FACT-CHECK: The Myth of Non-Citizens Voting en Masse, Dismantled With Hard Data

A law built on a ghost

On February 11, 2026, the House passed the SAVE America Act — Safeguard American Voter Eligibility — by a margin of 218 votes to 213. The premise: millions of undocumented immigrants are allegedly flooding the ballot box, corrupting American elections, and stealing votes from legitimate citizens. The premise sounds explosive. It is also, according to every serious study conducted over the past four decades, statistically false.

On June 4, 2026, the Senate rejected it: 48 votes in favor, 50 against. Four Republicans — Collins, Murkowski, McConnell, and Tillis — crossed the aisle. That defection, in today's political climate, speaks volumes. When senators from the same party break ranks over a voter fraud bill, it means the numbers don't hold up. It means someone looked at the data and blinked.

The bill that was going to fix everything

The SAVE Act would have required every voter registration applicant to present a passport or birth certificate in person. It would have abolished online registration, abolished mail-in registration, and exposed every election official who registered a voter incorrectly to criminal penalties. A sweeping, radical overhaul — to solve a problem that, according to the data, affects approximately 2.3 cases per year across the entire country.

At the same time, Judge Sooknanan ruled on June 22, 2026, that the extended use of the SAVE database — deployed against 67 million voter records — was illegal. One day after the Senate vote failed, a federal court confirmed that the enforcement arm of this fiction was itself operating outside the law.

I have spent years watching politicians manufacture crises from scratch. But few achieve this level of statistical audacity: building a law that could strip 21 million citizens of their right to vote in order to address a problem documented in fewer cases than a mid-sized city has car accidents on a long weekend.

A number that demolishes the narrative

In 2016, the Brennan Center for Justice conducted one of the most rigorous studies ever produced on non-citizen voting. Researchers contacted 44 election administrators across the United States, covering a combined total of 23.5 million votes cast. The result: 30 suspected cases of non-citizen voting. Thirty. Out of twenty-three and a half million.

That is a rate of 0.0001%. To put that in perspective: you are statistically more likely to be struck by lightning twice in the same year than to encounter a non-citizen voter at your polling station. The Brennan Center updated this research in 2024. The conclusion had not changed. The phenomenon exists at the margins of the margins.

The New York Times survey: 2 possible cases in 137.7 million votes

Following the 2016 presidential election — the one that supposedly triggered this entire moral panic — The New York Times conducted its own investigation. Reporters surveyed election officials in 49 states, covering 137.7 million votes cast. They found two possible cases of non-citizen voting. Two. Across forty-nine states.

Not two thousand. Not two hundred. Two. And the word "possible" was doing considerable work in that sentence — neither case had been confirmed as intentional fraud. This is the electoral crisis that justified hundreds of hours of congressional debate, the threat of criminal charges against election workers, and the potential disenfranchisement of tens of millions of American citizens.

When a New York Times survey of 137.7 million votes turns up two possible cases, the story is no longer about fraud. The story is about what someone needed this fraud to justify.

When the right's own evidence undermines the argument

The Heritage Foundation — a conservative think tank that has spent years documenting electoral fraud and is by no means sympathetic to the Democratic Party — maintains the most comprehensive database of proven fraud cases in the United States. It is the right's own scoreboard. Their number, after forty years of data collection: 68 proven cases of non-citizen voting.

Sixty-eight. In forty years. Across a country of 330 million people. That works out to an average of 1.7 proven cases per year. And of those 68 cases, only 10 involved undocumented immigrants. The rest were legal residents — green card holders, visa holders — who misunderstood their eligibility or were misinformed by registration forms. The deliberate, large-scale fraud that the SAVE Act was designed to stop does not appear in the Heritage Foundation's own numbers.

State-by-state: the reality behind the claims

The state-level data is equally deflating. In Ohio, since the year 2000, authorities identified 44 non-citizen voters — over a 26-year period in a state with millions of registered voters. In Colorado in 2012, an initial alarming claim suggested 11,805 non-citizens had voted. The actual confirmed figure, after investigation: 35. In Nevada in 2016: 3 cases in 1.1 million votes cast. In North Carolina, between 2002 and 2012 — a full decade — authorities found 58 cases in 19.5 million votes.

