Noncitizen Mass-Voting Fraud Claims

Noncitizen Mass Voting Fraud Claims

Origins of the Claim

Allegations that non-citizens vote in significant numbers in United States federal elections have circulated in American political discourse for decades, but gained substantial prominence following the 2016 presidential election. When Donald Trump won the Electoral College but lost the popular vote to Hillary Clinton by approximately 2.9 million votes, he publicly claimed — without evidence — that 3 to 5 million illegal votes had been cast, primarily by undocumented immigrants and non-citizen residents. This framing transformed a longstanding fringe concern into a mainstream political assertion.

What Proponents Argue

Proponents of the mass non-citizen voting theory contend that loose voter registration requirements, inconsistent ID verification, and limited cross-referencing of voter rolls against immigration databases create conditions in which large numbers of non-citizens could register and vote undetected. Some cite studies — most prominently a 2014 paper published in Electoral Studies by Richman, Chattha, and Earnest — as evidence that hundreds of thousands of non-citizens may have voted in recent elections. Proponents also point to "motor voter" registration, which automatically registers eligible residents when they obtain driver's licenses, arguing that states issuing licenses to undocumented immigrants create pathways for ineligible registration.

What Evidence Shows

The evidentiary record consistently contradicts the mass-fraud narrative. The Government Accountability Office, the Brennan Center for Justice, and the Heritage Foundation — ideologically spanning the political spectrum — have all studied the question and reached similar conclusions: confirmed cases of non-citizen voting are extremely rare.

The Heritage Foundation's Election Fraud Database, maintained specifically to document proven cases of electoral fraud, contains records of approximately 12 to 15 confirmed instances of non-citizen voting per election cycle nationally, against a backdrop of roughly 150 million votes cast. This represents a fraud rate measured in parts per million. The Brennan Center's research has found similar figures. State-level audits following high-profile elections — including those conducted in Georgia, Arizona, and Wisconsin after 2020 — found no evidence of systematic non-citizen participation.

The Richman et al. study cited by proponents has been extensively criticized by political scientists, including the authors of the Cooperative Congressional Election Study dataset the paper relied upon. Those authors found that the apparent non-citizen responses in the data were most likely the result of survey measurement error rather than actual non-citizen voters.

In January 2017, President Trump established the Presidential Advisory Commission on Election Integrity, co-chaired by Vice President Pence and Kansas Secretary of State Kris Kobach. The commission requested extensive voter data from all 50 states. Most states refused or provided only partial data. The commission was disbanded in January 2018 without issuing a final report or finding evidence of widespread fraud.

Why the Claim Persists

The claim persists in part because it is structurally difficult to disprove to a motivated skeptic. Absence of documented fraud can be reframed as evidence of inadequate detection. The claim also serves clear political purposes, providing justification for voter ID laws, purges of voter rolls, and restrictions on registration — policies that have measurable effects on turnout among specific demographic groups.

Current Verdict

Debunked. Every major institutional investigation of non-citizen voting has found it to be vanishingly rare. The specific claim that it occurs at a scale sufficient to influence election outcomes is unsupported by any credible evidence.

What Would Change the Verdict

A rigorously designed audit — with full access to voter registration data, immigration records, and ballot information — that documented systematic non-citizen participation in numbers exceeding confirmed cases by several orders of magnitude would require reassessment. No such study has produced those findings.

The Strongest Version of the Proponent Case

A fair accounting of this theory has to start with what its proponents actually point to, rather than a caricature of it. Four categories of evidence are cited most often, and each deserves to be stated in its strongest form before being weighed against the record.

The first is Kris Kobach's argument, made as Kansas Secretary of State and later as vice-chair of President Trump's 2017 Presidential Advisory Commission on Election Integrity, that documentary proof-of-citizenship requirements are necessary because existing self-attestation systems cannot catch noncitizens who register. Kobach did not rely only on the disputed Richman survey; he pointed to Kansas's own registration and DMV data and to cases his office had identified through jury-duty records, arguing these showed a real, if partially hidden, vulnerability in the registration system.

