DHS Chief Threatens Prison Over Voter Data

When a federal homeland security chief threatens state election officials with prison if they defy Washington’s data demands, you are no longer talking about routine voter-roll maintenance—you are in the middle of a live test of the boundary between federal power and the states’ constitutional control over elections.

Key Points

DHS Secretary Markwayne Mullin has publicly warned state and local election officials that refusing to cooperate with federal election security directives could expose them to fines, penalties, and even prison time.
The threats are anchored in a disputed DHS claim that roughly 250,000 noncitizens and 400,000 deceased individuals are on voter rolls in four Democratic-led states, figures derived from public data with no disclosed methodology.
State election officials and independent experts say noncitizen voting is vanishingly rare, challenge DHS’s numbers as speculative, and note that federal databases used for these sweeps have already been blocked by a judge as unreliable.
The episode fits a recurring pattern: expansive federal rhetoric about large-scale illegal voting followed by court setbacks, methodological debunking, and no evidence that noncitizen voting has altered any election outcome.

Mullin’s Escalation: From Election Integrity to Criminal Exposure

Markwayne Mullin’s recent remarks did more than repeat President Trump’s familiar narrative about insecure elections; they introduced a direct coercive threat aimed at the officials who actually run those elections. In a series of briefings and interviews, Mullin said that if election officials refuse to cooperate with DHS’s directives—sharing voter data, enrolling in federal screening tools, and running specified “scrubs” of their rolls—“those individuals can also be held accountable by fines, by penalties, and even depending on how far it goes, prison time.” These statements build on earlier Justice Department letters warning officials they could face prosecution if they knowingly leave noncitizens on their rolls.

Legally, Mullin is gesturing toward existing federal statutes: it is already a crime to knowingly facilitate illegal voting in federal elections, and illegal voter registration or casting a ballot while ineligible can carry penalties up to five years in prison and $250,000 in fines. The novelty is not the law; it is the posture. A homeland security secretary is publicly framing ordinary disagreements over data sharing and methodology as potential criminal conduct by state officials. That is a sharp departure from the usual cooperative, technical tone of federal–state work on election security.

The Numbers at the Center of the Clash

Mullin’s threats are tethered to a set of headline figures that have become the fulcrum of the dispute. DHS asserts that it has “identified over 250,000 potential non-citizens illegally registered to vote in just four U.S. states,” all led by Democratic governors: California, New Jersey, Nevada, and Pennsylvania. A related DHS report and Mullin’s public remarks also cite roughly 400,000 deceased individuals remaining on voter rolls in those states.

More detailed breakdowns, reported in sympathetic outlets, assign about 190,832 of those supposed noncitizen registrations to California, 35,152 to New Jersey, 15,903 to Nevada, and 14,576 to Pennsylvania. Yet when Mullin wrote to Pennsylvania officials, he described the 14,576 figure as an “initial estimate” based on an analysis of publicly available voter registration data rather than any verified citizenship records. Neither he nor the department has disclosed the underlying methodology, the specific databases used, or the rate at which past attempts with similar data have misidentified lawful voters as noncitizens.

This lack of methodological transparency is not a minor technical quibble. Public voter lists are intentionally scrubbed of key identifiers—such as driver’s license numbers and full Social Security numbers—to protect privacy. Analysts have repeatedly shown that when you try to infer citizenship status from these truncated records, especially by matching against external databases that are incomplete or out of date, the resulting estimates are highly inflated. It is precisely this pattern that leads independent experts to treat large top-line numbers about “potential” noncitizen registrations as suspicion flags rather than reliable fact.

State Officials Push Back: “Wildly Speculative” Claims

Officials in the states Mullin targeted have not simply shrugged off the federal claims; they have actively challenged both the numbers and the premise of the crackdown. California’s secretary of state has emphasized that noncitizen voting there remains “exceedingly rare” and that voter rolls are regularly maintained according to state law. Nevada’s secretary of state, Francisco Aguilar, was more blunt, calling the DHS figures “wildly speculative at best” and noting that the department “hasn’t shared anything that backs it up.”

Pennsylvania’s election authorities have taken a similar line: they acknowledge receiving Mullin’s letter and say they will review the assertions, but they also stress that they already conduct list-maintenance procedures and have seen no evidence of large-scale noncitizen voting. Rhode Island’s secretary of state, though not among the four named states, refused outright to share voters’ private data with the federal government despite threats of lost funding, signaling that resistance is not confined to the jurisdictions Mullin singled out.

This kind of pushback matters for two reasons. First, it highlights the constitutional reality that states, not DHS, administer elections and control voter registration data, subject to federal law but not federal micromanagement. Second, it underscores that the officials closest to the system—the ones who see actual ballots and registration forms—are not witnessing the kind of illegal voting crisis Mullin’s numbers imply.

What the Broader Evidence Shows About Noncitizen Voting

To understand the stakes, you have to zoom out from one day’s press conference to decades of empirical work on noncitizen voting. Across multiple studies, covering tens of millions of ballots in dozens of jurisdictions, researchers have found that noncitizen participation in federal and state elections is, in the words of one comprehensive review, “vanishingly rare.” A Brennan Center analysis of 42 jurisdictions overseeing 23.5 million votes in the 2016 general election identified only about 30 suspected incidents of noncitizen voting referred for further investigation. That is roughly 0.0001% of votes cast, and many of those cases were never confirmed as intentional fraud.