These numbers do not describe a crisis. They describe a statistical artifact so small it is barely distinguishable from administrative error, voter registration confusion, or database mismatches. The gap between the political rhetoric and the documented reality is not a rounding error — it is an abyss.

Colorado's 11,805 becomes 35 after investigation. Ohio's 40-year total barely reaches 44. Nevada delivers 3 cases in over a million votes. These are not the numbers of a rigged democracy. They are the numbers of a manufactured emergency — and everyone who reads them knows it.

A cure far worse than the disease

The SAVE Act's requirements were straightforward: to register to vote, you would need to present a passport or birth certificate in person at a government office. Online registration and mail-in registration would be eliminated. For tens of millions of Americans, this would have been a decisive barrier. The Brennan Center estimated that 21 million American citizens do not have ready access to one of these documents.

The Center for American Progress put the number of Americans without a valid passport at 140 million. The Brennan Center also noted that 69 million American women would face additional complications due to name changes after marriage — their birth certificate name no longer matching their current legal identity. These are not edge cases. These are the realities of everyday American life, distributed unevenly across racial, economic, and geographic lines.

Criminal penalties for election workers

Perhaps the most chilling provision of the SAVE Act was its treatment of election officials. Under the bill, any election administrator who registered an ineligible voter — even through honest administrative error — would have faced criminal charges. In an already-pressured environment where election workers across the country have faced threats, harassment, and resignation, this provision would have functioned as a deterrent not against fraud but against civic participation itself.

New Hampshire provided a concrete preview of these dynamics. Following early SAVE-related enforcement actions, confusion among eligible citizens created situations where legitimate voters were unable to register or vote. A federal court ruling on May 29, 2026 was required to reverse the damage. The law hadn't even passed — and its enforcement apparatus was already producing collateral disenfranchisement.

There is a word for a policy that demonstrably produces more harm to eligible voters than it prevents among ineligible ones: voter suppression. The SAVE Act's architects knew the numbers. They built the law anyway.

The strongest deterrent already exists

Lost in the political debate is a basic legal reality: non-citizen voting is already a serious federal crime. Any non-citizen who votes in a federal election faces up to five years in federal prison, immediate deportation, and a permanent bar to ever becoming a naturalized American citizen. This is not a fine. This is not a misdemeanor. It is a life-altering criminal conviction that permanently closes the door to legal residency and citizenship.

For a green card holder who has spent decades building a life in the United States, raising children, paying taxes, and waiting in line for citizenship, the rational calculus of casting a single ballot illegally is — by any reasonable analysis — insane. The incentive structure already makes non-citizen voting one of the most self-destructive acts imaginable. The SAVE Act was a solution in search of a problem that the existing criminal code had already made nearly impossible.

The asymmetry nobody discusses

Professor Steve Vladeck of Georgetown University Law School distilled the problem with surgical precision: "The problem this law seeks to fix occurs so rarely that the proposed solution would be far more harmful than the problem itself." That sentence should have ended the debate. It did not, because the debate was never primarily about fraud.

The asymmetry is stark: on one side, 68 proven cases in 40 years, with existing criminal penalties that are already severe. On the other side, a law that would have placed documentary barriers between 21 million citizens and their constitutional right to vote, criminalized good-faith errors by election workers, and eliminated the registration methods used by millions of low-income and rural voters. The trade-off was not even close.

A government that imposes five-year federal prison sentences on non-citizens who vote has already built a deterrent. What the SAVE Act added was not deterrence — it was obstruction, aimed at a different target entirely.

The 6.4% claim that collapsed under scrutiny

In 2014, political scientist Jesse Richman published a study claiming that approximately 6.4% of non-citizens had voted in recent American elections. For advocates of strict voter ID laws and non-citizen voting restrictions, this was a revelatory number — proof that the problem was not marginal but systemic. The study circulated widely in conservative media and was cited repeatedly in congressional testimony.

