Just 34% of people arrested by U.S. Immigration and Customs Enforcement agents during President Donald Trump's second term have a criminal conviction, according to new research published July 27 in the Proceedings of the National Academy of Science .
That's a near-historic low, the authors note.
"Our results reveal that the reality of immigration enforcement diverges sharply from the public narrative that ICE arrests are necessary to protect public safety by removing people who commit crimes," said lead author Chloe East, an associate professor of economics at the University of Colorado Boulder.
East and her colleagues analyzed 1.6 million ICE arrests from October 2015 to March 2026, looking at the location and method of each arrest and whether the person arrested had a criminal conviction.
They found that early in each of Trump's terms, his administration delivered on its promise to increase ICE activity. For example, the daily average number of ICE arrests in the first 14 months of Trump's first term was 436, up 43% from the daily average during President Barack Obama's final year in office. In the first 14 months of Trump's second term, average daily ICE arrests soared to 952 per day, up 214% from President Joe Biden's final year in office.
Increased enforcement did not, however, translate to more arrests of previously convicted criminals. Instead, the opposite happened.
Just one in three ICE arrests during Trump's first 14 months in office—for his second term—involved individuals with criminal convictions. That's down from 52% in Biden's final year in office, 70% in Trump's first year of his first term and 79% in Obama's final year.
"We found that there is an inverse relationship between the number of arrests that ICE makes and ICE's ability to target people with a criminal conviction," said East.
The study also identified stark changes in the way ICE arrests are being conducted, with more arrests now made in community gathering places like schools, workplaces and airports. In the past, ICE typically detained individuals who were already in law enforcement custody.
Prior to Trump's second inauguration, only 17% of ICE arrests were made in the community. That has risen to nearly half of all arrests.
Enforcement activity has not taken place evenly across the country, the study found. In 2025—26, democratic strongholds like Denver, Boston and Washington, D.C., were far more likely to see dramatic spikes in community arrests, the study found.
For instance, in the enforcement area that includes Denver, which has a democratic mayor and governor, ICE arrests increased 220% at the beginning of Trump's second term, with community arrests increasing 269%.
Previous studies have consistently shown that immigrants, including undocumented immigrants, are less likely to commit crimes than are people born in the United States.
East fears that by increasingly arresting those with no criminal record, heightened ICE enforcement efforts could make communities less safe by making law-abiding citizens afraid to report crimes.
Such efforts also hurt the economy, her previous research has found .
"When you remove thousands of people from the labor market through arrest, detention and deportation, that alone has a harmful effect. But we also find that those who remain work less because they are afraid," she said.
East hopes her findings will help lawmakers from all parties develop more effective immigration policies.
"Overall, our data provide important new evidence that simply increasing the funding and scale of ICE activity is unlikely to improve public safety."
