It is no secret that Donald Trump is concerned about noncitizens on the voter rolls.  It was evident in 2016, when he claimed that the government was letting immigrants “pour into the country” so that they could vote.  That pattern continued in his second term, as he and his Administration have treated noncitizen voting like a dog with a bone. 

One consequence has been to magnify this small-but-real administrative problem beyond all proportion and to make politically workable solutions harder to reach.  As tempting as it is to speculate about how the Administration’s fixation on noncitizen voting will affect the 2026 election outcome, we also need to consider what a Democratic Congress could do after November 3.

The available evidence points to relatively few noncitizens on the voter rolls and almost none voting.  To be clear, “relatively few” and “almost none” are not “zero.”  But even a homeopathic quantity of noncitizens on the rolls is a problem for at least four reasons.  First, it is illegal for noncitizens to register, much less vote, so none should be on the list.  Second, evidence of even a small number of noncitizens is ammunition for the 28% of registered voters (according to the 2024 Survey of the Performance of American Elections) who believe noncitizen voting is “very common” or occurs “occasionally.”  Third, any reluctance election officials appear to have in identifying noncitizens on the rolls—whether inadvertent or not—makes them look ineffective.  Finally, a noncitizen who registered, on purpose or not, may face serious immigration consequences.

I talk to election officials and attend their meetings.  Wherever they come from, I have yet to encounter one who does not want better tools to prevent noncitizens from getting on the rolls or to identify them when they do.

The profession is well aware of the imperfections of the existing tools used to remove voters because they have moved away or died:  National Change of Address (NCOA), motor-vehicle records, death files, and the like.  The 1993 National Voter Registration Act (NVRA) constrains list maintenance by requiring uniform, nondiscriminatory programs and prohibiting systematic removals within 90 days of a federal election.  For removals based on a change of residence, the NVRA also requires notice and a waiting period.

In this context, the Administration's decision to facilitate the use of the Systematic Alien Verification for Entitlements (SAVE) program, and to enhance it by allowing bulk matching and combining it with other federal records to better verify citizenship, can provide election officials with another imperfect but potentially valuable data source.  The evidence made public over the past year, however, shows that SAVE must be used at least as carefully as other list-maintenance data.

Utah provides one example of a serious use of government data to confirm voter citizenship. A Bipartisan Policy Center report describes a recent “citizenship verification audit” by the Utah lieutenant governor’s office.  That audit reported that data from the Utah Driver License Division (DLD) could not confirm the citizenship of 4% of the state’s voter records.  When Utah used SAVE to review 71,256 records the DLD couldn’t confirm, SAVE couldn’t confirm the citizenship of 8,917 of those voters.  After a further manual review of these remaining registered voters, Utah found 27 confirmed noncitizens and 25 probable ones.  The review began in April 2025 and was not completed until May 2026.

One sentence in the Utah report stands out:  “While we found information obtained from the DLD or SAVE to be reliable in confirming whether an individual is a citizen, it was not reliable in indicating an individual was not a citizen, without considering other factors.”

The Utah case provides context for the Department of Justice’s (DOJ) October 13 letter, which gave California 48 hours to investigate a list of 20,000 people whom the federal government described as likely noncitizens and who had been sent mail ballots.  The list went to Los Angeles County, even though it reportedly included voters from across the state.  The state said that neither the Department of Homeland Security nor DOJ had provided the underlying data or methodology for their allegations; DOJ sued when California did not comply and LA County has responded.

Others have weighed in on the legal deficiencies of the DOJ’s position.  For present purposes, the more important point is that the Administration is using federal citizenship-matching information not as a tool for regular state list maintenance, but to manufacture a late-breaking confrontation over voter rolls while voting is underway.

What is to be done?  It is easy to see the October 13 letter and similar actions as the first shot in what could become countless salvos against the election process, fueled by evidence from federal and commercial sources.  Because I am aligned with those election experts who are confident that regular order will prevail and the rightful winners will be seated in office in 2026, I believe it is important to look beyond Election Day and the inevitable challenges to the outcomes and consider what the 120th Congress should do. If Democrats control one or both chambers of the 120th Congress, they should consider how to minimize the damage caused by DOJ’s use of federal citizenship matching without prohibiting it altogether.

The obvious answer is to use the power of the purse to keep DOJ from turning federal citizenship matches into demands that state and local officials change their voter rolls.  If enacted, such a rider would not resolve every underlying legal dispute, but it would curb this form of federal pressure on state and local election officials.

One advantage of prohibiting DOJ from using the enhanced SAVE program for further action on the noncitizen voting front is that it would leave states free to use SAVE voluntarily in their own list-maintenance programs, subject to the NVRA and other applicable law.  It would also allow evidence to accumulate about how SAVE works in practice and how states respond to inconclusive matches.  That approach would not satisfy everyone.  But it would restore the proper balance:  states could keep their rolls accurate without DOJ turning citizen verification into a federal political weapon.