The second is the pattern of state list-maintenance audits since 2022, several of which have produced large initial numbers. Georgia's Secretary of State opened files on more than 1,600 registrants flagged by the SAVE (Systematic Alien Verification for Entitlements) database in 2022, and Ohio's Secretary of State referred over 1,000 flagged registrations to the U.S. Department of Justice in 2025, with state officials publicly describing some of these individuals as having cast ballots in prior federal elections. Proponents argue that officials would not run these audits, or refer cases to prosecutors and the DOJ, if the underlying concern were baseless.

The third is the Heritage Foundation's Election Fraud Database, which — unlike anonymous claims — documents named individuals, court dockets, and prosecutions going back decades. Proponents argue that a documented, prosecutable pattern of noncitizen voting, however small in absolute terms, demonstrates the vulnerability is real rather than hypothetical, and that the true number is necessarily an undercount because most jurisdictions do not systematically check citizenship after registration.

The fourth, and most recent, is a U.S. Census Bureau publication released in August 2026, titled "Noncitizen Voting in the 2020 Election, A Beginning Analysis," which states with "high confidence" that of roughly 128 million matched voter records, more than 24,000 represented noncitizens who cast ballots in the 2020 general election. President Trump and allied officials cited this figure as vindication of years of claims that federal voter rolls contain substantial numbers of ineligible voters, with more than 32 million additional records still to be analyzed at the time of publication.

What the Audits Actually Resolved Once Completed

The pattern across every state audit is that initial flagged totals shrink sharply once investigated, and the state audits proponents cite as their strongest recent evidence are also the clearest illustrations of that pattern.

Georgia's 2022 SAVE-database review flagged 1,634 registration attempts for further review. When Georgia's Secretary of State completed the follow-up investigation and reported results in October 2024, the number of individuals conclusively confirmed to be noncitizens registered to vote was 20, out of 8.2 million registered voters statewide — a rate of roughly two ten-thousandths of one percent. None of the 20 had cast a ballot in the 2024 general election; nine had voting histories from prior elections, and the remaining eleven had no record of ever voting. All 20 registrations were canceled and referred to local district attorneys, the Georgia Bureau of Investigation, and the State Election Board — referrals for potential prosecution, not findings of proven fraud by a court. A further 156 cases remained under investigation and had not been resolved either way at the time of reporting. Gabriel Sterling, chief operating officer of the Georgia Secretary of State's office, stated publicly that "there is no proof that there is this overwhelming number of noncitizens on the rolls."

Ohio's 2025 referral followed the same shape at a different point in the pipeline: Secretary of State Frank LaRose referred 1,084 flagged registrations to the U.S. Department of Justice, with his office stating that 167 of those individuals appeared to have cast ballots in federal elections since 2018, and referring an additional 135 cases tied to other alleged unlawful-voting patterns (double registration across states, voting twice in one election, and similar). These are referrals for federal investigation and potential prosecution; they are not adjudicated findings, and Ohio's own earlier list-maintenance cycle (2019-2023) had previously flagged 521 potential noncitizen registrations and confirmed only one. The Bipartisan Policy Center and the Center for Election Innovation and Research (CEIR), reviewing audits across multiple states — including Iowa, where an initial flag of 2,176 potential noncitizen registrations fell to 277 confirmed cases and 35 alleged ballots cast, and Michigan, where 16 confirmed noncitizen votes represented roughly 0.00028% of ballots cast in the 2024 general election — describe this as a consistent pattern: initial database-matching flags are dominated by clerical mismatches, outdated naturalization records, and identity-matching errors, and shrink by one or two orders of magnitude once individually verified.

