DMV Glitch Quietly Warps NJ Elections

Voters at polling booths with American flag in background — Photo: vesperstock / Shutterstock

The New Jersey case of 6,600 non‑citizens mistakenly added to voter rolls is best understood not as proof of systemic voter fraud, but as a stark illustration of how poorly designed registration technology can create real, but limited, risks for election integrity—and political narratives eager to exploit them.

Key Points

About 6,600 non‑citizens were wrongly registered in New Jersey between June 2023 and June 2024 due to a Motor Vehicle Commission software flaw, with fewer than 400 actually casting ballots.
The error arose inside an automatic voter registration workflow at the DMV, where applicants correctly answered they were not citizens but were nevertheless registered; the glitch was fixed in 2024, before the current governor took office.
Election‑law research and multi‑state audits show that such episodes almost always stem from bureaucratic or technical error, with non‑citizen voting representing a microscopic share of total ballots.
Partisan actors and some federal officials have amplified the New Jersey case to argue for sweeping claims of widespread illegal voting, but have not produced specific, verifiable evidence matching those larger numbers.
The durable lesson is that election systems need rigorous technology controls and transparent audits—not panic about rampant fraud, nor complacency about the reality of rare but consequential mistakes.

What Actually Happened in New Jersey’s Voter Registration System

According to Governor Mikie Sherrill, a “serious software error” in New Jersey’s Motor Vehicle Commission (MVC) registration system led to roughly 6,600 individuals who had indicated that they were not U.S. citizens being nevertheless added to the voter rolls between June 2023 and June 2024. The problem surfaced in the context of automatic voter registration: when residents applied for a driver’s license or ID, they were asked on a keypad whether they were U.S. citizens and whether they wished to register. Many answered “no” on the citizenship question, but the system still treated them as eligible registrants.

Critically, this was not a case of applicants lying about their citizenship status. State officials have consistently described these individuals as having answered truthfully and being registered “through no fault of their own.” The flaw was in the software logic and its integration with the voter registration database. The vendor’s program, which was supposed to prevent non‑citizens from being registered, did not reliably do so; the error persisted for about a year before being corrected in June 2024, and it predated Sherrill’s administration by nearly three years.

Magnitude of the Error: 6,600 Registrations, Fewer Than 400 Votes

The number that rightly grabs attention is the 6,600 erroneous registrations. The follow‑on number that matters for election outcomes is how many of those individuals actually voted. The state’s preliminary analysis indicates that fewer than 400 of the mistakenly registered non‑citizens cast ballots while on the rolls. Those voters were not concentrated in a single jurisdiction or party; they were scattered around the state, registered as Democrats, Republicans, and unaffiliated.

To understand scale, New Jersey logged more than 4 million votes in the 2024 presidential election alone. Even if every one of the roughly 400 non‑citizen votes were cast in that race—which is not what the data show—they would represent well under one in 10,000 ballots. Sherrill has stated that the state has “no evidence at this time that any elections were swayed” by these votes. Independent outlets reviewing the episode have reached similar conclusions: the numbers are accurate, the registrations were erroneous, and the evidence so far points away from organized fraud.

At the same time, the fact that hundreds of ineligible votes were counted is not trivial. Federal law bars non‑citizens from voting in federal elections, and one improper ballot can matter in close local contests. That is why the state ordered all 6,600 registrations removed and opened an investigation into how the error persisted for a year and why it took another two years for the public to be informed.

Mechanism: How DMV Software Can Create Ineligible Voters

New Jersey is one of more than twenty jurisdictions that tie voter registration to motor‑vehicle transactions, relying on electronic prompts and data matching to enroll eligible voters efficiently. When such systems work properly, they reduce paperwork, increase registration among eligible citizens, and standardize data. When they fail, the very efficiency they create becomes a liability. A single misconfigured rule can quietly scale an error across thousands of records in months.

In the New Jersey case, the critical failure was at the point where citizenship status was captured and transmitted. Applicants interacted with a multilingual PIN pad, answered “no” to a citizenship question, yet were still flagged for addition to the statewide voter file. That suggests either an inverted or ignored Boolean flag in the software, a faulty integration with the election database, or a misinterpretation of a field by downstream systems. The vendor—identified in coverage as IDEMIA—was supposed to provide a safeguard that filtered out non‑citizen records before they reached election officials; instead, its product became a conduit for error.

Election‑administration scholars have warned for years that combining “licenses for all” policies (which rightly allow non‑citizens to obtain driver’s licenses) with automatic voter registration requires meticulous citizenship checks and regular audits. Where those controls are weak or untested, non‑citizens can be added to the rolls without intent or awareness—exactly what New Jersey’s experience now illustrates.

How Rare Is Non‑Citizen Voting? The Broader Evidence

It is easy to seize on the New Jersey numbers and extrapolate to claims of widespread illegal voting. The broader empirical record does not support that leap. Reviews of alleged non‑citizen voting in multiple states—using forensic audits, federal criminal prosecutions, and independent research—repeatedly find that non‑citizen votes represent a microscopic fraction of total ballots cast.

