballots cast from 2000–2014.”
statistic, article, points to no. Claim: 31 credible impersonation instances among over 1 billion ballots 2000–2014; DOJ unit found roughly 0.00000013% fraud rate.
4apnews.comNews and mediaquality 1.00
https://apnews.com/article/election-2024-voter-fraud-trump-harris-a3b4c2db17217311770259193c115b80
Backs: “Post-2020, an Associated Press review of every potential fraud case in six contested swing states found fewer than 475 potential cases out of more than 25 million ballots—far too few to affect results (Biden’s combined margin in those states exceeded 300,000 votes). Most involved isolated individuals, not schemes.”
statistic, article, points to no. Claim: AP review found fewer than 475 potential fraud cases among 25 million ballots in six swing states; safeguards make large-scale fraud difficult.
5mysanantonio.comNot yet classified
https://www.mysanantonio.com/news/article/voter-fraud-claims-persist-the-numbers-tell-a-22437199.php
Backs: “Hearst Television’s investigation found 513 proven cases (convictions or court actions) since 2016, out of an estimated 685 million ballots cast. Only a tiny handful led to overturned local results.”
statistic, article, points to no. Claim: Hearst found 513 proven fraud cases since 2016 out of an estimated 685 million ballots cast.
6brookings.eduNonprofits and think tanks
https://brookings.edu/articles/how-widespread-is-election-fraud-in-the-united-states-not-very
Backs: “Pennsylvania had 39 cases over 30 years across >100 million votes. No major election outcomes were altered.”
evidence, article, points to no. Claim: Analyses of Heritage database data show minuscule fraud rates, such as Arizona's 0.0000845%, with no major outcomes altered.
7newsweek.comNews and mediaquality 0.82
https://newsweek.com/voter-fraud-us-elections-data-research-1978342
Backs: “Pennsylvania had 39 cases over 30 years across >100 million votes. No major election outcomes were altered.”
evidence, article, points to no. Claim: Heritage database figures for Arizona and Pennsylvania show very few cases relative to total votes.
8nytimes.comNews and mediaquality 0.86
https://www.nytimes.com/interactive/2026/10/15/us/politics/noncitizen-voting-data.html
Backs: “Brennan Center surveys and other reviews find rates like 0.0001% or lower in large samples; it has never been shown to tip a significant election.”
evidence, data page, points to no. Claim: Noncitizen voting is rare, with dozens of convictions over decades; recent federal charges remain small in number.
9brennancenter.orgNonprofits and think tanks
https://www.brennancenter.org/our-work/research-reports/noncitizen-voting-vanishingly-rare
Backs: “Brennan Center surveys and other reviews find rates like 0.0001% or lower in large samples; it has never been shown to tip a significant election.”
statistic, article, points to no. Claim: Brennan Center surveys find noncitizen voting rates around 0.0001% or lower and never shown to tip a significant election.
10frontiersin.orgAcademic
https://www.frontiersin.org/journals/political-science/articles/10.3389/fpos.2024.1493897/full
Backs: “This erodes trust even when data does not support widespread problems.”
evidence, article, neutral. Claim: Many Americans, especially along partisan lines, believe fraud is common, eroding trust even when data doesn't support it.