What share of ICE arrests under Biden and Trump involved individuals with criminal convictions, by fiscal year?
Executive summary
The available public datasets and independent analyses do not provide a single, consistent fiscal‑year series that cleanly answers “what share of ICE arrests involved individuals with criminal convictions” for each fiscal year; ICE publishes arrest and detention totals but not a standardized, year‑by‑year conviction breakdown in one table [1]. Independent researchers and policy groups reach different conclusions: the Center for Immigration Studies (CIS) reports that under the Biden administration fewer than half of interior ICE arrestees had criminal convictions while roughly two‑thirds of those detained and removed under the early Trump II period had criminal histories [2], whereas other trackers such as TRAC and commentary from Cato find mixed patterns and point to large shares without convictions in various snapshots [3] [4] [5].
1. The core data problem: ICE totals exist, conviction slices do not consistently
ICE’s public “Enforcement and Removal Operations Statistics” dashboard publishes arrests, detentions and removals every fiscal period and describes typical criminal history categories (DUI, drug possession, assault, traffic offenses), but it does not lock a standardized, multi‑year table of “share with convictions” in a way that every analyst can replicate without additional processing [1]. Multiple organizations rely on ICE’s semi‑monthly spreadsheets, FOIA pulls or leaked internal datasets to construct conviction‑rate estimates, which produces different results depending on definitions (e.g., “conviction,” “pending charges,” “criminal history,” and whether border encounters are included) [2] [1].
2. What analysts report for Biden‑era arrests (FY2021–FY2024): fewer than half with convictions in some analyses
CIS’s review of ICE interior arrest data concluded that between February 2021 and December 2024 just over 500,000 interior arrests occurred and that fewer than half of those interior arrestees under the Biden administration had convictions or pending criminal charges—an estimate CIS highlights in contrasting Biden policy to later enforcement rhetoric [2]. TRAC’s reporting complicates the picture by focusing on daily averages and cumulative arrest/removal totals rather than a straight conviction share, noting Biden’s FY2024 averaged about 759 ICE arrests per day [6] [7]. ICE itself notes that ERO historically arrests a mix of people including those with non‑violent convictions such as DUI and drug possession [1], which means “criminal” in many datasets includes relatively low‑level offenses.
3. What analysts report for the Trump II period (FY2025 onward): higher claimed criminal shares but conflicting snapshots
CIS reports that roughly two‑thirds of the aliens arrested by ICE and detained/removed under the early Trump II period had criminal histories, using ICE’s contemporaneous dashboards and supplementary detention spreadsheets [2]. The Trump administration and DHS issued public statements touting large increases in arrests and emphasizing criminality, but independent trackers such as TRAC have repeatedly warned that administration claims overstate comparisons or mix apples and oranges, and TRAC’s semi‑monthly tabulations find that arrest and removal rates under Trump II are close to Biden’s recent daily averages rather than wildly higher [7] [3]. Leaked or nonpublic datasets cited by Cato suggest in later 2025 snapshots that a large fraction of detainees had no convictions (for example, Cato reports 73% with no convictions in one dataset), directly contradicting the administration’s framing and underscoring that results depend on which dataset and time window are used [5].
4. Why estimates diverge: definitions, timing, and data transparency
Disagreement derives from at least three concrete factors in the sources: differing definitions (counting “pending charges” as criminal or not), carryover from previous fiscal years in ICE’s reporting windows, and reliance on semi‑monthly cumulative spreadsheets that are periodically revised or “locked” only at fiscal‑year end, which makes snapshot comparisons fragile [1] [7] [3]. Advocacy and policy groups use selected slices—interior arrests, detainees in custody, or removals—each of which has a different criminality profile, and administrations selectively emphasize those slices that bolster their narrative [2] [8].
5. The bottom line: no definitive fiscal‑year table exists in public sources; available analyses point to variation rather than a single trend
Based on the reviewed reporting, one cannot produce a single authoritative fiscal‑year table of “share of ICE arrests with criminal convictions” without undertaking a fresh, reproducible data extraction and defining uniform inclusion rules; CIS’s analysis finds under‑50% conviction shares for Biden interior arrests while it and other sources indicate higher criminal‑history shares in some Trump II snapshots, yet TRAC and Cato analyses highlight contradictory evidence and methodological caveats that undermine an absolute claim [2] [7] [5] [1]. Any precise year‑by‑year percentage requires transparent, identical definitions applied to ICE’s underlying spreadsheets or to a released ICE variable that, as of the cited reporting, is not consistently published in a locked, multi‑year table [1].