The Kansas Case Study

Kobach's own signature policy provides the most thoroughly litigated test of the underlying claim. When Kansas's Documentary Proof of Citizenship law was challenged in federal court in Fish v. Kobach, the trial record — developed using Kansas's own registration and DMV data, the same kind of data proponents point to elsewhere — showed that 39 noncitizens had successfully registered to vote in Kansas, out of a statewide voter roll of roughly 1.76 million, a rate of about 0.002%. U.S. District Judge Julie Robinson wrote that the state's data revealed "no iceberg, only an icicle, largely created by confusion and administrative error," rejecting Kobach's extrapolation of the disputed national noncitizen-voting estimate to Kansas's actual numbers. The same law, meanwhile, had blocked or suspended registration for 31,089 applicants — 12.4% of all new registrants in the state during 2013-2015 — for failure to produce citizenship paperwork, the overwhelming majority of whom were later shown to be eligible U.S. citizens. The Tenth Circuit Court of Appeals unanimously affirmed the district court's ruling in 2020. The case is frequently cited by both sides: proponents note that noncitizens were in fact found on the rolls; the court record shows the number was several orders of magnitude smaller than the political claims built on it, and that the verification regime built to catch them disenfranchised roughly 800 eligible citizens for every actual noncitizen it might have stopped.

The 2026 Census Bureau Report

The Census Bureau's August 2026 publication is the most significant new document in this debate and deserves scrutiny on its own terms rather than dismissal by association. The report's core claim — 24,000-plus noncitizen voters identified with "high confidence" among 128 million matched records — was quickly challenged by data professionals and former Census Bureau officials on methodological grounds rather than political ones. Critics noted the publication lacks named authors and a stated margin of error, unusual omissions for a Census Bureau technical release. Beth Jarosz, a senior fellow at Georgetown University's Massive Data Institute, said the absence of that information was "glaring" and left her thinking the Bureau might be "hiding something." Terri Ann Lowenthal, who served as staff director of the U.S. House subcommittee that oversaw the Census Bureau from 1987 to 1994, said the report was "beneath the Census Bureau's stellar reputation" and that the absence of named career staff suggested the work might not reflect "the sound, reputable analysis for which the Census Bureau is known." The underlying voter data was supplied to the Bureau under an $885,000 contract by Data-Clear, a private Virginia-based business whose president, Carolyn Carlson, said she was surprised her company's data had been used this way. Justin Levitt, an election-law specialist at Loyola Law School and a former Biden White House voting-rights staffer, noted that in a record-matching exercise of this scale, a false-positive rate of even 0.019 percent "would turn the total conclusions to zero," given known issues with matching names and addresses across databases that are not updated in real time, including cases where an individual had naturalized after the citizenship snapshot used for matching. Notably, even taking the 24,000 figure at face value, it represents roughly 0.02% of the estimated 160 million ballots cast in 2020, in an election decided by margins of several million popular votes and tens of thousands of votes in the closest swing states — well below any threshold that would have been outcome-determinative.

Weighing the Record

Taken together, the recent audits do not strengthen the mass-fraud claim; they narrow it further. Every jurisdiction that has completed a full audit cycle — Georgia, Ohio's earlier cycle, Iowa, Michigan, Kansas under judicial scrutiny — has found confirmed noncitizen voting at rates measured in thousandths or ten-thousandths of a percent, consistent with the Brennan Center's 2017 finding of roughly 30 suspected incidents across 23.5 million votes surveyed (0.0001%) and with the Heritage Foundation's own fraud database, which after decades of collection contains on the order of dozens to roughly one hundred confirmed noncitizen-voting cases nationally. The 2026 Census Bureau figure is the largest number produced by any single analysis to date, but it has not yet been replicated, has not named its authors, and has been directly disputed on methodological grounds by people with no evident stake in the outcome — the same evidentiary standard, incidentally, that proponents rightly apply to the 2014 Richman study when arguing it overstated noncitizen voting. Pending independent replication of the Census Bureau's matching methodology, the audits that have gone through full investigation and adjudication remain the most reliable evidence available, and they consistently return the same verdict: the vulnerability proponents describe exists at the margins of the system, and has not been shown to occur at a scale capable of altering a federal election outcome.