Trump Keeps Railing Against Non-Citizen Voting. Research Shows It’s Extremely Rare

President Donald Trump has repeatedly insisted that non-citizen voting is a widespread and pressing issue in U.S. elections, a claim he has returned to in his recent push for stricter voter ID requirements.

In his State of the Union Address on Tuesday night, the President urged lawmakers to pass the Safeguard American Voter Eligibility, or SAVE America, Act, a bill that would require voters to provide proof of citizenship and photo identification to cast ballots in all 50 states, among other changes.

[time-brightcove not-tgx=”true”]

“Perhaps most importantly, I’m asking you to approve the SAVE America Act to stop illegal aliens and others who are unpermitted persons from voting in our sacred American elections,” Trump said. “That cheating is rampant in our elections. It’s rampant.”

Read more: Read Donald Trump’s Full State of the Union Speech

In the past, Trump has baselessly claimed that thousands, if not millions, of undocumented immigrants voted against him in his presidential bids. Following the 2016 election, Trump claimed that between 3 million and 5 million “illegal votes” caused him to lose the popular vote to Hillary Clinton. When he lost the 2020 election to President Joe Biden four years later, he alleged that Biden won in key states due to tens of thousands of non-citizen votes.

It’s already illegal for non-citizens to vote in federal U.S. elections, however, and research and recent state investigations have shown that noncitizen registration and voting occurs only very rarely, accounting for just a sliver of registered voters and an even more miniscule fraction of ballots cast. 

Here’s what to know.

‘Vanishingly small numbers’ that grow smaller with thorough investigation

The preliminary results of recent investigations into potential non-citizen voting that have been conducted in multiple states have yielded numbers that are a fraction of a fraction of those Trump has claimed.

Louisiana Secretary of State Nancy Landry announced in September that an ongoing investigation had found 390 non-citizens on the state’s voter rolls. That number represents roughly 0.01 percent of the 2.9 million registered voters in Louisiana. Of the 390 non-citizen registrants identified by the investigation, Landry said 79 had voted in at least one election.

Utah’s election office, meanwhile, said last month that it had not discovered any cases of non-citizens casting ballots in the state after months of investigating.

“We have not yet encountered anyone who voted illegally,” Utah Lt. Gov. Deidre Henderson said about her office’s review of the 2.1 million people on the state’s voter rolls.

Henderson noted that 486 voters had registration information that was incomplete or inaccurate and that a third of that group registered decades ago, when the state did not have the same requirements. 

“I anticipate that the vast majority of these 486 are in fact citizens and just need to update their information,” Henderson said. Even if every one of those voters were found to be non-citizens, they would account for only 0.02 percent of voters registered in Utah. 

Analysis by the Center for Election Innovation and Research (CEIR), a nonpartisan group that advises state and local election officials, has also found that the number of registered voters  who are flagged as possible non-citizens in states’ cursory reviews of voting rolls—which some states publicly announce prior to more thorough investigation—are often inflated.

In a review of claims of non-citizen voting by states through Dec. 31, 2025, CEIR’s analysis revealed “that initial claims alleging large numbers of noncitizen registrants or voters are almost certain to be a misleading overestimation of the actual number of records with unconfirmed citizenship,” the group said.

Investigations into claims of large numbers of possible non-citizens registered to vote, CEIR’S research shows, have typically demonstrated that at least some initial flags “were based on outdated, incomplete, or improperly matched data that incorrectly labeled eligible citizens as possible noncitizens.” 

But, the group found, “the small numbers that follow from proper investigation generally receive far less public attention” than the inflated earlier numbers.

In one example pointed to by CEIR, the Iowa Secretary of State’s office announced in October 2024 that 2,176 voter registration records were potentially linked to non-citizens. That figure was already extremely small relative to the total number of registered voters in the state, accounting for roughly one-tenth of one percent. But the revised number, published in March 2025, was far smaller: 277 confirmed non-citizen ballots, or one-eighth of the earlier figure.

“Even when taken at face value, the largest claims never identify numbers of possible noncitizen registrants or voters that amount to more than a few tenths of one percent of the number of eligible voters in a state,” CEIR states.

The group noted that in a recent update to its research published this month, it paid particular attention to states that are using a U.S. Citizenship and Immigration Services (USCIS) program in their efforts to verify voter eligibility. USCIS announced it was making its Systematic Alien Verification for Entitlements (SAVE) program—a tool intended to enable federal, state, and local governments to verify peoples’ citizenship and immigration status—available to be used in this way last spring. More than 20 states currently have agreements with the agency permitting them to do so; Landry, the Louisiana Secretary of State, highlighted that SAVE had been used in her state’s ongoing investigation. But CEIR found that “the widespread use of SAVE by states has not changed this pattern” of more thorough investigations finding that some early flags were caused by data shortcomings.

Effective safeguards already in place 

CEIR’s analysis has also found that safeguards and procedures states already have in place are effective in preventing non-citizens from registering and rooting out non-citizens who do make it onto voting rolls.

In one example cited by the group, a Michigan Department of State audit that began in December 2024 compared 7.9 million active state driving records for people of voting age against the 7.2 million active registered voters in the state and identified only 15 instances of “individuals who appeared to be noncitizens and who had cast a ballot in the 2024 general election.”

This number represents 0.00028% of the more than 5.7 million total ballots cast in  Michigan in that election. Of those 15 cases, 13 were referred to the state Attorney General for potential prosecution. 

“This is a serious issue, one we must address with a scalpel, not a sledgehammer,” Michigan Secretary of State Jocelyn Benson said in a statement at the time. “Our careful review confirms what we already knew – that this illegal activity is very rare. While we take all violations of election law very seriously, this tiny fraction of potential cases in Michigan and at the national level do not justify recent efforts to pass laws we know would block tens of thousands of Michigan citizens from voting in future elections.”

“The limited scope of even the most sweeping claims indicates that existing safeguards and procedures are broadly effective,” CEIR states. “The small number of noncitizens actually found underscores the robust checks already in place.”

The stricter requirements that would be introduced under the voter identification bill Trump is pushing lawmakers to pass, meanwhile, would likely prove challenging for many citizens to meet. 

A 2023 Brennan Center for Justice survey found that about 21.3 million voting-age American citizens, or 9.1 percent, do not have documents proving citizenship readily available, while roughly half of U.S. citizens do not have a passport.