The evidence of voter fraud in U.S. elections, gathered across more than two decades of academic studies, government investigations, court rulings, and partisan audits, points to the same conclusion: fraud occurs but is extraordinarily rare, typically at rates between 0.0001 percent and 0.0025 percent of ballots cast, and no documented case has ever changed the outcome of an American election.1Brookings Institution. How Widespread Is Election Fraud in the United States? Not Very
How Rare Is It, By the Numbers
The Brennan Center for Justice, which has tracked fraud research for more than 20 years, found voter impersonation rates between 0.0003 percent and 0.0025 percent of ballots cast.2Brennan Center for Justice. Debunking the Voter Fraud Myth A Department of Justice review of the 2002 and 2004 federal elections put the rate at 0.00000013 percent and found no evidence of any coordinated effort to alter outcomes.3Brennan Center for Justice. Resources on Voter Fraud Claims
A 2014 Washington Post investigation found 31 credible instances of in-person impersonation out of more than a billion ballots cast between 2000 and 2014.2Brennan Center for Justice. Debunking the Voter Fraud Myth A 2012 Arizona State University study found 10 cases of voter impersonation nationwide over 12 years; its 2016 follow-up found zero successful prosecutions for impersonation across the five states studied.3Brennan Center for Justice. Resources on Voter Fraud Claims
Even the Heritage Foundation’s Election Fraud Database, the most frequently cited source for proven cases, supports the same picture when its numbers are placed against the number of ballots cast. Heritage catalogs 1,620 proven instances across multiple decades.4Heritage Foundation. Election Fraud Cases by Category Heritage itself describes the database as a “sampling” and states that it is “not making any definitive claims about the extent of election fraud in our country.”5Heritage Foundation. About the Election Fraud Database Using those cases, the Brookings Institution calculated a fraud rate in Arizona of 36 cases over 25 years and 42 million ballots, or 0.0000845 percent. Pennsylvania produced 39 cases from more than 100 million votes over 30 years.1Brookings Institution. How Widespread Is Election Fraud in the United States? Not Very
What the Fraud That Does Happen Looks Like
Fraud is not one thing. The categories differ sharply in how often they occur and how easy they are to detect, so the evidence breaks down by type.
In-person impersonation, the form most often invoked to justify photo ID laws, is the rarest. The Fifth Circuit found only two convictions out of 20 million votes cast in Texas over a decade. The Fourth Circuit found that North Carolina could not identify a single person charged with in-person fraud. The record before the Supreme Court in Crawford v. Marion County contained no evidence of impersonation fraud ever occurring in Indiana.6Brennan Center for Justice. Briefing Memo: Debunking the Voter Fraud Myth
Double voting runs higher but still barely registers. A 2016 working paper set an upper limit of 0.02 percent for the 2012 election and noted that much of what it captured likely reflected measurement error rather than intentional fraud.3Brennan Center for Justice. Resources on Voter Fraud Claims
Mail ballot fraud, often cited as the vulnerable channel after the pandemic expanded absentee voting, is similarly minimal. A Brookings analysis of the 2016 through 2022 general elections found an average mail voting fraud rate of 0.000043 percent, roughly four cases per 10 million mail ballots cast.7Brookings Institution. Mail Voting in the U.S.: Data Points to Very Low Fraud
Some of the most serious fraud prosecutions involve neither voters nor ballots but campaign operatives. In Michigan, Willie Reed and Shawn Wilmoth ran signature-collection businesses that charged 2022 gubernatorial campaigns nearly $350,000 and knowingly delivered tens of thousands of forged nominating petition signatures. Reed received 24 to 240 months in prison and was ordered to pay $333,817 in restitution; Wilmoth received 4 to 20 years.8Michigan Attorney General. Second Man Sentenced in 2022 Signature Collection Election Fraud Scheme
Individual voter prosecutions tend to be small and self-defeating. In June 2026, Brenda Knuth of Hoopeston, Illinois, was sentenced to three years of probation and a $2,500 fine for voting twice in November 2024. She voted early and then again on Election Day, telling the FBI she wanted to “test the system” to see if her early vote counted.9State Journal-Register. Illinois Woman Sentenced for Voting Twice in 2024 Federal Election Cases like this show the system catches fraud; they also show how isolated and small-scale the real cases typically are.
Noncitizen Voting
Noncitizen voting has become the central fraud claim in recent political debate, so it deserves separate treatment. The evidence is consistent across sources.
A 2014 academic paper concluded the likely percentage of noncitizen voters in recent elections was zero.2Brennan Center for Justice. Debunking the Voter Fraud Myth A Brennan Center survey of 42 jurisdictions covering 23.5 million votes in 2016 found a suspected noncitizen voting rate of 0.0001 percent, with 40 of 42 jurisdictions reporting zero incidents.10Migration Policy Institute. Noncitizen Voting in U.S. Elections A Michigan Department of State review of the 2024 general election identified more than a dozen instances of noncitizens voting out of 5.7 million total votes, a rate of 0.00028 percent.11NPR. Noncitizen Voting: CEIR Review The Heritage Foundation’s own database documented 23 instances of noncitizen voting nationwide between 2003 and 2022.
