How do ICE arrest rates and the share without criminal convictions vary by field office or state?
Executive summary
ICE arrest rates and the proportion of arrestees without U.S. criminal convictions vary sharply by field office and state: some offices (like Dallas) report a majority of arrestees with no convictions, while other states and regions show higher shares of arrests tied to pending charges or convictions, and local policies—sanctuary laws, cooperation with federal authorities—and ICE’s shifting tactics drive much of that geographic variation [1] [2] [3]. National datasets assembled by researchers and ICE itself show a recent and large rise in arrests of people without criminal convictions, but precise shares depend on definitions (convictions vs. pending charges vs. “other immigration violators”) and the snapshot date used [4] [5] [6].
1. Arrest volumes and geography: field offices are not uniform
ICE organizes enforcement across 25 field offices and those offices report very different arrest volumes and mixes of cases; ICE’s public statistics and the Deportation Data Project both publish arrest counts by field of responsibility, allowing analysts to compare offices, but raw arrest rates per capita and the profile of arrestees diverge strongly from state to state and office to office [6] [7]. Some field offices have recorded steep increases in community or “at-large” arrests conducted in neighborhoods rather than jails, boosting overall interior enforcement numbers in certain states while other regions—particularly parts of Northern California—have maintained comparatively low arrest rates, in part because of local limits on cooperation [8] [3].
2. Who lacks a conviction? National snapshots that disagree because of definitions and timing
Multiple independent trackers find that a large share of people arrested by ICE lack U.S. criminal convictions, but the headline percentages vary by dataset and date: TRAC and advocacy compilations put the no-conviction share very high (TRAC/related reporting finds roughly 71–74% without convictions in late 2025) while other analyses note nearly half or similar large fractions depending on whether pending charges are counted separately [4] [5] [9]. Cato and other analysts have highlighted that at various points in 2025 the share of arrests without convictions rose dramatically [10], and Stateline’s read of TRAC data found periods where only a tiny fraction of detainees had convictions during certain short windows—illustrating that short-term snapshots can produce extreme-seeming percentages [11].
3. Local contrasts: Dallas, Utah and Northern California as case studies
Concrete office- and state-level contrasts illuminate the national pattern: Dallas’s ICE office saw a majority of arrestees without convictions in 2025, with pending charges becoming the largest single category in some months, signaling a pivot away from jails-based arrests [1]. Utah’s rise in ICE activity shows the opposite mix in some reporting—more than 80% of ICE arrests there in 2025 were described locally as involving pending charges or convictions—illustrating that state-level enforcement mixes can run counter to national trends and reflect targeted campaigns and DHS messaging [2]. Northern California, in contrast, has maintained a low ICE arrest rate overall; even so, reporting there found that nearly half of people arrested by ICE in the Bay Area lacked criminal convictions, a pattern tied to “community” arrests as the agency seeks removals under new enforcement priorities [3].
4. Why the variation exists—and the data limits that matter
Variation by field office and state reflects at least three forces visible in the records and reporting: changes in ICE tactics toward at-large community arrests (which capture people who often lack U.S. convictions), differential local cooperation and sanctuary policies that limit jails-to-ICE transfers, and administrative classification choices (ICE’s “Other Immigration Violators” category includes repeat re‑entrants and foreign fugitives who may have no U.S. criminal record) [8] [3] [6]. Data limits complicate precise comparisons: ICE and DHS sometimes aggregate “pending charges” with convictions in public statements, researchers use different cut‑off dates and merges, and some academic and nonprofit counts rely on leaked or FOIA-derived files that cover different time windows, producing conflicting percentage points even while agreeing on the broad trend of more non‑convicted arrestees [12] [7] [10].
5. Bottom line and reporting caveats
The bottom line is clear: field offices and states vary substantially—some offices report most arrests involving people without convictions, others report larger shares with pending charges or convictions—and the recent national surge in non‑convicted arrestees reflects both operational shifts at ICE and how the agency classifies cases; however, precise cross‑office rankings and exact percentages depend on which dataset, dates, and definitions are used, and that ambiguity should temper absolute claims from any single source [4] [5] [6].