Similar findings recur in other independent examinations. The Bipartisan Policy Center, after reviewing available data and state investigations, concluded that “there is no evidence that noncitizen voting has ever been significant enough to impact an election’s outcome.” The American Immigration Council’s analysis of the Heritage Foundation’s own fraud database—often cited by advocates of stricter rules—found extraordinarily few substantiated noncitizen cases compared with the hundreds of millions of ballots cast. A nationwide review published in 2025 again reinforced the same point: noncitizen voting happens occasionally but at “minuscule” levels and not in any coordinated fashion.

Taken together, these studies do not claim perfection; some noncitizen voting does occur, and any illegal voting is serious for the individuals involved. But the pattern is unmistakable. Whenever states have conducted detailed investigations—comparing voter rolls against up-to-date citizenship records, reviewing individual cases, and sometimes prosecuting offenders—the result has been a tiny handful of cases, not the hundreds of thousands cited by political actors. That empirical backdrop sharply undercuts narratives of systemic illegal voting.

Unreliable Databases and Judicial Roadblocks

The reliability of the tools DHS wants states to use is a central point of contention. Mullin has urged states to adopt a federal system to identify noncitizen voters and suggested that those who decline will become enforcement “priorities.” But federal databases commonly used for such matching—such as SAVE (Systematic Alien Verification for Entitlements)—were designed for benefit eligibility, not for high-stakes voter-roll sweeps, and they are known to contain outdated or incomplete information about individuals’ citizenship status.

That design mismatch has real consequences. Because these systems often lag behind naturalization and can misclassify long-time citizens, their use for voter purges produces frequent false positives: people flagged as noncitizens who are in fact eligible voters. PBS NewsHour’s review of recent litigation reports that a federal judge blocked the Trump administration from using these records for its noncitizen voter initiative, specifically because of privacy concerns and the high rate of misidentification. Courts have also repeatedly rejected efforts by federal officials to compel states to hand over expansive voter data, leading to multiple lawsuits the administration has lost over access to detailed rolls.

These setbacks do more than slow federal plans; they erode the legal foundation of Mullin’s threats. If the databases underlying DHS’s claims are judicially recognized as unreliable for identifying noncitizen voters, then punishing state officials for refusing to rely on them is difficult to square with rule-of-law commitments. The judiciary is, in effect, signaling that caution and skepticism about these tools is not obstruction—it is prudence.

Foreign Interference Rhetoric and the National Security Frame

Mullin and Trump have consistently framed election security not just as a procedural matter but as a frontline national security concern. Their speeches reference alleged Iranian cyber activities against U.S. voter systems and broader foreign efforts to access voter data and compromise elections. In Trump’s earlier prime-time address, he cited declassified documents to claim that China had illicitly acquired 220 million U.S. voter files and that Venezuela’s regime could tamper with voting machines—a narrative meant to justify sweeping federal interventions.

The underlying intelligence materials, however, don’t support those dramatic claims. Detailed reporting shows that China obtained publicly available voter registration information from a handful of states—data that campaigns, journalists, and private firms can legally purchase—rather than hacking into systems or altering votes. The Venezuela-related intelligence, likewise, described capabilities relevant to Venezuelan elections, not proven or even plausible mechanisms for altering U.S. vote tallies. To date, Trump’s own allies have conceded that there is “zero evidence that a foreign power flipped a vote” in recent federal elections.

For election professionals, the concern is not that foreign actors are harmless; it is that inflated, poorly grounded rhetoric about foreign interference is being used to justify domestic policy moves that centralize power in Washington, weaken state autonomy, and cast doubt on the legitimacy of future results. The empirical track record—no documented vote flipping, rare noncitizen voting, extensive use of paper ballots and chain-of-custody procedures—points toward resilience, not imminent collapse.

Where the Real Line of Disagreement Lies

Both sides of this dispute claim to care about secure elections. The genuine disagreement is about scale, method, and power. The Trump administration, through Mullin, asserts a large-scale registration problem based on opaque analysis of public data and moves quickly to threaten funding cuts and criminal exposure for officials who balk at federal demands. States and independent experts counter that the numbers are methodologically unsound, that noncitizen voting is empirically rare, and that using unreliable federal databases to purge rolls risks disenfranchising lawful voters.

On the evidence available, the weight lies clearly with the skeptics. There is no published DHS methodology for the 250,000 figure, no list of names, no demonstration of how many of those registrants ever cast a ballot, and no clear legal path for compelling states to redesign their election procedures on the basis of those estimates. There is, by contrast, a substantial body of empirical work, judicial decisions, and practical experience from state officials pointing to tiny levels of noncitizen voting and cautioning against overbroad purges.

For citizens who care about both ballot integrity and constitutional structure, the lesson is straightforward. Guarding elections against fraud—including rare noncitizen cases—is necessary, but it must rest on accurate data, transparent methodology, and respect for the states’ primary role in running elections. When national security rhetoric and splashy numbers are deployed without that foundation, and backed by threats of prison for local officials, the risk is not that fraud will go unaddressed; the risk is that trust in the system will be eroded by the very institutions claiming to protect it.