Census Bureau Publishes Unsigned August 18 Brief Claiming 24000 Noncitizen Voters

The brief combined commercial voter files with federal administrative records from Social Security and immigration databases. No authors, acknowledgments, or contact points appeared, violating every prior Census research output format. Former chief scientist John Abowd documented that earlier linkage tests produced false-match rates high enough to explain most of the flagged records as data errors rather than verified noncitizen votes.

Record-linkage false-positive rates documented in Census methodology papers exceed the 24000 figure once scaled to the full voter file. Amy O’Hara, former administrative-data research lead, stated the report lacks methodological rigor and accountability. The separate privacy-method decision further altered differential-privacy parameters for block-level demographic releases used in redistricting.

These interventions concentrate damage on voting and local population estimates while sparing headline macroeconomic series. Political authorship presented under the Census imprimatur collapses the separation between statistical production and policy advocacy that has protected data integrity across administrations.

States and localities will receive 2020 census-derived files carrying altered privacy protections and contested citizenship signals. External audits of linkage pipelines and renewed congressional oversight of statistical-agency leadership vacancies become probable within the next fiscal year.