20,000 ‘Noncitizen’ BALLOTS Get Counted Anyway

People counting ballots at a table — Photo: Alexander Kalina / Shutterstock

The fight over alleged noncitizen ballots in California was never really about a tidy number; it was about timing, method, and the hard boundary between federal suspicion and state-run election machinery. When the Justice Department dropped its lawsuit after losing emergency relief, it didn’t resolve whether the underlying list was accurate—it confirmed that in American elections, process and proof govern remedies, especially when clocks and constitutional roles collide.

The Short Version

DOJ alleged more than 20,000 Californians “likely” to be noncitizens were registered and had been sent mail ballots, and sought to set those ballots aside pending verification.
Courts refused emergency relief during active voting; the Supreme Court declined to intervene, and DOJ voluntarily dismissed, ending the case without a merits ruling.
California officials challenged the list’s accuracy; a Los Angeles County sample review found many inactive registrations and indicators of U.S. birth, and most with no voting history.
The dispute illustrates a recurring dynamic: administrative data can flag possibilities, but mid-election remedies require precise evidence, clear authority, and lead time.

What DOJ Alleged—and the Remedy It Sought

The Justice Department filed a federal complaint asserting it had identified more than 20,000 people “likely” to be noncitizens on California’s voter rolls and that these individuals had been sent vote-by-mail ballots. The government said it transmitted the names to state officials and asked a court to order those ballots set aside until citizenship could be verified—an intervention tailored to preserve votes if lawful and exclude them if not. As emergency relief goes, that’s a concrete, reviewable ask: isolate a defined set of ballots and pause tabulation pending verification.

Two important caveats were baked into DOJ’s own framing. First, the list was probabilistic, not a set of adjudicated findings—“likely noncitizens,” not confirmed ineligible voters. Second, the allegation blended steps in the election pipeline: registration, ballot issuance, ballot return, and tabulation are distinct events. DOJ’s filings focused on who was registered and sent a ballot; they did not, in public materials, demonstrate how many of those ballots were actually returned or counted.

How the Courts Weighed Timing, Authority, and Evidence

Courts do not take mid-election surgery lightly. A district judge rejected the requested injunction, citing state control over election administration. A Ninth Circuit panel then declined to grant emergency relief, faulting the showing required for an extraordinary remedy during an active election—and noting that some of the timing problem was of DOJ’s own making. The Supreme Court denied the emergency application with a one-line order. None of these decisions decided the citizenship status of anyone on the list; they decided whether the federal government had shown enough, quickly enough, to justify segregating ballots in-flight.

Once ballots were mixed with others in county counts—“irreparably commingled,” in DOJ’s words—the practical remedy DOJ asked for evaporated. The government voluntarily dismissed the case, ending it before any court reached the core evidentiary question of who on the list was ineligible or whether any ineligible ballots had in fact been counted. Procedural defeat does not equal factual disproof; it does, however, tell you that emergency relief has a high bar and that late-arriving claims, however serious, struggle to clear it.

California’s Counter-Case: Accuracy and Usefulness of the List

California officials pushed back on both substance and usability. Secretary of State Shirley Weber argued the federal government had not provided evidence supporting its broad allegation and that the thumb-drive submission lacked the detail necessary for an actionable inquiry. Los Angeles County’s registrar reported a preliminary review of 200 names from the list: 112 were canceled or inactive (so no ballot mailed), nearly 40% of records with a reported birthplace indicated U.S. birth, and 71.2% had no voting history—figures that suggest a high false-positive rate if the public narrative presumes returned, countable ballots by noncitizens.

These data points are not a merits verdict on the full list; they are a snapshot that calls the list’s precision into question and underscores why courts hesitated to upend tabulation midstream. They also capture a chronic tension in election data work: administrative datasets flag anomalies, but they rarely prove disqualifying facts without person-level validation and clear chain-of-custody across disparate systems.

Mechanics Matter: From Database Match to Counted Vote

Four steps define the path from name to numerator: registration, issuance, return, and count. A person may be erroneously registered yet never receive a ballot (inactive or canceled status), may receive but never return one, or may return a ballot that is subsequently challenged and not counted. A bulk match across federal immigration or benefits databases and state voter rolls can raise plausible concerns—especially when status changes over time—but it inevitably carries false positives absent rigorous, current, person-level confirmation and well-understood error rates. That is why any large “possible noncitizen” headline must be decomposed into how many ballots, if any, entered the count.

Emergency election relief demands more than potential anomalies; it demands concrete risk to the count that is both imminent and identifiable, plus a remedy that is justiciable and feasible without disenfranchising eligible voters. That bar is highest when voting is already underway and the requested fix would alter routine canvassing procedures across a state.

The Justice Department on Friday voluntarily dismissed its California ballot lawsuit, ending its bid to set aside mail ballots from roughly 20,000 voters on its likely noncitizen list.

DOJ told the court the ballots can no longer be segregated before the Nov. 13 hearing. Judge…

— Frontier Signal (@FrontierSigHQ) November 7, 2026

What the Dismissal Does—and Doesn’t—Settle

The voluntary dismissal settled the immediate fight over ballot segregation; it did not settle the evidentiary dispute over the list’s accuracy or the number of ballots, if any, that should have been excluded. If anything, the case demonstrates why front-loaded, cooperative validation beats back-loaded litigation. If the federal government believes it has reliable indicators, the right forum is months before ballots go out—through structured data exchanges, audit protocols, and privacy-protective sampling that quantify false positives and separate registration housekeeping from ballot-count risk.

Lessons for Future Integrity Disputes

Three durable lessons emerge. First, precision beats scale. A smaller, validated set with documented methodology and error rates will always travel farther in court than a big round number with opaque matching rules. Second, calendar is destiny. Remedies that are sensible on paper—temporary segregation, targeted verification—become impossible once ballots are mixed. Third, state-federal roles are not a technicality. States run elections; federal actors seeking mid-cycle intervention must show clear statutory authority and a remedy that respects that structure.

What a Credible Path Forward Would Look Like

A serious approach would include: a defined, pre-election validation window; documented matching criteria and quality controls; independent auditing of false-positive rates; privacy-screened, person-level sampling against birth, naturalization, and immigration records; and clear reporting that disaggregates registration anomalies from returned and counted ballots. If agencies want courts to act, they must also preserve a merits forum by filing early enough that a neutral fact finder—special master or otherwise—can test the evidence before ballots are commingled.