The real conflict in California is not whether noncitizens ever get on the voter rolls—they sometimes do—but whether the federal government proved a scale and urgency that justifies extraordinary intervention into a state’s ballot handling before votes are counted.
The Short Version
- Federal prosecutors sued California after identifying more than 20,000 “likely noncitizens” on the voter rolls and sought to set aside associated ballots pending verification.
- Alongside the lawsuit, DOJ charged 28 lawful permanent residents with allegedly registering and, in many cases, voting unlawfully; 11 were arrested in a coordinated sweep.
- California officials rejected the federal list and emergency remedy, arguing the data and methodology were insufficient and that ballot sequestration would disenfranchise lawful voters.
- Evidence shows noncitizen voting exists but at very low observed rates, and large-scale identification is technically difficult; the legal question of pre-count intervention is distinct from the empirical one.
What DOJ actually alleged—and why that matters
The Justice Department said it filed a civil complaint asserting that more than 20,000 people “likely” to be noncitizens are registered to vote in California, developed in concert with the Department of Homeland Security and delivered to state officials with a demand: set aside ballots linked to those registrants until citizenship can be verified. In parallel, prosecutors announced 28 criminal cases—mostly lawful permanent residents—alleging illegal registration and, for many defendants, participation in federal elections; 11 were arrested in an operation spanning all four California-based U.S. Attorney’s Offices. These are concrete, named cases. They also form the crux of DOJ’s argument that its larger list is not hypothetical.
Precision of claim is crucial. The 20,000 figure describes “likely noncitizens” on the rolls, not proven unlawful voters; registration is not the same as a counted ballot. The criminal filings are more specific—individual charges backed by documents and alleged conduct—but 28 cases do not numerically validate a 20,000-person statewide estimate. That gap goes to remedy. DOJ asked a court to segregate ballots associated with those registrants now, then verify later.
California’s counter: data sufficiency, legal authority, and voter protection
California’s election leadership did not merely express skepticism; they directly challenged the list’s utility and DOJ’s proposed remedy. Secretary of State Shirley Weber said DOJ had not provided evidence for its 20,000-person allegation and that the methodology was unexplained, unverifiable, and inadequate for investigative use. The Secretary of State’s office characterized the demand to seize and sequester ballots as lacking concrete evidence and incompatible with state and federal law governing election administration.
State officials additionally argued that voter rolls alone do not establish citizenship status and that California already enforces eligibility rules. Attorney General Rob Bonta framed the announced prosecutions as isolated misconduct in a state with roughly 23 million registered voters, not evidence of systemic fraud; the cases may be real, he suggested, but statistically they are “a drop in the bucket” relative to statewide participation. Senator Alex Padilla added a jurisdictional critique, asserting DOJ had neither proven widespread fraud nor authority to direct California’s election machinery.
Mechanics: why “likely noncitizen” lists are hard to get right
Large-scale citizenship screening is a collision of database uncertainty and election law. Government and commercial files vary in freshness, structure, and purpose; immigration status changes over time; names collide; and identifiers are incomplete. Confidence thresholds that are tolerable in investigative triage can be too noisy to justify ballot-level remedies. Without disclosure of match fields, deduplication, and validation error rates, a 20,000-person “likely” list can be either a critical early-warning or a broad net that catches naturalized citizens and stale records—sometimes both.
Empirically, well-run post-election audits and prosecutor referrals over the past decade typically find vanishingly few noncitizen ballots relative to totals. One survey of jurisdictions covering 23.5 million votes in 2016 identified roughly 30 referred incidents of suspected noncitizen voting—on the order of one in a million ballots. That base rate does not negate individual violations, and it does not guarantee perfect detection; it does explain why courts are cautious when the proposed remedy risks sidelining lawful votes at scale before facts are individualized.
The legal fork: criminal enforcement versus pre-count ballot intervention
Two legal tracks traveled together here. The criminal cases are conventional: identify an ineligible registrant, reconstruct the registration pathway and any ballots cast, and prosecute if the facts and intent support charges. If sustained, they vindicate the proposition that noncitizen voting happens, albeit infrequently. The civil suit is different. DOJ’s emergency motion sought to obligate California to set aside ballots tethered to a mass list without prior voter-by-voter adjudication. That asks a court to balance the risk of counting unlawful ballots against the countervailing risk of chilling or delaying lawful votes—while squaring state prerogatives in election administration with federal preemption claims anchored in federal election crimes.
Court skepticism in similar disputes tends to turn on tailoring: is the remedy no more intrusive than necessary, grounded in verifiable, individualized evidence, and consistent with statutory limits on list maintenance and ballot secrecy? California’s resistance focused Squarely on those points—methodological opacity and disenfranchisement risk—rather than denying that any violations can occur.
What the numbers do—and do not—establish
The announced 28 cases are the strongest, most specific pieces of the federal narrative; they name defendants, allege false citizenship attestations, and, in most instances, identify ballots cast. They do not, by themselves, scale to 20,000 unlawful ballots. Nor does the 20,000 “likely” registrant figure, absent public validation of match methodology and error rate, prove that all or most of those individuals are currently noncitizens or that they voted. This asymmetry does not trivialize the cases; it clarifies what each datapoint can legitimately carry.
On the other side, California’s statistical framing—that 28 alleged cases among 23 million registrants is de minimis—properly guards against hyperbole. It does not dissolve the state’s obligation to investigate credible leads or correct erroneous registrations. Both can be true: the base rate is very low, and the law still forbids even a single unlawful ballot.
A constructive path: transparency, adjudication, and narrowly tailored safeguards
The way to resolve disputes like this without corroding public confidence is procedural clarity, not rhetorical escalation. Three steps would help. First, disclose the list-building protocol under protective order: fields used, source systems, confidence thresholds, and validation results, including measured false positives and deduplication. Second, run a controlled, joint audit—DOJ, DHS, California, and county officials—with due-process safeguards for naturalized citizens and a clear appeal channel. Third, tie any interim ballot handling to individualized determinations: only segregate when a verified, current noncitizen status is established for a specific registrant and when records indicate a ballot has actually been returned for counting.
Criminal cases should proceed on their evidentiary merits; civil remedies should be calibrated to demonstrated facts, not estimates. That is how you defend eligibility rules and voter confidence at the same time. Done well, the result is not a compromise with illegality; it is a disciplined separation of genuine violations from database noise—ensuring unlawful ballots are not counted, and lawful ballots are never sidelined because of an opaque algorithm.
Sources:
nypost.com, justice.gov, abc10.com, abc7.com, latimes.com
