Local Immigration Enforcement

What the evidence says

State-level cooperation agreements between local law enforcement and federal immigration authorities produce a measurable increase in enforcement activity. After a state signs a 287(g) Memorandum of Agreement (MOA), the probability that ICE issues a detainer request rises by 6.24 percentage points, the probability a noncitizen is transferred into ICE custody rises by 1.81 percentage points, and the probability of removal rises by 2.36 percentage points (Nunez CDV, 2026). These effects are not uniform: in counties that already had a prior 287(g) agreement in place, the same MOA signing produces larger jumps — detainers rise 11.47 percentage points versus 4.71 in counties without prior 287(g) history, transfers rise 5.66 percentage points, and removals rise 5.54 percentage points. In counties without prior 287(g) infrastructure, only the detainer effect reaches statistical significance; transfers and removals show no measured increase. The mechanism the evidence points to is capacity: existing local-federal enforcement infrastructure amplifies what a new formal agreement can do. A separate, earlier case study of Nashville's 287(g) program (2007–2012) found it identified over 10,000 removable immigrants for deportation, and that the majority identified were Latino residents arrested for minor violations rather than serious crimes (Armenta, 2017) — describing who local enforcement programs draw in, not just how many.

The downstream health and mental-health evidence is consistent in direction across several independent studies and populations. Secure Communities increased the share of Latino immigrants living with noncitizen family members reporting mental health distress by 2.2 percentage points (a 14.7 percent relative increase), and Task Force Enforcement under 287(g) worsened mental health distress scores by 15 percent (0.08 standard deviation) (Wang, 2018). Jail Enforcement under 287(g) increased the share reporting fair or poor health by 1 percentage point (11.1 percent relative) and decreased the share reporting very good or excellent health by 4.8 to 7.0 percentage points (Wang, 2018). A separate study of ICE arrest rates found increases of 0.19 additional poor mental health days per month among Hispanic adults, a 2.2 percent relative increase in the likelihood of reporting any poor mental health days, and a 2.4 percent increase in composite mental health problems — with no measured effect among non-Hispanic adults and effects concentrated among Hispanic women (0.15 additional poor mental health days per month) (Aslim EG, 2026). The absence of a comparable effect among non-Hispanic adults, with confidence intervals reported as precisely centered around zero, supports a mechanism tied to group-specific exposure or fear rather than a general community-wide effect.

Local enforcement also produces social isolation among immigrant populations, measured directly in a 2004-respondent survey: 39 percent reported being afraid to leave the house because of local law enforcement involvement in immigration activity, 44 percent felt more isolated, and 39 percent felt under more suspicion (Simmons, 2021). Isolation was higher for undocumented respondents, for those with children under 18, and for those with personal or vicarious experience of detention or deportation, and it was higher in Phoenix and Houston — cities with stricter local enforcement regimes at the time of the survey — than in Los Angeles or Chicago. Gender showed no measured difference in isolation, breaking from earlier research suggesting women had social-network advantages. Thirty percent of respondents reported a close friend or family member deported in the past two years.

Enforcement activity also reaches children and prenatal care. Immigration raids increased daily student absences by 22 percent, with effects over three times larger for K-5 students (β = 0.269) than for high school students (β = 0.078), no measured decline in the effect from the initial raid period into the following month, and no measured effect in a comparison county over 200 miles from the raid (Dee, 2025) — indicating a geographically bounded, age-graded, and non-transitory effect. On prenatal care, 287(g) implementation increased the disparity between Hispanic/Latina and non-Hispanic/Latina mothers in late prenatal care entry by 1.5 percentage points and in inadequate prenatal care by 1.1 percentage points (Rhodes, 2014).

Several claims describe fear-driven avoidance of care and services without quantified magnitude: fear of enforcement is linked to increased preventable morbidity and mortality through deterred medical care-seeking (Sakran, 2026), anti-immigration policies are linked to decreased access to health services and increased depression, anxiety, and PTSD (Martínez, 2013), and immigration enforcement policies are linked to adverse maternal outcomes and reduced care-seeking among undocumented immigrants (anon, 2026) — all reported with low confidence in magnitude and, in two cases, hedged sourcing language. No claim in this set measures effects of local immigration enforcement on crime rates, program cost, local labor markets, or the size of the unauthorized population — these are absent from the evidence gathered here.

Where the evidence splits

Rhodes (2014) reports two different results for the same policy from the same study. Comparing Hispanic/Latina and non-Hispanic/Latina mothers directly, 287(g) implementation is associated with a 1.5 percentage point increase in the late-prenatal-care disparity and a 1.1 percentage point increase in the inadequate-prenatal-care disparity. But a difference-in-difference analysis of matched counties (n = 14) in the same study finds no significant difference in prenatal care entry or adequacy pre- and post-implementation. The source does not reconcile these directly, but the difference in method is stated: the null result comes from a small matched-county (n = 14) difference-in-difference design, while the disparity estimates come from a larger analysis comparing mothers within counties. The split is between aggregation level and sample size, not between different populations or time periods.

Thresholds and dose-response

In counties with a prior 287(g) agreement, State MOA signing produces larger increases in enforcement outcomes than in counties without one: detainer requests rise 11.47 percentage points (versus 4.71 percentage points in counties without a prior agreement), transfers to ICE custody rise 5.66 percentage points (versus a non-significant estimate in counties without a prior agreement), and removals rise 5.54 percentage points (versus a non-significant estimate) (Nunez CDV, 2026). The condition that changes the magnitude is the presence of preexisting local-federal enforcement infrastructure at the time of MOA signing, not a continuous rate or numeric cutoff.

Open empirical questions

• No source here measures the effect of local immigration enforcement on crime rates, arrests for serious offenses, or public safety outcomes in the enforcing jurisdiction.
• No source measures the fiscal cost of 287(g) or MOA programs to local or state budgets, or compares that cost to any measured benefit.
• No source measures whether local enforcement changes the size of the unauthorized immigrant population in a jurisdiction, or migration flows into or out of it.
• No source establishes how long the measured health, mental health, or isolation effects persist after an enforcement program ends or an MOA is rescinded.