# Restoring Election Integrity: Oversight of Voter Fraud in New Jersey To date, all evidence suggests that noncitizens voting in American elections is a minute phenomenon. September 15, 2026 • Testimony By [Stephen Richer](https://www.cato.org/people/stephen-richer) **Subcommittee on Oversight** **Committee on the Judiciary** **United States House of Representatives** Chairman Jordan, Ranking Member Raskin, Chairman Van Drew, Ranking Member Crockett, and Members of the Subcommittee, thank you for the opportunity to speak with you today. I served as the elected, Republican, Maricopa County Recorder from January, 2021 until January, 2025, where I was responsible for, among other things, voter registration and early voting for the 4.6 million people in that county. I am now a legal fellow at the Cato Institute, a lecturer and fellow at the Harvard Kennedy School’s Ash Center, and the CEO of Republic Affairs, a consulting firm for civil society nonprofits. I received my master’s degree and law degree from The University of Chicago. Let me begin by saying that I do not think there is anything inherently wrong with requiring documented proof of citizenship to vote in U.S. elections. Arizona, my home state, has had a [documented proof of citizenship requirement](https://www.azcleanelections.gov/federal-only-voters) since January 2005. It is imperfect, and implementation has presented many challenges, but it is a fundamentally sound system. But as with almost all public policy proposals, there are tradeoffs. There are costs, and there are benefits. To make these calculations, we must first have a clear understanding of the facts. To date, all evidence suggests that noncitizens voting in American elections is a minute phenomenon. Every well-documented voter roll audit from the past six years has found that non-citizens make up a fraction of a percent of registered voters, and noncitizens who actually vote make up an even smaller percentage. I am aware of at least 29 audits performed by states and counties since the beginning of 2021 that have searched voter roll lists for noncitizens. I have made those available on the [Declare News website](https://declarenews.org/map-investigative-data/). The most recent audit is one published earlier this month by the [Georgia Department of Audits & Accounts](https://www.audits2.ga.gov/reports/summaries/voter-registration-and-list-maintenance/). That audit found 407 possible noncitizen voters over the past 30 years. Again, that is just possible noncitizen voters. Not confirmed. And from over the past 30 years. They did not all vote in the same election. Georgia has approximately 8 million registered voters. President Trump received approximately 2.66 million votes in Georgia in 2024. Vice President Harris received approximately 2.54 million votes in Georgia in that election. In May of this year, Utah [published](https://ltgovernor.utah.gov/wp-content/uploads/CITIZENSHIP-FULL-SUMMARY-2.pdf) its findings from a year-long audit of the state’s 2.1 million registered voters. The state found 27 confirmed noncitizens on the voter rolls. That’s.001 percent. In August, Yavapai County, Arizona [published](https://www.yavapaivotes.gov/files/assets/elections/v/1/documents/my-experience-using-save-final.pdf) its findings from a comprehensive audit of its approximately 200,000 registered voters. It found 30 registered noncitizens, four of whom had voted in a past election. And in Michigan, the Secretary of State recently audited the approximately 725,000 thousand registered voters in Macomb County. That audit found [four](https://www.michigan.gov/sos/resources/news/2026/01/29/secretary-benson-macomb-county-clerks-reckless-accusations-put-eligible-michigan-voters-at-risk) “apparent noncitizens,” including [one](https://bridgemi.com/michigan-government/michigan-macomb-clerks-noncitizen-voter-claims-overstated/) who had actually voted. No thorough, professional audit has found that noncitizens constitute anything even close to a tenth of a percent of the state’s voter population. Recent federal law enforcement efforts, many of which have been done at the behest of President Trump, affirm the comparatively microscopic nature of noncitizen voting. According to reporting from [The New York Times](https://www.nytimes.com/2026/09/02/us/politics/trump-election-voter-fraud.html), [Fox](https://www.foxnews.com/politics/watch-illegal-alien-busted-battleground-state-allegedly-committing-voter-fraud), and [Reuters](https://www.reuters.com/world/us/thousands-agents-diverted-trump-immigration-crackdown-2025-03-22/), the federal government has mobilized hundreds of highly-skilled law enforcement agents – the type of people who often investigate