Last week, I published data leaked from the Department of Homeland Security (DHS) showing that 73 percent of individuals booked into Immigration and Customs Enforcement (ICE) detention facilities since October 1, 2025, had no criminal convictions of any kind—only 5 percent had violent criminal convictions. It included this chart:
In response, DHS called the report’s main chart “made up” with “no legitimate data behind it.” Yet we provided the raw data in the report from DHS. The broad results were checked against three separate sources, including publicly available data from DHS. Nonetheless, DHS claimed that it was inaccurate and that “~70% of illegal aliens arrested have active criminal charges or criminal convictions.”
DHS inflates the number of “criminal” arrests by including people with pending criminal charges, because many criminal charges do not end in convictions, and an ICE arrest denies them due process and their day in court. Nonetheless, DHS’s claim conflicts with the data in our report, which finds that nearly half (47 percent) of ICE book-ins had not even been charged with a crime.
DHS’s claim focuses on arrests, rather than individuals booked into ICE detention. ICE arrests exclude individuals who were arrested by Border Patrol and then referred to ICE detention. News reports indicate that Border Patrol’s interior arrests are even less focused on criminals than ICE’s are, so ICE’s preferred measure is biased in favor of finding a higher percentage of criminals.
Yet a new dataset, available on December 1, detailing all ICE arrests in early October, refutes DHS’s response and vindicates our report’s conclusions. The data were obtained directly from ICE via Freedom of Information Act requests by the Deportation Data Project, a collaboration of UC Berkeley Law School and the UCLA School of Law’s Center for Immigration Law and Policy.
The data—again directly from ICE—show that 71 percent of ICE arrests from October 1 to October 15, 2025, had no criminal convictions, and that 45 percent of ICE arrests in October 2025 had no criminal convictions or pending charges. This compares with the 73 percent and 47 percent of ICE book-ins, respectively, for the period from October 1 to November 15, 2025. The slightly lower percentage makes sense given that Border Patrol’s activities targeting non-criminals are excluded from ICE arrests.
The share of criminal convicts arrested by ICE has halved over the last year, from a weekly average of 59 percent this time last year to just 27 percent for the week ending October 15, 2025. Even if we look at ICE’s preferred measure, the share of ICE arrests without criminal convictions or even pending charges rose from just 14 percent to 47 percent, comparing the same weeks in each year.
The number of arrests of individuals without criminal convictions has risen by 585 percent year over year, and the number of arrests of people without criminal convictions or criminal charges has increased by 1,200 percent. This is the true measure of the opportunity cost, as ICE has a significantly larger budget. It could be sending more agents to track down the nearly 500,000 immigrants who ICE claims have criminal convictions and are removable, but it apparently prefers to grab easy targets to meet arbitrary arrest quotas.
As noted in our original report, the broad outlines of these facts were already confirmed by publicly available DHS data on the population currently detained by ICE (69% no convictions) and removed by ICE (70% no convictions).
Despite this manifest change in priorities, the administration continues to claim that it is arresting “the worst of the worst.” But in January, the president ordered ICE to stop prioritizing people based on their level of public safety threat. In May, the White House further ordered ICE to stop focusing on criminals or even specific people and just go out on the streets and arrest people, which has led to its demographic and racial profiling program.
Given these policies, it is no surprise that ICE has become much less focused on people with criminal histories, especially people with serious, violent criminal histories. Arrests of criminal convicts have increased, but only because DHS has diverted so much of the rest of the federal and state criminal law enforcement from active criminal investigations to arrest immigrants indiscriminately. As the New York Times reported in multiple pieces last month, federal criminal law enforcement is ignoring its basic tasks to conduct mass deportations, including by letting sex traffickers go free.
Many of these immigrant “convicts” were convicted of marijuana convictions years ago and were legal immigrants, like 75-year-old Isidro Pérez, Chris Landry, and Will Kim, a researcher who grew up in the United States. In the first few months alone, ICE deported over 600 people who had nothing more than a marijuana possession charge—77 percent of whom were 5 years old or older.
That ICE would lie about the Cato Institute’s research and its own data is concerning, but the fact that it is wasting billions of dollars on deporting peaceful people who are a fiscal and economic benefit to the United States and are valued by their US citizen families and friends is worse. ICE should reprioritize serious public safety threats and be transparent about its arrests going forward.