How frequent is voter fraud in federal elections
Executive summary
Extensive, peer-reviewed and investigative reporting finds voter fraud in U.S. federal elections is vanishingly rare—measured in fractions of a percent, and typically so small it could not have changed outcomes in contested federal races [1] [2] [3]. Multiple long-form analyses and databases show isolated incidents across decades but no evidence of widespread, coordinated fraud that would flip a presidential or congressional result [4] [5].
1. What the numbers say: microscopic rates, not mass conspiracies
Comprehensive reviews put the rate of voter fraud in federal contests at essentially zero for practical purposes: one well-known Brennan Center analysis estimated fraud rates between roughly 0.0003% and 0.0025% based on decades of data, and investigators have repeatedly found only a few hundred potential problematic votes across tens of millions cast in recent presidential contests [6] [7] [3]. The Heritage Foundation’s own historical listings—when reanalyzed—produce only tiny absolute counts relative to the billions of votes cast, and academic work concludes that the statistical claims offered as evidence of large-scale fraud do not hold up under scrutiny [4] [8] [2].
2. How reporters and researchers count cases—and what they find
Investigations such as the Associated Press’ post-2020 audit identified fewer than 475 potential instances of fraud across six battleground states—an amount far smaller than margins that decided those states—while national research consistently describes fraudulent voting as “infinitesimally rare” or “very rare,” with many flagged incidents turning out to be mistakes, administrative errors, or lawful confusion rather than intentional, coordinated theft [3] [9] [1]. Election officials and scholars emphasize that decentralization, transparency, and multiple verification steps mean most attempted fraud is either detected, reversed, or too few in number to influence federal results [10] [11].
3. Types of fraud observed—and their typical scale
The handful of documented offenses tend to be straightforward, local crimes—double voting, voting on behalf of deceased persons, or mismatched absentee ballots—rather than systemic manipulation of machines or mass noncitizen voting; courts and prosecutors have pursued these cases, but they occur at a scale that risks local offices or contests more than national outcomes [12] [5] [13]. Studies and fact-checking organizations note that many high-profile allegations (e.g., machine switching, millions of illegal ballots) failed statistical or evidentiary tests and were dismissed in multiple lawsuits and academic analyses [2] [14].
4. How evidence and claims diverge: politics, perception, and consequences
Political actors and media ecosystems have amplified rare incidents into narratives of stolen elections; scholarly and journalistic reviews show this amplification often outruns the underlying facts—leading to policy changes and public distrust that are disproportionate to the empirical record [5] [1]. At the same time, reporting also documents honest limits in the public record: databases and audits vary in scope and methodology, so precise counts differ across sources even as the qualitative conclusion—fraud is minimal—remains consistent [4] [8].
5. Bottom line: frequency, impact, and where questions remain
For federal elections, voter fraud is exceedingly uncommon—measured in fractions of a percent and almost never large enough to alter presidential or congressional outcomes—though isolated cases can and do occur and can matter more in close local contests [6] [3] [13]. Existing evidence comes from a mix of investigative reporting, academic analysis, and official records; it robustly undercuts claims of mass fraud but does not eliminate the need for continued auditing, transparency, and prosecution of genuine offenses, nor does it resolve all methodological debates about how best to count every single irregularity [2] [10] [14].