- President Trump's recent clarification that ICE arrest policy remains unchanged, allowing enforcement against all illegal immigrants rather than limiting to those with criminal records.
- The statement addresses September 27, 2026 reports of internal guidance that briefly appeared to restrict non-criminal arrests and collateral detentions, prompting swift DHS denial and reaffirmation of broad authority.
- Data from Trump's second term shows ICE arrests have surged with a notable portion involving individuals lacking U.S. criminal convictions, underscoring the policy's focus on overall immigration violations beyond public safety threats alone.
ICE arrest numbers have risen sharply since early 2025, while the share of people taken into custody who already have a U.S. criminal conviction has fallen. The gap between official talking points and the underlying data is mostly about definitions and scale. Volume. ICE Enforcement and Removal Operations (ERO) made 113,431 administrative arrests in FY 2024. Independent analyses of FOIA data put interior arrests after January 20, 2025 at several times that pace: one peer-reviewed study of 1.6 million records found an average of about 952 arrests per day from January 2025 through March 2026, versus about 304 per day in the year before the second inauguration. DHS later reported monthly records of roughly 43,900 in June 2026, 50,208 in July, and 50,925 in August. Book-ins to ICE detention, a common proxy when arrest tables lag, reached about 365,500 in calendar 2025.
Criminal history. ICE groups people as convicted, pending U.S. charges, or “other immigration violators” (no known U.S. conviction or pending charge). In FY 2024, ICE said 71.7% of ERO administrative arrests had a conviction or pending charges; about 51% had an actual prior conviction. Under the later, higher-volume period, that conviction share dropped. USAFacts reported 38% with a prior conviction in 2025 and 32% in April 2026. The Deportation Data Project put the conviction share among book-ins at an all-time low of 27% in July 2026. A PNAS analysis found 34% with a conviction across the first 14 months of the second term, versus 52% in Biden’s last year and 70–79% in earlier high-priority periods. CBS, using an internal DHS document covering the first year back in office, reported that about 40% of ~393,000 arrests had no U.S. criminal record and that under 14% involved violent charges or convictions. How the numbers can both be “true.” DHS often cites “criminal aliens” as people with either convictions or pending charges, and sometimes treats illegal presence itself as the relevant offense. That produces figures near 60–70% “criminal.” Researchers who count only recorded U.S. convictions get closer to one-third. Among those who do have convictions, traffic, drug, and immigration offenses (illegal entry/re-entry) have long made up a large share; violent crimes are a smaller slice. Illegal presence is generally a civil violation; illegal re-entry after removal is a crime. Foreign convictions are often not in the U.S. files ICE publishes.
How people are arrested. The mix shifted toward at-large/community arrests and 287(g) partnerships with local police, not just jail transfers after someone has already been booked for a crime.
Community arrests rose from about 17% of the total in the year before the second term to nearly half in 2025–early 2026. When ICE expands street and collateral arrests, the pool of available convicted targets is finite, so the conviction rate falls even if the number of convicted people arrested also rises. That pattern showed up in both Trump terms, but more sharply in the second because the overall volume increase was larger. What this means for the policy fight. The administration’s legal position is that anyone unlawfully present is subject to arrest and removal. The data is consistent with that: a large and growing share of arrests are of people whose only recorded U.S. offense is an immigration violation. The same data does not support the narrower claim that most current arrests are of people already convicted of serious violent crimes. Both the surge in total arrests and the drop in the conviction share are large enough, and appear in official, FOIA, and academic sources, that they are not a rounding issue. Official ICE statistical dashboards have lagged the monthly press totals, so the most current criminality splits still come mainly from FOIA-processed datasets rather than a fully current ICE year-end report.
Criminal history. ICE groups people as convicted, pending U.S. charges, or “other immigration violators” (no known U.S. conviction or pending charge). In FY 2024, ICE said 71.7% of ERO administrative arrests had a conviction or pending charges; about 51% had an actual prior conviction. Under the later, higher-volume period, that conviction share dropped. USAFacts reported 38% with a prior conviction in 2025 and 32% in April 2026. The Deportation Data Project put the conviction share among book-ins at an all-time low of 27% in July 2026. A PNAS analysis found 34% with a conviction across the first 14 months of the second term, versus 52% in Biden’s last year and 70–79% in earlier high-priority periods. CBS, using an internal DHS document covering the first year back in office, reported that about 40% of ~393,000 arrests had no U.S. criminal record and that under 14% involved violent charges or convictions. How the numbers can both be “true.” DHS often cites “criminal aliens” as people with either convictions or pending charges, and sometimes treats illegal presence itself as the relevant offense. That produces figures near 60–70% “criminal.” Researchers who count only recorded U.S. convictions get closer to one-third. Among those who do have convictions, traffic, drug, and immigration offenses (illegal entry/re-entry) have long made up a large share; violent crimes are a smaller slice. Illegal presence is generally a civil violation; illegal re-entry after removal is a crime. Foreign convictions are often not in the U.S. files ICE publishes.
How people are arrested. The mix shifted toward at-large/community arrests and 287(g) partnerships with local police, not just jail transfers after someone has already been booked for a crime.
Community arrests rose from about 17% of the total in the year before the second term to nearly half in 2025–early 2026. When ICE expands street and collateral arrests, the pool of available convicted targets is finite, so the conviction rate falls even if the number of convicted people arrested also rises. That pattern showed up in both Trump terms, but more sharply in the second because the overall volume increase was larger. What this means for the policy fight. The administration’s legal position is that anyone unlawfully present is subject to arrest and removal. The data is consistent with that: a large and growing share of arrests are of people whose only recorded U.S. offense is an immigration violation. The same data does not support the narrower claim that most current arrests are of people already convicted of serious violent crimes. Both the surge in total arrests and the drop in the conviction share are large enough, and appear in official, FOIA, and academic sources, that they are not a rounding issue. Official ICE statistical dashboards have lagged the monthly press totals, so the most current criminality splits still come mainly from FOIA-processed datasets rather than a fully current ICE year-end report.