Large initial claims have repeatedly shrunk under scrutiny. In Texas, an initial 2024 claim of 1,930 potential noncitizen voters narrowed to investigations of around 100 individuals. In Alabama, a 2024 effort to remove more than 3,000 alleged noncitizens from the rolls was halted by a judge after it was determined that thousands of the flagged voters were U.S. citizens. A July 2025 Center for Election Innovation and Research report concluded that most noncitizen voting allegations stem from “misunderstandings, mischaracterizations, or outright fabrications,” and that actual noncitizen registration, when it occurs, is typically the product of bureaucratic errors.12NPR. Noncitizen Voting: CEIR Review
Noncitizen voting in federal elections has been illegal since 1924. A 1996 federal law set criminal penalties including up to five years in prison and potential deportation.10Migration Policy Institute. Noncitizen Voting in U.S. Elections
What Courts and Audits Found After 2020
The 2020 presidential election produced the largest real-world test of fraud claims in modern American history. More than 60 lawsuits alleging fraud or seeking to overturn results were filed. Judges, including those appointed by Republican presidents, consistently rejected the claims. Some were dismissed for lack of standing or insufficient evidence; others were decided on the merits and found wanting.13Campaign Legal Center. Results of Lawsuits Regarding 2020 Elections
The judicial language was often sharp. In Bowyer v. Ducey, an Arizona federal court found that claims relied on “anonymous witnesses, hearsay, and irrelevant analysis of unrelated elections.” In Costantino v. City of Detroit, the court cited a lack of essential details like locations, frequency, and names of people allegedly involved in misconduct. In Ward v. Jackson, the court found that Maricopa County’s ballot duplication process was 99.45 percent accurate, with inaccuracies attributable to human error rather than fraud.13Campaign Legal Center. Results of Lawsuits Regarding 2020 Elections
The highest-profile state-level audit cut the same way. The 2021 Cyber Ninjas review of Maricopa County’s 2020 ballots, commissioned by the Republican-controlled Arizona State Senate and funded by Trump allies, concluded that Biden had actually received 99 more votes than the official count showed, while Trump received 261 fewer. Republican Senate President Karen Fann acknowledged that the final totals were “close” to the official tally and that the certified winners had won.14CNBC. Cyber Ninjas Audit of Arizona Votes Still Shows Biden Won
Some of the lawyers who filed these cases faced professional consequences. In August 2021, U.S. District Judge Linda Parker sanctioned Sidney Powell and eight other attorneys over a Michigan case she called “a historic and profound abuse of the judicial process.”15Minnesota Lawyer. Michigan Lawyers Misconduct Hearing Over 2020 Election Challenge The attorneys were ordered to pay more than $150,000 in costs, complete 12 hours of continuing legal education on pleading standards and election law, and were referred to their state bar organizations.16ABA Journal. 6th Circuit Won’t Pause CLE Sanctions Against Sidney Powell The Sixth Circuit upheld the sanctions, finding the claims “baseless” and based on “facially unreliable expert reports,” and the Supreme Court declined to review the case in February 2024.17Bloomberg Law. Supreme Court Rejects Sidney Powell Over 2020 Election Sanctions
The Statistical Claims and Why They Failed
Many 2020 fraud arguments took a statistical form, and these have been evaluated in peer-reviewed research. A 2021 study in the Proceedings of the National Academy of Sciences by Andrew Eggers, Haritz Garro, and Justin Grimmer of Stanford’s Hoover Institution systematically examined the most prominent claims and concluded: “None of them is even remotely convincing.”18PNAS. No Evidence for Systematic Voter Fraud
On the bellwether-counties argument, critics noted that Biden won only one of 19 counties that had correctly picked every presidential winner since 1980. Using a model based on 2016 voting patterns, the researchers calculated that Biden would be expected to win roughly 1.65 of those 19 counties, making his result statistically unremarkable.18PNAS. No Evidence for Systematic Voter Fraud
On the “one-in-a-quadrillion” odds argument from Charles Cicchetti, cited by the Texas Attorney General in a suit against Pennsylvania, the researchers identified a misapplied hypothesis test. When they ran the same methodology against all state-year election combinations since 1960, it rejected the null hypothesis in 1,488 of 1,498 cases, which would absurdly imply fraud in nearly every American election.18PNAS. No Evidence for Systematic Voter Fraud
On Dominion voting machines, the unadjusted coefficient for Dominion use was a statistically insignificant 0.007. Once the researchers controlled for 2016 vote shares and state-level differences, the coefficient flipped to negative 0.006, slightly in Trump’s favor. On absentee ballot patterns in Fulton County, Georgia, and Allegheny County, Pennsylvania, claims disappeared when the researchers applied a fixed-effects model that properly accounted for county- and state-level variation.18PNAS. No Evidence for Systematic Voter Fraud
Grimmer’s related work in the Journal of Legal Analysis concluded that the “vast majority of election policies can’t possibly change even the closest elections.”19Stanford University. Debunking Evidence in Election Fraud Cases
Recent Federal Investigations and What They Have Produced
Beginning in 2025, the Trump administration opened a broad federal effort to investigate the 2020 and, increasingly, the 2024 elections. The investigations have revived allegations previously examined and rejected.
