How many illegal immigrants are in the us
Executive summary
Mainstream demographic researchers converge on a current unauthorized (often called “illegal” or “undocumented”) population in the low-to-mid millions, with leading survey-based estimates clustering around 11–14 million people, while alternate methodological approaches and partisan organizations produce higher figures—ranging from about 14 million up to more than 20 million—so the best answer is a range with clear caveats about methods and recent policy shifts [1] [2] [3] [4] [5].
1. What the major nonpartisan researchers say: roughly 11–14 million
Two of the most-cited, methodologically transparent estimates sit in the 11–14 million band: state-level compilations and residual-method updates point to roughly 11–12 million undocumented people in recent years (Center for Migration Studies / Migration Research Initiative summaries) [1] [6], and the Pew Research Center—after incorporating revised Census migration inputs—reported a record 14 million unauthorized immigrants in 2023, noting that new Census estimates of international migration materially increased their totals for 2021–2023 [2] [7].
2. Why those estimates differ: methods, undercounts and new Census inputs
Differences among estimates reflect measurement choices: survey-based residual methods start from Census/ACS counts of foreign-born residents and adjust for legal populations, while demographic and operational models use enforcement, removal and overstay data to reconstruct flows; surveys can undercount unauthorized populations and recent Census revisions increased migration measures, which pushed Pew’s total upward—illustrating how inputs and undercount adjustments produce notably different totals [8] [2] [7].
3. Alternative and higher estimates: modeling and advocacy figures
Some academic models and advocacy groups produce substantially higher numbers: a 2018 Yale-affiliated demographic model estimated about 22.1 million undocumented immigrants by using operational enforcement and demographic flows rather than survey residuals [4], and organizations such as FAIR publish larger tallies—for example an 18.6 million estimate in 2025—often relying on different assumptions about net flows and recent enforcement outcomes [5]. These higher figures are methodologically plausible but contested by mainstream demographic teams who flag sensitivity to assumptions about entries, exits and undercount rates [8] [4].
4. The policy context that changes the count quickly
Legal and administrative actions can alter who is counted as “unauthorized” almost overnight: revocations of temporary protections, changes to parole or asylum rules, and mass removals or expedited border processing affect both the stock and visibility of the population, and agencies such as the CBO and Brookings see net migration and enforcement policy changes materially reducing or reshaping migration flows in 2025–26—factors that complicate any single-point estimate [9] [10] [11].
5. A reasoned number to report now—and the honest caveats
Given concordance among multiple independent survey-based teams and the recent Pew revision incorporating new Census measures, describing the current unauthorized population as approximately 11–14 million is the most defensible, empirically grounded statement, while acknowledging credible upward outliers (roughly 18–22 million) exist depending on modeling choices; any headline figure must therefore be paired with the estimation method, the year of reference, and caveats about recent policy-driven status changes that can raise or lower the count quickly [1] [2] [4] [5].
6. What reporting and readers should watch next
Future revisions from the Census Bureau, updates to DHS/OHSS population estimates, and forthcoming CBO or Brookings analyses of net migration and enforcement impacts will be the key inputs that either consolidate the 11–14 million consensus or lend weight to higher-model estimates; without those inputs, public debate risks substituting methodological preference for empirical grounding, and partisan tallies should be treated as claims tied to specific assumptions rather than settled fact [12] [9] [10].