So let’s take up the challenge. It’s true that some laws do get violated a lot. Is the law against non-citizen voting among them? Does enough such voting go on to sway many electoral outcomes? While it’s impossible to prove a negative, there are ways to assess the probabilities. And as we do, we will find ourselves circling around to another question: If Musk or Trump or Vance have good evidence that this is happening, why haven’t they presented it?
Imagining Fraud
The prohibition might seem hard for a would-be voter to miss. “Are you a citizen of the United States of America?” That’s the very first question on the federally prescribed form states must accept to register voters, and it’s followed by the instruction that if your answer is “no,” you should not fill out the rest of the form. Those who persist anyway will find, alongside the signature line, a warning that the “penalty of perjury” applies to false statements. “If I have provided false information, I may be fined, imprisoned, or (if not a U.S. citizen) deported from or refused entry to the United States,” a voter must sign. That might seem like a pretty strong disincentive for someone who daily fears being separated from family and livelihood by being thrown out of the country (or incarcerated).
If law enforcement can show that a noncitizen has so much as registered, let alone cast an actual vote, they’ve got possible grounds for a prosecution. And if, as Musk’s rants might have it, illegal votes are being mustered in a conspiracy so vast that “every” deportation means a lost vote, that’s a much bigger deal. Organizing the scheme would likely require the participation of many thousands of persons to coordinate the coaching and generation of individualized perjury while trying to make sure no one talks. Were any street-level operatives to be apprehended on such charges, the likeliest way to cop a favorable plea deal would be to offer to inform on higher-ups, and if some of those higher-ups were politically influential people, so much the more explosive. Rolling up a conspiracy of this sort that could be traced to a party leader or elected official would be a prosecutor’s dream, quite possibly career-making.
For four years, the Department of Justice reported to Donald Trump, who had inveighed against voter fraud. So far as I have been able to tell from news reports, its biggest resulting prosecution of noncitizen voting came in 2020 in North Carolina, where a federal grand jury, following a DHS investigation, indicted 19 persons of varying nationalities for voting in the state’s federal election. That’s 19 persons too many to have voted, assuming the charges panned out, but it’s unlikely that it changed any outcomes given that more than 3.6 million persons cast their ballot in North Carolina’s 2018 election.
Federal and State Investigations Fetch Up Little
One might also pause to note that the Trump administration created a commission on voter fraud, which, like every other player that has investigated the issue, was unable to document large-scale lawbreaking. (The Heritage Foundation’s much-cited database of voting irregularities, when recently checked, included about 85 cases involving noncitizens since 2002.)
State-level prosecutions in this area are equally rare. Are states, too, somehow in on the plot? It seems hard to believe all of them could be. To begin with, many states with large noncitizen populations like Texas and Florida have been run by Republicans for decades, as have their attorney generals’ offices. It’s also rash to imagine that Democratic law enforcers, merely because they are Democrats, would overlook misconduct by members of their party. The fact is that Democrats regularly prosecute fellow Democrats for election-related as well as garden-variety misconduct, even as Republicans regularly prosecute fellow Republicans.
(Partisan politics, in general, is overrated as a source of voter fraud. One of the ironies one finds while on the election integrity beat is that many of the genuine instances of organized ballot fraud and sharp practice are found in low-turnout local races, especially primaries and nonpartisan races in places like Philadelphia and Bridgeport, where they have the effect of boosting one candidate against a candidate of the same party.)
Some states report on the number of noncitizen cases they have found. Former Nevada Secretary of State Barbara Cegavsje, a Republican, said a statewide audit turned up three noncitizen voters in 2016, while the GOP-majority North Carolina Board of Elections reported that 41 noncitizens had cast ballots the same year in that much larger state. A dedicated ballot integrity task force appointed by Florida Gov. Ron DeSantis does not report having made noncitizen voting a major focus of its investigations. Meanwhile, a 2022 Georgia audit found that 2,258 persons lacking the proper citizenship qualification had attempted to register over a 25-year period, according to database matching, but that the state’s screening methods had prevented them from completing those registrations and so none had voted.
Mistakes and False Positives
Wait a minute. Database matching? Screening methods? Yes, you might never know it from the memes, but states can and do employ database checks to detect and prevent the addition of noncitizens to voting rolls. Unfortunately none of these checking methods are perfect, and they generate both false positives and false negatives. Aside from larger databases, administrators may (for example) obtain from court authorities a list of persons who ask to be excused from jury duty on the grounds that they are not citizens. Not infrequently these people turn out to have been fibbing to the court clerk to get out of jury service and are in fact native-born citizens—thus generating a false positive. Other false positive matches can arise because someone omits to check the “citizen” box on a driver’s license application even though they are in fact a citizen, or because databases take a while to catch up after someone becomes a U.S. citizen through naturalization.
