What are reliable data sources for unauthorized immigrant population estimates during 2017–2020?
Executive summary
Reliable estimates of the unauthorized (undocumented) immigrant population for 2017–2020 rest on a handful of recurring data sources and methods: primarily the U.S. Census Bureau’s surveys (the American Community Survey and, to a lesser extent, the Current Population Survey), administrative records compiled by the Department of Homeland Security and Social Security, and research groups that apply residual and imputation methods such as the Migration Policy Institute, the Center for Migration Studies, and Pew Research—each with documented strengths and limits [1] [2] [3] [4].
1. Core survey backbone: American Community Survey (ACS) and Current Population Survey (CPS)
The ACS is the principal survey used for stock estimates because of its large sample size and rich demographic detail; most institutional residual estimates for stock counts through 2017–2020 are built from ACS microdata [1] [5]. The CPS is more timely and useful for detecting short-term trends but is much smaller and less reliable for level estimates of unauthorized populations—researchers use CPS cautiously for trend signals rather than definitive counts [1] [6].
2. Administrative pillars: DHS Yearbook / Office of Immigration Statistics and visa‑overstay records
DHS administrative data — including legal admission counts in the Yearbook of Immigration Statistics and visa‑overstay flow estimates — supply the “lawful” side of the equation and information on annual inflows and outflows that feed residual methods; the Office of Immigration Statistics publishes component data used by analysts for 2017–2020 work [1] [7]. DHS visa‑overstay compilations are particularly important because visa overstays accounted for a growing share of the unauthorized population by 2017 [7] [8].
3. Residual and imputation methods used by major research centers
The residual method—subtracting estimated lawful foreign‑born from total foreign‑born counts—underlies DHS, Center for Migration Studies (CMS), Migration Policy Institute (MPI), and Pew estimates for 2017–2020; these groups also apply imputation rules to assign likely legal status to survey respondents and reconcile administrative flows with survey totals [3] [2] [1]. Results from these institutions generally lie in a similar band, though point estimates differ by methodology and timing [9].
4. Innovative cross‑checks: Social Security and linked administrative approaches
Researchers have developed novel approaches linking survey responses to Social Security Administration records to detect mismatches (a signal of potential unauthorized status) and to validate residual results; such methods were proposed and evaluated in Social Security Bulletin work and offer an administrative cross‑check to survey‑based estimates for periods including 2017 [8] [7].
5. What the major producers reported for 2017–2020 (and their caveats)
Leading estimates for the late 2010s place the unauthorized population around roughly 10–11 million, with Pew reporting about 10.5 million in 2017 and DHS/CMS estimates producing comparable numbers depending on timing and assumptions; analysts caution that 2020 was an atypical year and survey disruptions meant some standard estimates were not produced or required adjustment [9] [3] [1]. Multiple studies stress that components—overstays vs border entries—shifted during the period and that departures, deaths, legalizations and pandemic disruptions complicate stock comparisons [7] [5].
6. Limitations, disagreements, and implicit agendas to watch
All methods rely on imperfect survey coverage, assumptions about mortality and emigration, and administrative completeness; these technical choices produce nontrivial variation across analysts [9]. Researchers and advocacy or policy organizations sometimes emphasize different sources or models to support particular policy arguments—groups vary in how they weight DHS administrative flows, CPS signals, or alternative imputations—so reading methods transparently is essential [4] [2].
7. Practical guidance: which sources to trust for 2017–2020 work
For reproducible, peer‑reviewed estimates use ACS‑based residual estimates from DHS, Pew, MPI, and CMS as primary references and consult DHS administrative flow tables and visa‑overstay reports for the component logic; treat CPS and provisional CMS/CPS‑derived figures as trend indicators rather than definitive counts, and use SSA‑linkage studies as valuable methodological validation rather than standalone totals [1] [2] [3] [8] [6] [10].