It was fully invalidated in 2015. Multiple independent researchers demonstrated that Richman's methodology was critically flawed: his data drew from the Cooperative Congressional Election Study, a survey that includes a small number of respondents who misidentify their citizenship status. The "non-citizens" in his sample were almost entirely American citizens who had made a data entry error. Once the misclassified respondents were properly accounted for, the 6.4% figure evaporated. Peer review demolished it. The study's own co-authors issued corrections.

Why a debunked study keeps circulating

Despite its comprehensive invalidation, the Richman study continued to surface in legislative debates, think tank publications, and congressional testimony through 2025 and into 2026. This is not an accident of scholarly confusion. It is a case study in the political life of a useful lie: once a false statistic has been embedded in a political argument, correcting the record requires far more effort than spreading the original claim.

The SAVE Act's congressional sponsors never meaningfully engaged with the body of evidence that contradicted their premise. They cited the Richman figure. They cited isolated state investigations that had been either exaggerated or discredited. They did not cite the Brennan Center. They did not cite the Heritage Foundation's own forty-year database. When your own side's evidence refutes your conclusion, you do not cite your own side's evidence.

A study debunked in 2015 was still being cited to justify federal legislation in 2026. That is not a failure of information — it is a choice. Someone decided the lie was more useful than the truth, and built a law on top of it.

The enforcement mechanism that overreached

Parallel to the legislative battle over the SAVE Act, the Department of Homeland Security extended the operational use of the SAVE database — the federal citizenship verification system — to cross-reference state voter rolls. The scope was extraordinary: 67 million voter records were run through the system in an effort to identify non-citizen registrants. The technical capacity existed. The legal authority, it turned out, did not.

On June 22, 2026, Judge Sooknanan ruled that this use of the SAVE database was illegal. The ruling marked the second major judicial rebuke of the administration's non-citizen voter purge apparatus within a month — the first being the Democracy Docket-documented federal injunction against DHS's attempt to use citizenship databases to remove voters from state rolls. Two branches of the same enforcement strategy. Both struck down.

The New Hampshire precedent

The New Hampshire case illustrated what happens when the enforcement machinery operates ahead of the legal framework. Following early implementation of SAVE-related voter verification protocols, eligible American citizens found themselves unable to complete registration. The documentation requirements, combined with database mismatches, produced false positives — legitimate voters flagged as potentially ineligible. A federal court ruling on May 29, 2026 was required to restore their access.

This is the concrete mechanism by which voter suppression operates in the modern era: not through a single dramatic act of disenfranchisement, but through layers of administrative friction, database errors, documentation requirements, and deadline pressure that fall most heavily on the most vulnerable eligible voters. The phantom non-citizen voter is the justification. The suppressed citizen voter is the result.

Sixty-seven million records queried. The system ruled illegal. New Hampshire's eligible voters blocked. And at the center of all of it: 1.7 proven non-citizen votes per year. The disproportion is not a policy failure. It is the policy.

Collins, Murkowski, McConnell, Tillis

On June 4, 2026, the SAVE Act failed in the Senate: 48 in favor, 50 against. The four Republican senators who voted against it — Susan Collins of Maine, Lisa Murkowski of Alaska, Mitch McConnell of Kentucky, and Thom Tillis of North Carolina — come from different wings of the party, represent different constituencies, and operate under different political pressures. Their shared vote against this bill was not coordinated ideological dissent. It was four independent judgments reaching the same conclusion.

McConnell, in particular, is not a figure associated with protecting minority voting rights or questioning conservative electoral policy. His vote against the SAVE Act carries a weight that is difficult to dismiss. When the Senate Republican leader who built modern legislative obstruction into an art form votes against a voter fraud bill championed by his own party's president, the implicit message is unambiguous: the numbers don't work.

The 218-213 House vote tells a different story

In the House, the bill passed — but only barely. A margin of five votes in a chamber where the majority party controls the floor schedule, committee assignments, and procedural calendar is not a mandate. It is a managed outcome. The Republican leadership pushed the bill through because it could, not because the evidence demanded it.

