Los datos y documentos de origen (nombres de agencias, acuerdos MOA, registros oficiales) permanecen en inglés tal como se publicaron originalmente. Metodología.

287(g) Agreement Type

Jail Enforcement Model

179 participating agencies · as of June 5, 2026

See in glossary →

Overview

This model only authorizes officers to investigate immigration status after someone has already been booked into jail — no community patrols, no arrests based on suspected immigration status in the field, and no street stops. Officers need an independent reason to take someone into custody.

Local officers screen and identify people already booked into jail. They can interview detained individuals about their immigration status, run DHS database checks, issue detainers, and hold someone up to 48 hours past their scheduled release so ICE can take custody.

Officers can also serve administrative arrest warrants and initiate immigration removal paperwork.

Officers can

→ Interview detained individuals about immigration status
→ Run DHS database checks (IDENT/IAFIS)
→ Issue immigration detainers
→ Hold individuals up to 48 hours past scheduled release for ICE transfer
→ Serve administrative arrest warrants
→ Initiate removal proceedings

Training requirements

4-week (160-hour) course at the Federal Law Enforcement Training Center (FLETC) in Georgia. Biennial refresher training required (reduced from annual after 2020 MOA changes).

Background

The first agreement was signed in 2002 by the Florida Department of Law Enforcement (FDLE). The program grew to 72 agencies by 2011, then contracted to 37 under the Obama administration by 2017. Growth stalled until Trump's second term, when numbers climbed to approximately 150 before accelerating further.

As of June 5, 2026, 179 agencies in 31 states have signed this type of agreement.

Key findings & history

No performance metrics (GAO-21-186, 2021): ICE "does not have goals or measures to assess program performance." Biennial compliance inspections exist on paper but are inconsistently conducted.
Accountability rollback (2020): When ICE renewed 75 JEM MOAs in 2020, it quietly removed expiration dates, the requirement for individual officers to commit to a minimum 2-year term, and the entire section describing civil rights complaint procedures — including agency points of contact. Source: ILRC analysis.

Program-wide oversight findings

▪
DHS OIG (2010): More than half of immigrants identified through JEM were arrested for misdemeanors, primarily traffic offenses. ICE and local agencies were not complying with MOA terms. At least 33 recommendations were issued.
▪
ACLU "License to Abuse" (2022): Of 142 participating agencies, 65% had documented patterns of racial profiling, 77% operated detention facilities with documented inhumane conditions.
▪
UN Committee on the Elimination of Racial Discrimination (2022): Called on the Biden administration to terminate the program, describing it as "indirectly promoting racial profiling."
▪
Cato Institute (2018): Analysis of North Carolina counties found no causal relationship between 287(g) apprehensions and crime reduction. Unexpectedly found a correlation with increased assaults on law enforcement, attributed to community trust erosion.
▪
Detainer authority (legally contested): Courts in multiple jurisdictions have held that civil immigration detainers are not binding on local jails, and that holding someone solely on a detainer past their legal release date may violate the Fourth Amendment. ICE maintains they are lawful requests. See Morales v. Chadbourne, 1st Cir. 2015.

How the three models compare

	WSO	JEM	TFM
Setting	Jail only	Jail only	Anywhere
Can interrogate suspects	No	Yes	Yes
Can initiate removal	No	Yes	Yes
Can arrest without state charges	No — warrant required	Yes (in facility)	Yes (anywhere)
Training	8 hrs online	160 hrs (4 weeks)	40 hrs online
Compliance inspections	None	Biennial	Unclear
Agencies as of June 5, 2026	504	179	1,220