How would excluding undocumented immigrants from the census affect congressional seats and federal funding?
Executive summary
Excluding undocumented immigrants from the census would shrink counts in states and localities with large unauthorized populations and could shift seats in the U.S. House and the Electoral College, because apportionment and many federal formulas rely on total population; Pew estimated 14 million unauthorized immigrants in 2023, a concentration that grew in 2021–23 and varies by state [1]. Researchers and policy groups disagree about recent mobility and undercount magnitudes; some analyses show large recent declines in the foreign‑born population in 2025 while others emphasize methodological limits in measuring undocumented residents [2] [3].
1. Why the census matters: seats, dollars and political power
The decennial census supplies the population totals used to apportion 435 House seats among states and allocates Electoral College votes and many federal grants on population-based formulas. If a category of people is removed from the count, states with sizable undocumented populations would see their total headcounts fall relative to other states, potentially costing House seats and influence in federal funding allocation; Pew’s 2023 estimate of 14 million unauthorized immigrants shows the scale at stake [1]. Available sources do not provide a precise seat-by-seat apportionment projection based on removing undocumented residents, but the mechanism is direct: smaller totals change apportionment math and funding formulas tied to population [1].
2. Who would be most affected: states and localities with big recent growth
Undocumented immigrants are not evenly distributed; recent inflows disproportionately raised population counts in specific states and localities, meaning those places would bear most of the numeric loss if undocumented residents were excluded [1]. Multiple research efforts show a surge in arrivals in 2021–23 and indicate that the foreign‑born population’s growth has been concentrated and uneven—so the political and fiscal effects would be geographically concentrated rather than spread uniformly across the country [1] [4].
3. The contested numbers: how many undocumented people does the census miss?
Estimating the unauthorized population relies on surveys like the American Community Survey (ACS) and methods that subtract legally resident immigrants from self‑reported foreign‑born counts; Pew used ACS data to estimate 14 million unauthorized immigrants in 2023, but researchers warn about timing, survey reweighting and undercount issues [1] [3]. Critics and alternative analysts produce different totals—ranging widely in other sources—and the Census Bureau has revised methods for net migration, complicating year‑to‑year comparisons; these methodological disputes mean any projection of seat or funding changes depends heavily on uncertain baseline counts [3] [4].
4. Recent churn alters the calculus: departures, enforcement and reweighting
Some analysts report sharp 2025 declines in portions of the foreign‑born population tied to policy and enforcement changes, arguing that departures since January 2025 materially reduce the population used in apportionment and funding calculations [2]. The Congressional Budget Office also lowered near‑term net‑immigration projections after administrative actions in 2025, which affects longer‑run demographic baselines [5]. These dynamics matter because apportionment is based on a snapshot: rapid inflows or outflows, or changes in survey weighting, change which people are counted—and therefore which states gain or lose relative representation [5] [2].
5. Funding formulas and local budgets: more than headline counts
Federal program dollars are often distributed using population counts, not citizenship status. Excluding undocumented residents would lower eligibility bases for Medicaid, highway funding, community development grants and many other formulas that use population or per‑capita measures, reducing dollars to jurisdictions with many undocumented residents even if residents still use services (available sources do not mention a definitive list of every program that would change under exclusion). The scale of federal funding changes tracks the size of the excluded population and which states receive disproportionate shares of those populations [1].
6. Legal and practical hurdles: counting vs. excluding
The Constitution requires an “actual Enumeration” for apportionment, but how to treat non‑citizens has been litigated and politically fraught; Congress, courts and the Census Bureau would play roles in any redesign of who’s counted. Separately, operational problems—survey nonresponse, methodological revisions and timing of ACS releases—make implementing exclusion both legally and technically complex [3] [6]. The Census Bureau’s periodic method changes and the different strengths of CPS versus ACS data illustrate limits in the available measurement tools [3] [4].
7. Competing interpretations and the political stakes
Advocates for exclusion argue that non‑citizens should not determine congressional seats; opponents point to constitutional text and the fiscal harm to communities that would lose representation and funding. Research centers and policy groups differ sharply on the size and recent trend of the undocumented population—Pew’s 14 million estimate is widely cited but contested by other analyses—so debates over exclusion are driven both by values and by disagreement about the empirical baseline [1] [7] [2]. Readers should treat any concrete numerical forecast of lost seats or dollars as conditional on which population estimate and method is used [1] [3].
Limitations: This analysis relies on the supplied sources, which emphasize methodological disputes, recent administrative actions and different data products (ACS vs. CPS). Available sources do not give a definitive seat‑by‑seat map of what would change if undocumented immigrants were excluded; producing such a map would require applying a specific exclusion policy to state counts using detailed local population data [1] [3].