In January 2026, federal investigators seized more than 600 boxes of 2020 ballots from Fulton County, Georgia. Director of National Intelligence Tulsi Gabbard oversaw the operation and arranged for President Trump to speak directly to agents afterward.20Votebeat. FBI Investigation Into 2020 Election The warrant was issued by Missouri U.S. Attorney Thomas Albus rather than the U.S. Attorney for Georgia, bypassing customary protocol. Senators Sheldon Whitehouse and Richard Blumenthal called the investigation a “sham,” noting that the warrant affidavit cited witnesses known for promoting debunked election fraud claims and that the FBI special agent in charge of the Atlanta field office was reportedly forced out after refusing to participate.21U.S. Senate. Whitehouse, Blumenthal Call for Investigation Into FBI Seizure
In May 2026, Judge J. P. Boulee denied Fulton County’s request for the return of the ballots in a 68-page ruling. He acknowledged that the warrant affidavit was “far from perfect” and contained “troubling” and “defective” elements but concluded that the county had not shown its rights were “callously disregarded.” The court noted no evidence that the seized materials had been manipulated or that ballot secrecy was compromised.22ABC News. Judge Denies Request to Force FBI to Return Seized 2020 Ballots No charges have resulted. The five-year statute of limitations for most potential 2020 charges expired in 2025.20Votebeat. FBI Investigation Into 2020 Election
In April 2026, Assistant Attorney General Harmeet Dhillon sent a letter to the Wayne County, Michigan, clerk demanding all 2024 ballots, receipts, and envelopes. The letter cited three voter fraud convictions from the 2020 election and a dismissed civil case as justification. Michigan’s governor, attorney general, and secretary of state rejected the demand. Attorney General Dana Nessel called it “as absurd as it is baseless,” noting that none of the fraud examples cited originated from the 2024 election.23ABC News. Justice Department Demands 2024 Election Ballots From Wayne County FBI Director Kash Patel suggested in April 2026 that arrests were imminent, but as of mid-2026 none have been announced.20Votebeat. FBI Investigation Into 2020 Election
The administration also overhauled the Department of Homeland Security’s Systematic Alien Verification for Entitlements (SAVE) database to allow bulk citizenship checks against state voter rolls. More than 60 million voter records were run through the system, and roughly 21,000 were flagged as potential noncitizens.24NPR. Judge Rules SAVE Voter Data Use Unlawful The program mistakenly flagged American-born citizens and naturalized citizens as potential noncitizens.25Votebeat. Judge Rules Against Trump Overhaul of SAVE Database
On June 22, 2026, U.S. District Judge Sparkle Sooknanan declared the overhauled SAVE program unlawful. In a 75-page ruling, she found that federal agencies had “haphazardly combined and repurposed the private information of millions of Americans, including citizenship data that they knew to be unreliable,” in violation of the Privacy Act, the Social Security Act, and the Administrative Procedure Act.24NPR. Judge Rules SAVE Voter Data Use Unlawful By June 30, 2026, the government had disabled the bulk upload and Social Security number search functions.26Democracy Docket. DHS Appears to Have Shut Down SAVE for Checking Voter Citizenship The administration also lost every federal lawsuit it filed to obtain detailed state voter data. Judge Denise Casper permanently blocked an executive order requiring proof of citizenship for registration, and Judge Indira Talwani struck down an order seeking a national voter list, ruling it violated the separation of powers.27Los Angeles Times. Trump’s Fixation on Voting Has Had Mixed Results
The Bottom Line on the Evidence
Across every method of looking, the answer converges. Academic studies find rates measured in thousandths or millionths of a percent. The Heritage Foundation’s own database, when its proven cases are placed against ballots cast, confirms the same picture. Courts hearing 2020 fraud lawsuits rejected them, often in blunt terms. A Trump-aligned audit in Arizona confirmed Biden’s win. Peer-reviewed statistical research found that the quantitative claims against the 2020 result do not survive basic model specification. Federal investigations launched in 2025 and 2026, drawing on the same previously examined allegations, have so far produced no charges and one major court loss limiting the government’s use of voter data.
Fraud exists. The system detects it, and individuals are prosecuted for it. The Brookings analysis of all available evidence found that no U.S. election outcome has ever been altered by ballot fraud.1Brookings Institution. How Widespread Is Election Fraud in the United States? Not Very