The contrast between the two chambers is instructive: the Senate, with its longer terms, staggered elections, and historically higher threshold for floor votes, proved more resistant to legislation built on statistical fiction. Five Republican senators with more insulation from short-term political pressure made a different calculation — and two of them backed the Democratic position. The institutional design of American democracy, for once, did its job.

Five votes in the House. Fifty-to-forty-eight in the Senate. This is not a democratic mandate — it is a procedural squeeze play on the Constitution. The numbers in Congress were almost as thin as the numbers on non-citizen fraud.

The hidden cost of the cure

The Brennan Center identified 21 million American citizens who do not have ready access to a passport or a birth certificate — the two documents the SAVE Act would have required for voter registration. These are not hypothetical people. They are disproportionately elderly Americans who were born at home or in rural hospitals that no longer exist. They are low-income Americans who have never had reason or money to obtain a passport. They are Native Americans on reservations where document infrastructure has historically been inadequate.

Twenty-one million citizens. Against sixty-eight proven non-citizen voters in forty years. If the SAVE Act had become law and been enforced uniformly, it would have created — at minimum — a 308,823-to-1 ratio of harmed eligible citizens to prevented ineligible votes, based on Heritage Foundation data. This is the arithmetic of the bill. Its sponsors knew it. They chose not to discuss it.

The 69 million women facing name-change complications

The Brennan Center also flagged a distinct, massive population at risk: 69 million American women whose birth certificate name does not match their current legal identity due to marriage-related name changes. Under the SAVE Act's in-person documentation requirement, these women would face a bureaucratic labyrinth — potentially needing to produce both a birth certificate and a marriage certificate (or multiple marriage and divorce certificates) to prove identity continuity.

For a woman who has been married and divorced, then remarried, the chain of documentation required could span decades and multiple county courthouses. The Center for American Progress noted that 140 million Americans lack a valid passport — meaning for most citizens, the birth certificate would be the only viable document, and for 69 million women, that document alone would be insufficient. The practical barrier was not a side effect of the law. It was structural.

Twenty-one million citizens without the right documents. Sixty-nine million women whose documents don't match. One hundred forty million without a passport. And on the other side of the scale: sixty-eight proven fraud cases in four decades. I keep running the numbers hoping they'll change. They don't.

The Georgetown diagnosis

Professor Steve Vladeck of Georgetown University Law School — one of the country's most cited constitutional law scholars — articulated the problem in terms that resist misinterpretation: "The problem this law seeks to fix occurs so rarely that the proposed solution would be far more harmful than the problem itself." This is not a political statement. It is a legal and empirical assessment grounded in the same data that the Brennan Center, the Heritage Foundation, and the New York Times had independently produced.

The significance of Vladeck's framing is that it places the debate where it belongs: in the domain of proportionality. A law is not justified simply because it targets something real. It must target something real with a remedy proportionate to the harm. When the remedy harms 21 million citizens in order to prevent 1.7 verified infractions per year, the law fails the basic proportionality test. It fails it catastrophically.

The Democracy Docket ruling: judicial clarity

The Democracy Docket-documented federal court ruling of June 22, 2026 — blocking DHS from using citizenship databases to purge voter rolls — was not an ideologically motivated judicial act. It was the application of existing statutory and constitutional constraints to an executive overreach. The court found that the administration had extended the SAVE system beyond its authorized scope, deploying a citizenship verification tool designed for immigration benefits processing as a voter suppression instrument.

The ruling confirmed what legal scholars had been arguing since the SAVE database expansion was announced: that the administration was using the enforcement architecture of immigration law to operate inside the electoral system — without congressional authorization, without judicial oversight, and without the evidentiary basis that such an intrusion would require. The court said no. Clearly and on the record.

Georgetown says no. The federal judiciary says no. The Senate says no by fifty to forty-eight. The data has been saying no for forty years. The only thing still saying yes is a political machine that needs this crisis more than it needs the truth.

What the MPI data reveals

The Migration Policy Institute published a comprehensive analysis in February 2025 examining non-citizen voting in American elections. The institute's researchers identified a consistent pattern across state investigations: the overwhelming majority of non-citizens flagged as potential voters by automated database matching were either naturalized citizens whose records had not been updated, lawful permanent residents who had misunderstood their eligibility, or false positives generated by name-matching errors in state voter databases.

