Is voter fraud a problem in the US?

To answer this based on evidence, I'll check reliable data on the incidence and impact of voter fraud in US elections.I am fetching a couple of key primary analyses for more detail.No, voter fraud is not a significant or widespread problem in the United States. It occurs at extremely low rates—typically measured in thousandths or hundred-thousandths of a percent of ballots cast—and there is no credible evidence that it has altered the outcome of any modern federal election (presidential or congressional). Isolated cases happen, are investigated, and are prosecuted when proven, but they do not threaten election integrity at scale.[1]

Key Evidence on Incidence Rates

Multiple independent analyses, academic studies, government investigations, and even scrutiny of databases compiled by groups concerned about fraud converge on the same finding:

•Brennan Center for Justice (reviewing dozens of studies): Voter impersonation rates of roughly 0.0003% to 0.0025%. An American is more likely to be struck by lightning than to impersonate another voter at the polls.[2]

•Washington Post investigation (2014): 31 credible instances of in-person impersonation fraud out of more than 1 billion ballots cast from 2000–2014.[2]

•Department of Justice review of 2002 and 2004 federal elections: Fraudulent ballots at a rate of about 0.00000013%, with no evidence of coordinated efforts.[2]

•Brookings Institution analysis of the Heritage Foundation’s Election Fraud Database (a conservative compilation of proven cases spanning decades): In Arizona, 36 cases over 25 years and ~42.6 million ballots (rate of 0.0000845%). In Pennsylvania, 39 cases over 30 years and >100 million votes. Similar tiny numbers in other competitive states. No election outcomes altered by ballot fraud. Heritage’s full database (updated into the 2020s) lists roughly 1,500–1,620 proven instances (convictions, diversions, civil penalties, or official findings) going back to the 1980s (or earlier in spots) against billions of votes cast—equating to roughly 7 cases per 10 million ballots in federal general elections.[1]

•Recent national reviews: One investigation identified ~513 proven cases since 2016 amid >685 million votes cast. An Associated Press review of six 2020 battleground states found fewer than 475 potential cases amid >25 million ballots (far too few to affect outcomes, and not all were counted or for one side).[3]

Mail-in/absentee fraud is similarly rare (Brookings estimate using Heritage data: average ~0.000043%, or about 4 cases per 10 million mail ballots across 2016–2022 elections). Noncitizen voting is vanishingly rare—Heritage and others document only dozens of cases over decades (e.g., ~23 from 2003–2022 in one tally); surveys of election officials and audits find rates like 0.0001% or lower.[4]

Many alleged “fraud” cases turn out to be clerical errors, misunderstandings of eligibility (e.g., recent movers or felons with restored rights), or innocent mistakes rather than intentional fraud. Databases like Heritage’s are useful for tracking proven cases but are incomplete samplings; critics note they sometimes include older/local cases, count individuals in multi-person schemes separately, or mix categories, yet even taken at face value the totals remain minuscule relative to vote volume.[5]

2020 Election Specifically

The 2020 election faced intense scrutiny: dozens of audits, recounts (including hand recounts), signature reviews, and >60 lawsuits. Courts (including those with Trump-appointed judges) rejected fraud claims for lack of evidence. Trump’s own Attorney General William Barr and cybersecurity officials stated there was no widespread fraud sufficient to change outcomes. An Arizona audit (Cyber Ninjas, commissioned by Republicans) slightly increased Biden’s margin. Statistical claims of anomalies were systematically debunked in peer-reviewed work (e.g., PNAS). Net error rates from audits were on the order of thousandths of a percent.[6]

Context and Safeguards

•Any fraud is unacceptable and is a crime (federal and state). It is prosecuted, with penalties including fines and jail. Local races with tiny margins or historical vote-buying schemes have occasionally been affected, which is why vigilance, bipartisan observation, chain-of-custody rules, audits, and voter ID/signature verification (in many jurisdictions) exist.

•Deterrents work: Risk of detection and punishment is high relative to the benefit of one (or a few) illegal votes. Large-scale coordinated fraud would require many people and leave detectable traces, which investigations have not found.

•Claims of “millions” of illegal votes or systemic flipping of national results lack supporting evidence and have been repeatedly disproven. Perception of a big problem often stems from partisan rhetoric, isolated anecdotes amplified without context, or conflating administrative errors with intentional fraud.

In short, U.S. elections have robust (if imperfect and decentralized) integrity measures. Documented voter fraud is real but rare and inconsequential to outcomes. Focusing enforcement on actual cases while avoiding unsubstantiated claims that erode public trust is the evidence-based approach. For the latest case data, resources like state election offices, the Heritage database (with caveats), or academic reviews provide transparent tracking.

Citations (6)

1brookings.eduNonprofits and think tanks
https://www.brookings.edu/articles/how-widespread-is-election-fraud-in-the-united-states-not-very/
Backs: “modern federal election (presidential or congressional). Isolated cases happen, are investigated, and are prosecuted when proven, but they do not threaten election integrity at scale.”
2legalclarity.orgCommercial content
https://legalclarity.org/how-common-is-voter-fraud-what-the-evidence-shows/
Backs: “Brennan Center for Justice (reviewing dozens of studies): Voter impersonation rates of roughly 0.0003% to 0.0025%. An American is more likely to be struck by lightning than to impersonate another voter at the polls.”
3forwardky.comNot yet classified
https://www.forwardky.com/voter-fraud-claims-persist-the-numbers-tell-a-different-story/
Backs: “An Associated Press review of six 2020 battleground states found fewer than 475 potential cases amid >25 million ballots (far too few to affect outcomes, and not all were counted or for one side).”
4legalclarity.orgCommercial content
https://legalclarity.org/voter-fraud-statistics-types-rates-and-prosecutions/
Backs: “Noncitizen voting is vanishingly rare—Heritage and others document only dozens of cases over decades (e.g., ~23 from 2003–2022 in one tally); surveys of election officials and audits find rates like 0.0001% or lower.”
5brennancenter.orgNonprofits and think tanks
https://www.brennancenter.org/media/30/download
Backs: “Databases like Heritage’s are useful for tracking proven cases but are incomplete samplings; critics note they sometimes include older/local cases, count individuals in multi-person schemes separately, or mix categories, yet even taken at face value the totals remain minuscule relative to vote volume.”
6nytimes.comNews and mediaquality 0.86
https://www.nytimes.com/2026/07/16/us/politics/trump-investigations-fraud-claims-2020-election.html
Backs: “An Arizona audit (Cyber Ninjas, commissioned by Republicans) slightly increased Biden’s margin. Statistical claims of anomalies were systematically debunked in peer-reviewed work (e.g., PNAS).”