How do demographers estimate net unauthorized immigrant population change using survey and administrative data?
Executive summary
Demographers estimate net unauthorized immigrant population change by combining household survey totals for the foreign‑born with administratively based counts of lawful immigrants, subtracting the latter from the former (the “residual”), and then adjusting for undercount, deaths, emigration and other flows to derive annual net change [1] [2]. Alternative approaches—linking survey records to Social Security Administration files or imputing legal status via statistical models that use ACS and SIPP—serve as cross‑checks and to study characteristics, but all methods face substantial uncertainty from coverage error and imperfect administrative records [3] [4] [5].
1. How net change is framed: stocks, flows, and the residual logic
Estimates start with a stock: the total foreign‑born counted in a large survey like the American Community Survey (ACS) or the Current Population Survey (CPS), then seek to separate that stock into lawfully present and unauthorized populations by subtracting an administratively constructed lawful count from the survey total; the remainder is the estimated unauthorized stock, whose year‑to‑year differences—after accounting for births, deaths, emigration and adjustments—produce net change [1] [2].
2. The residual method in practice: subtracting the lawful from the measured
The canonical “residual” approach builds a deterministic estimate of the lawful foreign‑born using administrative admissions and status data (citizens, LPRs, refugees, etc.), projects survivorship and emigration for those legal cohorts, and subtracts that legal total from the survey’s foreign‑born total—then adds a coverage adjustment because unauthorized immigrants are undercounted in surveys [1] [2] [6].
3. Administrative data: the backbone and its limits
Administrative counts from DHS and other agencies provide concrete tallies of lawful admissions and program beneficiaries, allowing researchers to sum legal admissions by country and vintage and then pare them down for deaths and emigration to produce the “L” in the equation, but these records are incomplete for some humanitarian admissions and timing revisions to population estimates have forced reweighting of surveys in recent years [6] [7].
4. Alternative methods: record linkage and imputation as cross‑checks
Researchers have developed complementary techniques—matching CPS survey responses to Social Security Administration records to flag SSN nonmatches as potential unauthorized cases, and using logical or statistical imputation with ACS and SIPP data to assign probable legal status to noncitizen respondents—which largely confirm residual estimates while offering more detail on characteristics and potential bias patterns [3] [5] [4].
5. Turning stock estimates into flows: estimating entries, exits, and net change
To move from a one‑year residual to annual net change researchers reconstruct inflows (new unauthorized entries, visa overstays, changes in humanitarian admissions) and outflows (emigration, status adjustments, removals, deaths) using administrative flows and modeled emigration rates; several studies and state‑level projects have used the residual disaggregated by year of entry combined with records on legal entrants and removals to estimate annual net changes [8] [1].
6. Uncertainty, contested revisions, and methodological trade‑offs
All approaches acknowledge sizable uncertainty: survey undercoverage of undocumented populations, errors in matching administrative records, assumptions about emigration and mortality, and recent Census Bureau revisions to net international migration that required re‑weighting of surveys have produced materially different counts across major groups (Pew, MPI, DHS, SSA), prompting method updates and debates about upward revisions since 2021 [9] [7] [2].
7. What this means for relying on estimates
Policymakers and analysts should treat unauthorized population net‑change estimates as informed, triangulated approximations rather than precise counts: the residual method remains the workhorse because it combines survey breadth and administrative grounding, while record linkage and imputation methods provide crucial checks on bias and characteristics—but all depend on assumptions about undercount, emigration, and administrative completeness that drive the margins of error [1] [3] [5].