The MPI analysis underscored a technical reality that gets lost in political debates: database cross-referencing is not fraud detection. When a state runs its voter rolls against a DHS citizenship database, it generates matches based on partial data — name, date of birth, last four digits of a Social Security number. These matches produce errors. A naturalized citizen who registered before her citizenship status was updated in federal databases appears as a non-citizen. She is not committing fraud. She is being miscategorized by an imperfect system.

The Center for Election Innovation and Research

The Center for Election Innovation and Research, in an analysis published on February 10, 2026, reviewed the non-citizen voting claims that had circulated most widely in the months preceding the SAVE Act vote. Their finding: in every case where a specific claim of large-scale non-citizen voting had been made — and then investigated by election authorities — the actual confirmed cases were a fraction of the initial allegation, often by a factor of several hundred.

The pattern was consistent enough to constitute its own data point. Initial allegations tend to rely on raw database comparisons that do not account for naturalization, registration updates, or data entry errors. Subsequent investigations — when conducted rigorously — reduce the numbers by orders of magnitude. The gap between allegation and confirmation is not noise. It is the measure of the myth.

Every investigation starts with thousands. Every thorough investigation ends with dozens, or fewer. The pattern is so consistent it can no longer be attributed to coincidence. Someone is counting on the allegation. No one is waiting for the confirmation.

An attempt to restrict, not protect

The Brookings Institution published a direct assessment of the SAVE Act in March 2025. The conclusion, stripped of diplomatic hedging: the bill represented "an attempt to restrict voting rights" under the cover of fraud prevention. Brookings researchers noted that the bill's architecture — eliminating online and mail-in registration, requiring in-person documentary proof, imposing criminal liability on election workers — was structurally consistent with historical voter suppression strategies, regardless of its stated intent.

The Brookings analysis highlighted a crucial distinction that the bill's supporters consistently obscured: there is a difference between preventing non-citizen voting and making citizen voting harder. The SAVE Act accomplished almost nothing on the first dimension — non-citizens were already subject to severe criminal penalties, and the documented fraud rate was negligible. On the second dimension, it would have accomplished a great deal, concentrating its impact most heavily on the populations least likely to have the required documentation: the elderly, the poor, the rural, and communities of color.

The 2026 midterm context

The SAVE Act did not emerge in a political vacuum. It arrived in February 2026, with the midterm election cycle already underway and both parties calculating their positions on voter access. The bill's timing — introduced at the start of the registration season for 2026 — was not coincidental. Its implementation, had it passed, would have taken effect just as registration deadlines approached for an election in which House control was potentially at stake.

Brookings noted that the states where documentation barriers would be most acute — where populations with limited access to passports and birth certificates are most concentrated — also happened to be states where competitive House and Senate races were expected. The electoral logic was not hidden. It was embedded in the bill's geographic impact.

Brookings calls it an attempt to restrict voting rights. The data confirms it. The timing confirms it. The geographic impact confirms it. At what point does "voter fraud prevention" become something we're willing to name correctly?

218 votes: closer than comfort allows

The SAVE Act came within five House votes of becoming law. It was defeated in the Senate by two votes. That proximity is not a minor footnote — it is a warning about the structural vulnerability of American electoral democracy to legislation built on demonstrably false premises. The fact that four Republican senators had the integrity to vote against their party's president on this issue prevented a crisis. That four senators standing against a tide is now the margin of safety for 21 million citizens' voting rights is, itself, a crisis.

The SAVE database ruling on June 22 blocked the enforcement mechanism. The Senate vote on June 4 blocked the legislative vehicle. But the political infrastructure that produced the SAVE Act — the media ecosystem, the think tank network, the legislative drafters — remains intact. The myth of mass non-citizen voting did not die in the Senate chamber. It retreated to regroup.

