Is voter fraud a big problem?
Executive summary
The best available reporting and research show that voter fraud in U.S. elections is vanishingly rare and has not altered recent national outcomes; multiple studies put proven or referred instances at only a few hundred to a few thousand across decades while fraud rates per ballot are effectively zero [1] [2] [3]. That rarity does not mean fraud never occurs—small-scale cases can and do affect tight local contests—and allegations of widespread fraud have been amplified for political ends, influencing policy debates over voter ID and mail voting [4] [5] [6].
1. The empirical bottom line: fraud exists but is quantitatively tiny
Comprehensive reviews find extremely low rates of proven fraud: one Brookings synthesis cites a fraud rate of about 0.0000845 percent across large datasets and reports that no election outcome was altered by ballot fraud in the reviewed period [1], the Brennan Center’s review of multiple studies found rates effectively at zero in large samples [2], and NPR reported that analyses uncovered only about 475 potential cases across six battleground states—an infinitesimal share of more than 25 million ballots cast there [3].
2. Where fraud is found: mostly small-scale and local
When fraud shows up it is frequently tied to small local contests, administrative error, or isolated schemes rather than mass manipulation that could flip statewide or national races; Hearst’s investigative count identified only 14 cases in recent years where a court ordered a new election or overturned results, almost all at the municipal or county level [4], and Reuters and other analysts note that mail-ballot and in-person impersonation frauds are exceptionally rare and concentrated in few localities [7] [8].
3. Official tallies and databases: different emphases, similar conclusions
Conservative databases like the Heritage Foundation’s tally document roughly 1,500 proven instances over decades, and Newsweek and PBS reference those figures while noting convictions; even those counts amount to tiny fractions of the total votes cast and do not change the empirical assessment that large-scale fraud is not occurring [9] [6]. Independent fact-checks and academic work reach the same essential conclusion: isolated cases exist but systemic, outcome-changing fraud does not appear in the data [10] [2].
4. Why the perception of a big problem persists
High-profile allegations, political messaging, and selective use of statistics inflate public fears: prosecutions or referrals are sometimes cited as proof of widespread fraud without context about scale or whether votes were ever counted, and claims about vulnerabilities—such as a handful of successful impersonations in studies—can be amplified beyond their statistical significance [10] [2]. Policy proposals like strict voter ID laws are often justified by fraud concerns even as research finds impersonation and mail-ballot fraud are vanishingly rare [2] [11].
5. Policy trade-offs and open questions
The evidence supports safeguards—audits, signature verification, chain-of-custody rules—because even rare fraud is unacceptable, but lawmakers should weigh those protections against the demonstrated risk and the real possibility of disenfranchising eligible voters; the Brennan Center warns that policies aimed at preventing “phantom” fraud can suppress turnout, and several courts have struck down measures that lacked evidentiary support [5] [2]. Reporting and datasets are robust on frequency but less conclusive about every type of procedural vulnerability, so continued transparency, audits, and independent oversight remain necessary [8] [7].