The Census Bureau proposed counting only U.S. citizens and lawful permanent residents for congressional apportionment in 2030, a departure from the 2020 census that would exclude undocumented immigrants and many people living lawfully in the country on temporary visas.

The 23-page proposed rule, filed for public inspection Wednesday and scheduled for publication Thursday, would redefine “usual residence” for the once-a-decade count. The bureau says people without citizenship or permanent-resident status lack the durable allegiance needed to be considered residents for apportionment, even when they live and work in the United States. The proposal is not final and opens a 30-day public-comment period.

The same rulemaking would prohibit questions about race, ethnicity or sexual orientation on the short-form decennial census and other questionnaires used to enumerate the population. Questions about biological sex, date of birth and household relationships could remain. The restriction would not apply to the American Community Survey or prevent the bureau from obtaining demographic information from administrative records.

Together, the changes could reshape the population base used to divide 435 House seats among the states, draw congressional and legislative districts, enforce civil-rights laws and distribute federal funds. They would also place two legally distinct issues—who counts for apportionment and which demographic data appear on the census form—into one consequential administrative proceeding.

The proposal narrows who has a usual residence

For the 2020 census, foreign citizens living in the United States were counted at the address where they lived and slept most of the time, whether they were present legally or illegally. Foreign visitors on vacation or short business trips were excluded. The new proposal would replace that physical-presence standard with one tied to legal status: citizens and lawful permanent residents would be included, while undocumented immigrants and people whose status is less durable than permanent residency would be omitted from the apportionment base.

That broader exclusion would reach temporary workers, international students and others who are lawfully present but do not hold green cards, depending on the final rule’s definitions and implementation. The bureau invited comments on whether any other immigration status is sufficiently similar to permanent residency to qualify. It also said it could use records from the Treasury, Homeland Security, Health and Human Services, State Department, Postal Service, state governments and commercial providers to determine legal status.

The agency is separately considering whether to add a legal-status question to the short census form. The proposal does not yet mandate that question. Instead, it leaves open a choice between asking households directly, using administrative data, or combining both methods. That distinction matters because a legal-status question could affect response behavior, while administrative matching creates separate questions about coverage, accuracy and the treatment of conflicting records.

House seats and federal resources are at stake

The decennial census does more than report population. The official apportionment results determine how the 435 voting seats in the House are divided among the 50 states. The 2020 apportionment population was 331,108,434, including state residents and certain overseas federal employees and their dependents assigned to a home state.

Removing large groups from the 2030 base could change which states gain or lose seats because apportionment turns on relative population, not merely a national total. States with larger shares of noncitizens could lose population for the formula even though those residents use roads, schools, hospitals and other public services. The effects cannot be calculated responsibly from the proposed rule alone because the bureau has not published a final matching method, error rate or state-level exclusion estimate.

The count also supplies the population denominators used across government and the private sector. Demographers and civil-rights organizations told Reuters that excluding noncitizens and removing race and ethnicity questions could distort redistricting, anti-discrimination enforcement and funding decisions. That is their assessment of likely consequences, not a measured outcome. The administration argues that a shorter, less intrusive questionnaire would improve response rates and protect the core population count.

A constitutional dispute is built into the rule

Section 2 of the 14th Amendment directs that representatives be apportioned by “counting the whole number of persons in each State,” according to the constitutional text. The proposed rule argues that “persons in each State” historically meant inhabitants with an enduring tie or allegiance, not every person physically present. It relies in part on Supreme Court decisions recognizing broad federal discretion to define usual residence for groups such as military and civilian personnel stationed overseas.

Opponents are expected to argue that the distinction between “persons” and “citizens” is deliberate and that a person who normally lives in a state is part of its population regardless of immigration status. Earlier litigation has not produced a definitive Supreme Court ruling on the exact policy now proposed. In 2020, the justices dismissed a challenge to President Donald Trump’s effort to exclude undocumented immigrants because the case was premature, leaving the central constitutional question unresolved.

A separate 2019 decision addressed the administration’s attempt to add a citizenship question to the 2020 form. The Supreme Court agreed that the government generally may ask about citizenship, but it held that the Commerce Department’s stated rationale did not match the administrative record. The official opinion therefore makes clear that census choices can fall within federal authority and still fail review if the agency does not provide a reasoned explanation supported by evidence.

The new proposal anticipates litigation by stating that its residence criteria and demographic-question restrictions are independent and severable. If a court blocks one part, the bureau intends the rest to remain in effect. That drafting cannot prevent a lawsuit, but it could limit how much of the rule falls if challengers prevail on only one issue.

Race and ethnicity data would move off the short form

The rule would bar race and ethnicity questions from the decennial enumeration even though demographic questions have appeared in every census since 1790. The bureau says those questions can increase burden, raise privacy concerns and complicate disclosure-avoidance methods that protect individual responses. It argues that removing them from the short form could improve participation and reduce the uncertainty introduced when detailed data are protected against identification.

The bureau would continue collecting demographic information through the American Community Survey, which samples households every year, and through government records. That means the proposal would not eliminate federal race and ethnicity statistics. It would change their source, frequency and geographic precision, and it could separate the official apportionment count from some of the demographic detail historically released alongside it.

Whether administrative data can substitute for direct responses is an empirical question. Government records may cover some people well while missing others, and linking files across agencies can propagate errors when names, addresses or legal statuses change. The bureau’s proposal says its record holdings are extensive and growing, but it does not establish that every person excluded or classified under the new system can be identified without consequential error.

What happens before any change takes effect

The Federal Register recorded the proposal as filed at 8:45 a.m. Wednesday, with formal publication scheduled for September 10. Comments will be accepted for 30 days under docket USBC-2026-0628. The bureau must then review significant comments, decide whether to revise the text and publish a final rule before the policy can govern the 2030 count.

Legal challenges are highly likely given pending lawsuits over whether noncitizens must be included. An AP report in February described litigation brought by Republican-led states seeking to exclude undocumented immigrants, while immigrant-rights groups argued that the Constitution and census law require counting usual residents. Those cases and any challenge to this rule could determine how much discretion the Commerce Department has before preparations for 2030 become difficult to change.

For now, no resident has been removed from a census count and no state has lost a House seat or federal dollars because of this proposal. What changed Wednesday is the administration’s position moving from a presidential directive and courtroom argument into a detailed regulatory plan. Its final reach will depend on public comments, the evidence the bureau develops, the wording of any final rule and almost certainly federal judicial review.