On June 4, 2026, the U.S. Secretary of Commerce issued Directive DAO‑216‑26, which replaces existing confidentiality safeguards for publications from the Bureau of Economic Analysis and the U.S. Census Bureau with methods dating from the early 1970s and limits the use of modern privacy‑enhancing techniques.

Differential privacy, a statistical method that adds random noise to protect individual identities while preserving data utility, has been used by the Census Bureau since 2008 for commuting‑pattern data and was planned for the 2030 Census. The Census Act (13 U.S.C. § 9) requires that released data not allow identification of individuals.

The directive bans differential privacy and other noise‑infusion techniques, permitting only coarsening (aggregation, rounding, range use) and suppression as a last resort. It also restricts input noise infusion, swapping, and differential privacy, affecting products such as the Quarterly Workforce Indicators and planned BEA statistics.

The directive conflicts with the Census Bureau’s statutory duty to protect confidentiality while providing usable statistics. A case study based on County Business Patterns data shows that coarsened categories can be combined to reconstruct exact employee counts for individual businesses, a risk that noise infusion would mitigate.

The order bypassed required administrative procedures and mandates a return to traditional statistical methods, which reduces the level of detail available in published data.

Civil servants may comply by reducing data granularity or releasing information that can be re‑identified, potentially eroding public trust and response rates for the census.

The scientific community and statisticians have characterized the directive as driven by political considerations rather than empirical evidence and have called for increased investment in statistical agencies and for privacy methods to be selected by technical experts. Individuals are encouraged to contact their congressional representatives to express concerns and to support preservation of methodological documentation.