For example, Pennsylvania officials identified 544 votes by non‑citizens over a 17‑year period, out of more than 93 million votes—a rate near 0.0006%. A multi‑state study by the Brennan Center and others, examining 23.5 million votes in the 2016 general election, found roughly 30 incidents of suspected non‑citizen voting referred for further review, or about 0.0001%. A national review of state investigations concluded that “actual incidents of noncitizens voting is not a significant occurrence.”

Even organizations critical of current registration systems inadvertently confirm the rarity. A paper from the Public Interest Legal Foundation, often cited in debates over New Jersey, reported alleged non‑citizen registrations but showed that most involved individuals who accurately disclosed their status and were nonetheless processed by government officials. As the Brennan Center’s analysis of that paper put it, “the system predominantly works,” and the problems that do exist largely reflect bureaucratic missteps rather than deliberate schemes.

Against that backdrop, New Jersey’s discovery fits a recognizable pattern: a sizeable but finite pool of erroneous registrations created by a technical flaw, yielding a small number of improper votes, and no evidence that those votes were cast with fraudulent intent or changed any major race.

Competing Narratives: Fraud Allegations vs. Error Explanation

The episode has unfolded amid a national political struggle over election integrity, and it has been promptly enlisted into broader narratives. On one side, New Jersey officials and election‑law experts present the case as a serious but bounded failure of technology and administration: identify the flaw, remove the ineligible registrants, investigate vendor performance, and strengthen safeguards. Their stance is that transparency and remediation are the proper response, not claims of a stolen election.

On the other side, partisan actors and some federal officials have leveraged the news to argue that non‑citizen voting is widespread. DHS Secretary Markwayne Mullin has claimed that roughly 250,000 non‑citizens are registered across four states, including New Jersey, and has threatened to tie federal election funding to participation in federal data‑matching programs.[transcript 1][transcript 3] Yet when pressed by state officials, DHS has not disclosed the specific names or methodology behind those figures, and states such as Pennsylvania have called them “wildly speculative at best.”

Similarly, New Jersey’s Republican Party has publicized findings of “hundreds” of non‑citizens on the rolls based on county records, but has not released detailed documentation—names, dates, or verified citizenship status—that would allow direct comparison with the 6,600‑person software cohort or with official analyses. Media outlets ideologically invested in fraud narratives have framed the error as “blue state exposed” and a vindication of prior warnings, while glossing over the distinction between registration errors and intentional illegal voting.

The key point for a skeptical, data‑driven reader is that these broader claims have not yet been substantiated with the kind of specific, auditable evidence that New Jersey has provided for the 6,600 erroneous registrations and fewer than 400 votes. At present, the software‑error explanation is supported by named officials, a clear time window, and a plausible technical mechanism; the claims of much larger pools of non‑citizen voters rest on opaque data matching and have been challenged by state election administrators.

Accountability, Technology, and What Should Change

Even when the scale of improper voting is small, episodes like New Jersey’s erode public confidence if they are not handled decisively. Sherrill has directed her chief counsel to conduct a comprehensive investigation, ordered removal of all erroneously registered individuals, and begun the process of replacing the vendor responsible for the flawed system. The MVC is adding layers of safeguards, including more robust data review and improved interface design.

From an election‑administration standpoint, several durable lessons emerge. First, automatic voter registration systems must treat citizenship status as a hard gate, not a soft flag. That requires clear separation between licensing databases (which appropriately include non‑citizens) and voter eligibility rules, with regular reconciliation against authoritative sources. Second, any time a state expands or modifies registration technology, it should be accompanied by high‑frequency audits—sampling registrations, cross‑checking citizenship where lawful and appropriate, and stress‑testing edge cases before full deployment.

Third, the public conversation needs a more precise vocabulary. “Thousands of non‑citizens on the rolls” and “thousands of non‑citizens voting” are different claims, with different implications. The New Jersey case demonstrates that erroneous registration is possible at scale, but also that the step from registration to actual voting is much rarer and often unintentional. Conflating the two obscures where the real vulnerabilities lie, which is in upstream data handling and software quality rather than hordes of non‑citizens conspiring to tilt elections.

What It Means Going Forward

For citizens trying to make sense of competing headlines—Jake Tapper noting this is “not good for Democrats,” DHS alleging hundreds of thousands of non‑citizen registrants, governors insisting their elections remain secure—the New Jersey episode offers a useful lens. It shows that one can take election integrity seriously without endorsing sweeping, unsupported claims of rampant fraud. It also shows that a state can acknowledge real, uncomfortable errors without collapsing into crisis.

The durable truth, borne out by multi‑state data, is that almost all votes in American elections are cast by eligible citizens, and that the small number of exceptions typically arise from confusion or administrative mistakes, not coordinated schemes. The New Jersey software glitch is a cautionary tale about the importance of robust technology and transparent oversight, not a demonstration that the broader system is rotten. Strengthening that system will require investment in technical competence and institutional accountability more than it will require new waves of fear about non‑citizen hordes stealing elections.