How have state-level variations in ICE arrests correlated with city-specific crime statistics since 2024?
Executive summary
State-level variation in ICE arrests since 2024 has been large and appears driven more by political, administrative and operational factors than by straightforward differences in city crime rates; researchers point to state politics and non‑citizen population shares as statistically significant predictors of arrest intensity [1], while ICE’s own dashboards and watchdog datasets show that many interior arrests are administrative rather than uniformly tied to violent crime [2] [3]. Available reporting and analyses caution that direct, robust correlations between state ICE‑arrest rates and city‑level crime statistics are not fully established in the public data because of mismatches in scope, timing and the type of arrests recorded [4] [5].
1. Political geography, not public safety, explains much of the state variation
A UCLA technical brief that modeled state variation in ICE arrests finds statistically significant associations with political variables — notably the share voting for Trump in 2024 — and the share of non‑citizens in a state’s population, suggesting enforcement intensity reflects policy choices and political context more than crime trends alone [1]. Prison Policy’s state‑by‑state per‑capita arrest rates further illustrate wide differences across states and emphasizes that high local jail arrests do not necessarily mean immigrants are more criminally involved — Texas data cited by the group even show lower arrest rates for undocumented people than for U.S.‑born residents [4].
2. ICE’s public data complicates any neat crime‑correlation story
ICE’s Enforcement and Removal Operations (ERO) dashboards distinguish administrative immigration arrests from criminal arrests and show that a large share of ERO activity is administrative enforcement of immigration law rather than prosecution for violent crimes, with datasets broken down by convictions and pending charges [2] [3]. Multiple outlets and analysts have noted that headline percentages offered by DHS or ICE (for example, claims that 70% of those arrested had convictions or pending charges) are contested and depend on how “criminal” is defined, with watchdogs finding many arrested persons had no convictions [6] [7].
3. City‑level crime data and ICE datasets are misaligned and incomplete
Reporters and researchers warn that ICE releases, local police crime stats and federal crime reporting operate on different timelines, geographies and definitions, making causal inference precarious; the Marshall Project flagged that detention and arrest data reporting was interrupted and incomplete after October 2024, limiting longitudinal linkage to municipal crime trends [5]. Where local analyses exist — for example Connecticut’s CTData showing a sharp rise in ICE arrests concentrated in Hartford in early 2025 — the reporting documents local enforcement surges but does not demonstrate that those surges tracked increases in violent crime in the city itself [8].
4. Where crime and enforcement overlap, the causal story is ambiguous
ICE and other reporters note that many ERO arrests historically involve people with convictions for DUI, drug possession or non‑violent traffic offenses rather than a broad uptick in violent crime [2] [3]. Conversely, some local law‑enforcement partnerships and task forces target foreign fugitives and serious offenders, meaning localized spikes in ICE arrests in particular jurisdictions may sometimes correspond with criminal investigations — but disentangling task‑force activity from politically driven enforcement requires data linkage not yet publicly available [2] [9].
5. Conclusions, caveats and the policy angle
The best reading of current reporting is that state‑level ICE arrest rates since 2024 correlate more clearly with political and administrative variables than with simple measures of city crime, but the inability to consistently link ICE person‑level enforcement records to municipal crime datasets prevents a definitive statistical claim about causation [1] [5] [4]. Reporting organizations and academics urge caution: claims that enforcement is primarily about violent crime are contested by researchers who find many arrests involve people without convictions and by watchdogs tracking surges of administrative arrests [7] [3]. Hidden agendas are present on both sides — political actors pushing narratives about “crime and border security” and advocates highlighting civil‑liberties harms — and the uneven public data mean journalists and policymakers must demand standardized, linkable datasets to test whether enforcement is responding to public‑safety needs or political directives [1] [5].