The next version is already being written

The legislative history of voter access restrictions in the United States follows a consistent pattern: a bill is defeated, modified, rebranded, and reintroduced. The underlying argument shifts — from poll taxes to literacy tests to photo ID requirements to citizenship documentation — but the structural intent remains consistent: to make voting more difficult for populations that tend to vote against the party advancing the restrictions. The SAVE Act failed on June 4, 2026. The political conditions that generated it have not changed.

What has changed, slightly, is the record. The Senate floor debate, the judicial rulings, the expert testimony, the Brennan Center data, the Heritage Foundation's own numbers — they are now all part of a documented counter-narrative. The next version of the SAVE Act will have to contend with a more robust evidentiary record. It will not be stopped by that record. But it will be harder to pass in silence.

The bill failed by two votes in the Senate. The database was ruled illegal. And yet I find no comfort in these outcomes — because both were closer than they should have been. A democracy should not require this kind of last-minute rescue from a fiction this well-documented.

What forty years of data actually show

After four decades of investigation, the empirical record on non-citizen voting in American elections is remarkably consistent. The Heritage Foundation's 68 proven cases. The Brennan Center's 30 suspected incidents in 23.5 million votes. The New York Times' two possible cases in 137.7 million votes. The state-level investigations that reduce initial claims by factors of hundreds. The Migration Policy Institute's documentation of database mismatches. The Center for Election Innovation and Research's pattern of evaporating allegations. Every credible data source points in the same direction: non-citizen voting in the United States is vanishingly rare, and the existing criminal penalties already make it one of the most irrational acts an immigrant could commit.

Against that record stands the SAVE Act: a law that would have required 21 million citizens to produce documents they don't have, criminalized the honest mistakes of election workers, eliminated the registration methods used by tens of millions of Americans, and deployed a federal database against 67 million voter records — all to prevent something that happens, on average, fewer than two times per year. The numbers were always there. They were simply inconvenient.

The real question this leaves open

Professor Vladeck's formulation stays with you: "The problem this law seeks to fix occurs so rarely that the proposed solution would be far more harmful than the problem itself." The question that formulation doesn't answer — and that no data point can answer — is whether the architects of the SAVE Act ever actually believed the problem was real. Whether they were deceived by a debunked 2014 study, or whether they knew the numbers and built the law anyway. That distinction matters. A democracy misled by bad data can correct course when the data is corrected. A democracy misled by its own leaders requires something more fundamental.

By Maxime Marquette, columnist

Editorial stance

I am not a journalist but a columnist and analyst. My expertise lies in observing and analysing the geopolitical, economic and strategic dynamics that shape our world. My work is to take political strategies apart, to understand global economic movements, to put the decisions of international actors in context and to offer analytical perspectives on the changes that are redefining our societies.

I do not claim the cold objectivity of traditional journalism, which stops at reporting facts. I claim analytical clarity, rigorous interpretation and a deep understanding of the complex issues that concern us all. My role is to give meaning to the facts, to place them in their historical and strategic context, and to offer a critical reading of events.

Method and sources

This text keeps the basic distinction between verified facts and interpretive analysis. The rule never changes: a factual claim is published only when a verifiable source supports it, and the sources this article actually used are listed under Sources, never here.

Primary source families used by the publication, when the story has them: official statements from governments and international institutions, public statements by political leaders, reports from intergovernmental organisations, dispatches from recognised international news agencies.

Secondary source families: specialised publications, internationally recognised news media, analyses from established research institutions, reports from industry organisations.

When an article cites statistical, economic or geopolitical data, it comes from the institutions that produce it (intergovernmental organisations, central banks, national statistics offices), and the exact institution is named under Sources.

Nature of the analysis

The analyses, interpretations and perspectives in the analytical sections of this article are a critical, contextual synthesis based on the information available, the trends observed and the expert comments cited in the sources consulted.

My role is to interpret these facts, to set them in the context of today's geopolitical and economic dynamics, and to give them a coherent meaning within the larger story of the changes shaping our time. These analyses reflect expertise built through continuous observation of international affairs and an understanding of the strategic mechanisms that drive global actors.

This article describes a documented state of affairs on its publication date, not a prediction: later developments may shift its outlook. No update is promised in advance; when an article is corrected or completed, the change is dated in the text.