Power Grab/Fire anyone who can stop him
The Census Bureau publishes a brief claiming over 24,000 noncitizens voted in 2020, and Trump says it “proves” it
The unauthored August 18, 2026 brief matched a commercial voter file to federal records; outside experts and former bureau officials say its method lacks the identifiers and error rates to support the claim.
What happened
On August 18, 2026, the Census Bureau released an unauthored brief saying it had linked over 24,000 voter records from the 2020 election to noncitizen status. Titled “Noncitizen Voting in the 2020 Election, A Beginning Analysis,” it matched a commercial voter file against federal administrative records.
The brief said over 128 million of nearly 160 million voter records were citizens with high confidence, and left over 32 million records to analyze. President Trump posted that the brief “proves” more than 24,000 noncitizens voted.
David Becker of the Center for Election Innovation and Research said the analysis “fails the laugh test” because Census data lack identifiers to match voter files. Former bureau officials told NPR it lacked error rates and a discussion of bias. NPR reported the team had ties to the America First Policy Institute.
Who benefited
Trump gained a federal statistical agency’s name on a claim he repeated as proof that over 24,000 noncitizens voted. He posted it as proof. The brief carries no author, no error rates and no bias discussion, which makes it hard to check and easy to quote.
Who wanted something
The team behind the brief had ties to the America First Policy Institute, NPR reported. The Census Bureau published it as “A Beginning Analysis,” with over 32 million records still unexamined. The claim of over 24,000 noncitizen voters was out before the analysis of those records was finished.
Who lost out
Voters, and the Census Bureau’s standing as a source of neutral numbers, lost out. Over 128 million of nearly 160 million records were judged citizens; over 32 million remain unanalyzed. Former officials said the brief had no error rates, so no one can say how many of the 24,000 linked records are wrong.
Why it matters
A federal agency published a count of noncitizen voters without an author, error rates or a bias discussion, and the President cited it as proof. Experts say the data cannot support the match. The brief gave the effort to portray noncitizen voting as a crisis an official-looking figure.
Sources U.S. Census Bureau (primary)NPRCenter for Election Innovation & Research (primary)AP