Elections · analysis
Noncitizen voting: rare, and the crackdowns hit citizens
Noncitizen voting is already a federal crime, up to a year in prison and deportation, and it is prosecuted: the Heritage database lists 100 proven cases in 43 years, and the Justice Department says it has charged 70 people in voting-fraud cases in 2026. Since 2024, states have searched their rolls harder than ever, and every audit tells the same story. The noncitizen registrants that states report run from about 0.0002% to 0.013% of the rolls, and fewer still ever voted. Meanwhile the checks themselves flag eligible citizens, most often naturalized Americans whose records lag, by the hundreds or thousands. That is the case against proof-of-citizenship mandates in one table.
This section makes an argument, unlike the rest of this site: that proven voter fraud is far too rare to justify new barriers to voting. Every number is sourced, and the strongest evidence on the other side is shown, not hidden. The other side →
Official state audits and reviews. Share columns divide by the records checked (or, where a state didn't report that, marked *, its 2024 registered voters from the EAC). “Voted” windows differ a lot: read the last column. Click a header to sort.
| State (audit) | Records checked | Noncitizens registered | % of rolls | Of them, voted | Voted, % of rolls | Over what period |
|---|---|---|---|---|---|---|
| Alabama SAVE check | 3,868,040* | 186 | 0.0048% | 25 | 0.00065% | ever |
| Georgia 2024 citizenship audit | 8,200,000 | 20 | 0.00024% | 9 | 0.00011% | ever; none in Nov 2024 |
| Idaho 2024 SAVE check | n/a | n/a | n/a | n/a | n/a | n/a |
| Iowa 2024–25 audit | 2,300,000 | 277 | 0.0120% | 35 | 0.0015% | Nov 2024 general |
| Louisiana 2025 SAVE check | 2,900,000 | 390 | 0.0134% | 79 | 0.0027% | at least once since the 1980s |
| Michigan 2025 review of Nov 2024 | 7,200,000 | n/a | n/a | 16 | 0.00022% | Nov 2024 (of 5.7M ballots) |
| North Carolina 2016 post-election audit | 4,800,000 | n/a | n/a | 41 | 0.00085% | Nov 2016 (of 4.8M ballots) |
| Ohio 2024 + 2025 referrals | 8,074,098* | 1,084 | 0.0134% | 167 | 0.0021% | federal elections 2018–2024 |
| Oregon 2024 DMV error review | 3,000,000 | 1,617 | 0.0539% | 10 | 0.00033% | DMV error; 5 of the 10 were citizens when they voted |
| Pennsylvania 2026 Auditor General review | 200,000 | 1 | 0.00050% | n/a | n/a | n/a |
| Texas 2025–26 SAVE check | 18,000,000 | n/a | n/a | n/a | n/a | n/a |
| Utah 2026 full-roll audit | 2,069,640 | 27 | 0.0013% | 13 | 0.00063% | since 2018 |
| Virginia 2024 removals (EO 35) | n/a | n/a | n/a | n/a | n/a | n/a |
Texas, Idaho, and Virginia reported flags and referrals but no confirmed registrant or ballot counts; see their state pages. North Carolina's July 2026 SAVE run flagged 1,599 registrants, still under review; Georgia's 2026 run lists 432 potential noncitizen voters since 1996, also under review. Oregon's row counts people its DMV registered by mistake without proof of citizenship, some of whom are citizens.
From flagged to confirmed to voted, in the two audits that reported all three stages. Each state is drawn on its own scale, with everyone flagged as the full bar.
Proven noncitizen cases by state, Heritage database, 1982–2025.
Proven noncitizen cases by year of outcome, 2000–2025. The 2018 spike is mostly North Carolina cases from its 2016 audit.
Nationwide, 2025–26
The expanded federal SAVE system has run 65 million+ voter records from 26 states and flagged 28,635 as “potential noncitizens” 0.044%, before any review (per the Solicitor General's Supreme Court filing; the Court let SAVE checks resume on Sept 25, 2026). The Justice Department says it has charged 70 people in voting-fraud cases in 2026, including 16 in a single September sweep. Votebeat reviewed 41 recent cases against noncitizens: at least 23 defendants held green cards and 8 cases involved undocumented immigrants, from at least 18 countries. It found no sign of an organized effort. Immigration agents made 160 arrests from about 1,600 voter-fraud leads, many for related crimes such as falsely claiming citizenship.
The big numbers that didn't hold
Large flag counts rarely survive review. Iowa's 2,176 flags became 277 confirmed. Alabama's 2024 purge list of 3,251 included thousands of citizens, by the state's own admission, and a federal judge halted it; a separate check through the SAVE database removed 186 registrants who did not send proof of citizenship. Texas's 2024 claim of 6,500 “potential noncitizens” removed included people who simply didn't answer letters; only 581 were removed as noncitizens. A July 2026 DHS claim of more than 250,000 noncitizen registrants in four states rested on what the department's own letters to those states called a “preliminary review” that found there “may be as many as” that number. The other direction applies too: audits catch only people whose records match, so the confirmed counts are floors, not ceilings.