What percentage of ICE arrests come from local jail bookings versus border apprehensions?
Executive summary
ICE’s own and independent analyses show that a majority of people booked into ICE custody are arrested after being identified in local jails and prisons (the so‑called Criminal Alien Program) rather than being initial border apprehensions — recent breakdowns put the jail/prison share in the majority (roughly six in ten in some year‑level summaries), and month‑by‑month snapshots during 2025 show an even larger interior share (for example, October 2025 bookings were overwhelmingly from ICE arrests rather than CBP apprehensions) [1] [2]. However, precise percentage splits vary by time period, geography and how “arrest source” is coded, and public datasets contain gaps and definitional ambiguities that make a single, definitive national percentage impossible to state from the available reporting alone [3] [4].
1. What the numbers being compared actually are — “ICE arrests,” “book‑ins,” and “border apprehensions”
Public reporting and agency tables use multiple overlapping measures — ICE posts arrest and book‑in tables, CBP posts apprehensions and arrests, and DHS OHSS posts “book‑ins by initial arresting agency” — and researchers warn these are different slices of the same pipeline: a Border Patrol “apprehension” may never become an ICE “book‑in” if expelled at the border, and many ICE arrests recorded as “in custody” originate as detainers placed on people held in local jails [5] [6] [4]. Deportation Data Project documentation stresses linked identifiers let analysts trace pathways but also warns that isolating community arrests from jail arrests is imperfect because of how “apprehension method” is coded [7] [4].
2. What recent public analyses report about the split — majority from jails and prisons
Analysts using ICE’s arrest tables have repeatedly found the majority of ICE administrative arrests stem from incarceration settings: CNN’s review and interviews reported that in 2024 about 62% of ICE arrests came from prisons and jails while roughly 27% were community arrests, with about 7% miscoded or “other” [1]. Independent month‑level reporting likewise shows interior enforcement dominating some recent months — TRAC’s October 2025 snapshot found ICE itself arrested 36,635 people while CBP arrested 4,989 of the 41,624 booked into ICE detention that month, implying roughly 88% of those book‑ins originated via ICE (interior) rather than CBP at the border for that month [2].
3. Why the split changes over time and place — policy, funding and local cooperation
ICE and DHS themselves note that enforcement patterns shift with border flows, agency priorities and available detention capacity — spikes in border crossings typically push up detention numbers, while policy directives and funding can expand interior arrests [5]. Journalistic analysis shows a political geography to the pattern: red states saw a larger proportion of arrests inside jails and prisons while blue states showed more community arrests, reflecting differences in local cooperation with ICE and sanctuary policies [1]. Independent researchers also document that changes in administrative practices under recent administrations produced massive increases in non‑criminal interior arrests, reshaping the share coming from local lockups [8] [9].
4. Data limitations and why exact national percentages are elusive
Multiple sources warn the public tables are imperfect: Prison Policy notes roughly 15% of ICE arrest rows lack state identifiers and analysts infer locations from other fields, introducing uncertainty [3]; Deportation Data Project and OHSS documentation explain “apprehension method” coding isn’t a perfect marker for community versus custody arrests and recommend joins across tables to trace pathways [4] [6]. ICE’s public statistics affirm the agency “cannot attest to subsequent transmissions” and that higher border crossings correlate with detention but don’t map one‑to‑one to book‑ins [5]. These gaps mean any published percentage must be read as a best‑estimate for specified periods and methods, not an immutable fact.
5. Bottom line and how to read future claims
The balance of evidence from ICE data and independent analysis is clear: most ICE‑booked arrests in recent years have originated from local jails/prisons rather than direct border apprehensions — commonly reported shares are in the 60% range for multi‑year snapshots and can be far higher in particular months (e.g., ~88% in October 2025) — but national totals fluctuate with policy, borders and data coding, and researchers must use careful joins and caveat missing or miscoded rows [1] [2] [3] [4]. When evaluating new claims, demand the time period, the fields used to define “source,” and whether CBP apprehensions that never resulted in ICE book‑ins were excluded or included [6] [5].