Keep Factually independent
Whether you agree or disagree with our analysis, these conversations matter for democracy. We don't take money from political groups - even a $5 donation helps us keep it that way.
What percentage of ICE arrests are for violent crimes?
Executive summary
Available reporting shows a wide range of ways to count “ICE arrests.” Multiple data analyses find that a majority of people booked into ICE custody in 2025 had no violent convictions — one analysis says more than 93% had no violent convictions (Cato), another finds 6.9% of those with convictions had violent offenses (CBC citing Cato) — while agency and DHS statements emphasize that roughly 70% of ICE arrests involve people with criminal charges or convictions (ICE/DHS/Fox) [1] [2] [3] [4]. Discrepancies reflect different time frames, definitions (book-ins vs. detained population vs. convictions), and how immigration or traffic offenses are categorized [5] [6].
1. Definitions matter: what counts as an “ICE arrest”?
ICE produces multiple datasets and categories — Enforcement and Removal Operations (ERO) administrative arrests, “book-ins” to custody, detained populations, and press accounts of “criminal illegal aliens” — and journalists and researchers use different denominators. ICE’s own materials break arrests into people with convictions, pending charges, and those with only immigration or traffic violations [5]. Analysts warn that comparing figures without matching definitions (for example, counting every booking since Oct. 1 versus the snapshot of who is detained) produces different percentages [6].
2. Two competing headline figures: ~70% vs. large majorities with no violent convictions
DHS and ICE public statements have emphasized that about 70% of ICE arrests are of people “charged or convicted of a crime,” a message relayed by outlets like Fox and quoted in DHS releases [3] [7]. By contrast, independent analyses of ICE “book-ins” for FY2025 show that 65% had no criminal convictions and more than 93% of book-ins were never convicted of a violent offense — i.e., violent convictions make up only a small share of bookings [1] [2]. Both sets of numbers can be true simultaneously depending on which subset and time window is measured [5] [6].
3. The nuance on “violent crimes” specifically
When reporters isolate violent convictions, the available analyses point to a small share of ICE bookings. The Cato analysis reported “more than 93%” of ICE book-ins had no violent convictions, and CBC cited the same Cato data showing 6.9% of those with convictions had committed a violent crime; Stateline’s breakdown similarly found the share of people convicted of violent crime fell from 10% to 7% in a specific comparison period [1] [2] [8]. These are not contradictory if you note that many ICE encounters involve people without convictions at all or with nonviolent convictions [1] [8].
4. Why numbers shifted in 2025: policy and enforcement changes
Multiple reports attribute shifts in ICE arrest composition to policy changes under the Trump administration’s second term — officials and former ICE leaders say priorities and tactics changed, expanding arrests beyond the traditional “worst first” targeting to broader sweeps that catch people without serious criminal records [9] [10]. Journalistic investigations and court actions (e.g., Illinois lawsuits) document large operations that swept a range of people into custody and prompted judges to demand lists and threat levels for thousands of arrestees [11] [10].
5. Agency messaging and political context
DHS and ICE public statements highlight operations targeting violent offenders and present 70% as a key statistic to defend enforcement; those statements coincide with press releases touting removal of “the worst of the worst” [7] [12]. Independent analysts and civil‑liberties-minded outlets emphasize FOIA-derived booking data showing most bookings lack violent convictions [1] [6]. Each side has an implicit agenda: DHS and ICE aim to justify aggressive enforcement, while advocacy groups and some researchers stress civil‑liberties harms and the inclusion of nonviolent migrants in sweeps [3] [1] [10].
6. What the data gaps and limits are
Available sources note gaps: ICE has multiple data products and some previously public detail has been reduced or altered, complicating apples‑to‑apples comparisons [5] [6]. Journalists and courts have had to rely on FOIA-driven datasets like the Deportation Data Project; DHS and ICE provide different snapshots (press statements, news releases, statistics pages) that emphasize varying metrics [6] [5]. The sources you provided do not include a single authoritative table reconciling every category across the same time frame; therefore precise, universally applicable percentages for “violent-crime arrests” depend on your chosen dataset and definition [6] [5].
7. Bottom line for readers
If you ask “what percentage of ICE arrests are for violent crimes?” the best-supported conclusion from available reporting is: violent convictions make up a small minority of ICE book-ins or arrests in 2025 — independent analyses put violent-conviction shares in the single digits (about 6–10% depending on period) and report that over 90% of book-ins had no violent convictions — even as government statements emphasize that roughly 70% of arrests involve people with either convictions or pending charges when broader classifications are used [1] [2] [3] [8]. Choose your source and definition carefully: the headline percent swings substantially with methodological choices [6] [5].