Even with encouragement—such as states offering ways for poll workers to report concerns of unlawful voting, or setting up hotlines for the public, as Florida has done—these methods turn up very few cases fit for prosecution. Where bad data is not the problem, mistakes can be: in “motor-voter” registration, for example, it can happen that someone mistook the boxes on the DMV form and inadvertently checked the voter registration box while trying to register a vehicle. But someone who didn’t intend to register is not a likely voter.
States aren’t the only ones using database matching to try to identify unlawful noncitizen voters. Amateur “Stop the Steal” fraud-chasers do so too. One report, from a group led by a member of Trump’s former election integrity committee, claimed that noncitizens had “succeeded in registering to vote by the thousands” in Virginia. Oops! Four persons born in the United States but identified as otherwise in the report came forward and sued for defamation, saying the report had in effect tagged them as felons based on data or recordkeeping deficiencies. The case eventually settled.
One study that Trump supporters often cite did arrive at an anomalously high estimate that as many as one in six noncitizens vote. There is reason to believe that data quality problems account for that extreme, outlier result, and in fact the lead scholar in the work, after being granted access to Arizona data as part of recent litigation, embraced a drastically reduced estimate of noncitizen voting incidence, saying that it had been too small to affect even the quite close 2020 Arizona presidential.
What do election administrators themselves think? For a 2017 report, the left-of-center Brennan Center interviewed election officials in 42 jurisdictions, many from some of the largest non-citizen populations in the country. Together the officials had collectively overseen the tabulation of 23.5 million votes in the 2016 general election. How many incidents of suspected noncitizens voting did they refer for further investigation or prosecution? Some 30! All but two of the officials reported that their jurisdiction used anti-fraud safeguards. While a few of the officials believed it would be a good idea to tighten precautions further, “no administrator reported that noncitizen voting was common.” In the three states Trump had specifically accused of being tainted by illegal voting, California, Virginia, and New Hampshire, Brennan reports that no official it interviewed “identified an incident of noncitizen voting in 2016.”
Statistical Proof
A common way of monitoring the integrity of elections is to apply statistical tests, looking for instances in which electoral results differ inexplicably from otherwise logical patterns. In this case, given common knowledge that certain localities have a high share of noncitizen residents, we can ask whether the number of votes cast relates reasonably to the known number of legally eligible voting-age citizens, applying plausible turnout expectations, or whether it comes in unaccountably higher than that, which might suggest seepage from the pool of ineligible persons. Tests of this sort “show nothing like widespread noncitizen voting,” Brian Quinn, a specialist in the evaluation of public sector statistics based in Wisconsin, told me. California is an example: It is a state where certain counties are universally assumed to be home to far more than their share of immigrants, including urban Los Angeles County as well as counties in the agricultural Central and Imperial valleys. “If a material number of those noncitizens were voting, total voters as a share of known eligible voters would be high.” It’s actually quite low as a share of legally eligible voters, probably due largely to low median age.
The 22nd California congressional district, for example, represented by Republican David Valadao, consists in substantial part of agricultural areas in the Central Valley with a very high share of noncitizens. It cast only about 100,000 votes in the closely contested 2022 midterms, an election in which many other districts around the state recorded 200,000 or even 300,000 votes. The same pattern, Quinn says, in which participation rates fail to come in suspiciously high when compared with the pool of lawful voters, “is consistent in every state. You can see it in Texas, Arizona, Florida, New York, and so forth.”
The claim that illegal voting is swaying American elections is nothing if not sensational. Those who levy sensational charges should bear the burden of proving them. But they haven’t. It’s just assertion after assertion, with no refutation of the considerable evidence to the contrary.
Over the past four years, through a long succession of court cases, audits, and studies, the props have been kicked out from under #StopTheSteal contention one after another: that voting machine tabulations are being hacked, that hordes of dead or nonexistent persons or ineligible felons vote, and on and on. Now we’re on to a claim of massive noncitizen voting that cleverly dovetails with public anxiety over immigration generally.
Bogus claims of widespread voter fraud, even when they do not stoke hatred and fear of the foreign-born, are grossly irresponsible. They exacerbate polarization and malign honest election administrators. Most of all, they undermine public confidence in our election system. The more people believe elections are rigged, the more they are likely to turn their discontents in a direction other than electoral politics. Some will go the passive route of resignation, withdrawing from civic involvements, making themselves the perfect subjects for strongman rule. Others will turn to militia activity or outright violence.
Either way, the consequences for the American experiment in liberal democratic self-rule will be unfortunate.