financial fraud or terrorism – to investigate election fraud, especially non-citizen voting. The results have been entirely consistent with the audits I just mentioned. According to [reporting from ProPublica](https://www.propublica.org/article/noncitizen-voting-trump-homeland-security-investigations), between January 2025 and May 2026, the federal government charged 41 noncitizens with voting illegally or other election-related crimes. Numbers produced since then have been similar in scope. On September 6, FBI Director Kash Patel took to social media to [announce](https://x.com/FBIDirectorKash/status/2096595077296537795?s=20) charges against five individuals across three states who had voted in either the primary or general election in 2022 or 2024. Director Patel [further wrote](https://x.com/FBIDirectorKash/status/2096595077296537795?s=20) that “Election integrity is a top priority for this FBI. We will investigate every credible allegation of election fraud and illegal voting.” Seemingly, the absence of widespread unmasking of noncitizen voting is not for want of effort from the FBI. Those who allege widespread noncitizen voting have little support for the position. They often point out that in most states, documented proof of citizenship isn’t required. And while that’s true, that doesn’t mean that noncitizens are actually registering or voting. Or they point to indiscriminate analyses that use commercial data and soft data matching techniques – that is, they use voter files without personal identifying information such as social security number or a driver’s license number – that allege a certain number of noncitizens voting, but never show their work and never investigate individual voter files. Those types of reports seem to be more for show than for election integrity. Turning to the specific situation of New Jersey. It is the largest finding of registered noncitizens of which I am aware. But by all reports, it is not an example of immoral criminals who wanted to cheat the system and were successful. It was instead the result of a software problem caused by imperfect human beings. What we know so far is that 6,600 noncitizens visited the New Jersey Motor Vehicle Commission to get a driver’s license, as they are lawfully allowed to do. They checked a box saying they are not citizens. They did the right thing. Owing to a programming error, that selection got misrecorded. They were instead marked as citizens residing in New Jersey. Accordingly, consistent with New Jersey law, election officers received the information from the Motor Vehicle Commission and registered those non-citizens. This was a mistake. And it should be fixed, as with any other mistake that happens in government. But some have used this software error to make bold, sweeping statements about American elections. Some have said, “see, this is why we need documented proof of citizenship.” But that wouldn’t have fixed the problem. I know this because in Arizona, we have documented proof of citizenship, and yet we had a very similar problem when the software tying the motor vehicle division with the state election office said certain voters had provided documented proof of citizenship, when in fact they hadn’t. And some say “see, this is proof of widespread noncitizen voting.” But, again, this was not the result of some nefarious plot, but instead a coding mistake. 6,600 people got improperly registered. [340](https://www.njspotlightnews.org/2026/08/nj-to-disclose-records-on-how-foreign-nationals-registered-to-vote/) of those voted in at least one election. But for context, there are 6.68 million registered voters in New Jersey. And 4.3 million of them voted in November 2024. Even in this scenario, this worst case scenario, noncitizens made up a fraction of a percent of voters. American elections are accessible, secure, and accurate. We should be proud. Nonetheless, we should continue to improve American elections. And we should have conversations about documented proof of citizenship. But we must be honest about the reality of noncitizens voting. Thank you, Stephen Richer ##### About the Author [![Stephen Richer](/sites/cato.org/files/styles/author_picture/public/2025-04/Richer_Stephen.jpg?itok=xnKCC9Fz)](/people/stephen-richer) ##### [Stephen Richer](/people/stephen-richer) Senior Practice Fellow in American Democracy, Ash Center for Democratic Governance and Innovation, Harvard Kennedy School [![Creative Commons License](/build/cato_2020/images/creative-commons.svg)](http://creativecommons.org/licenses/by-nc-sa/4.0/) This work is licensed under a [Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License](https://creativecommons.org/licenses/by-nc-sa/4.0/).