Show Your Work: The Math Behind My Claim that 92% of ICE's Detention Growth Comes from People with No Criminal Convictions

The Wall Street Journal cited my analysis showing detention growth is driven almost entirely by people without criminal convictions. Here's exactly how I got those numbers.

On January 27, the Wall Street Journal’s Editorial Board cited my analysis of immigrant detention data in an opinion piece titled “Mass Deportation by the Numbers.” They referenced my previous finding that detention growth during roughly first quarter of fiscal year 2026 (September 21, 2025 to January 7, 2026) was driven almost entirely by people without criminal convictions. The full paragraph reads as follows:

“Syracuse professor Austin Kocher, who tracks official ICE data, finds that between Sept. 21, 2025, and Jan. 7, 2026, single-day ICE detentions increased 11,296. But only 902 of those were convicted criminals, 2,273 had pending criminal charges and 8,121 were other immigrant violators. ICE arrests have been trending upward since January 2025, but criminal arrests have plateaued.”

In the spirit of full transparency, and to further drive this finding home, today I want to explain the step-by-step process for how I got those numbers and present it visually in a way that hopefully everyone can see and understand for themselves. I used biweekly detention spreadsheets published by ICE on their website to do this analysis. Note that 🚨 ICE has begun withholding this spreadsheet from the public 🚨 in defiance of Congress’s mandate that they publish it regularly.

My goal for the analysis was to evaluate what changed in ICE’s detention numbers during the first quarter of FY 2026 (i.e., from October 1 to December 31). The closest we get in ICE’s spreadsheets is September 21, 2025 and January 7, 2026—close enough for our purposes.

First, let’s look at the overall change in detention numbers. The single-day detained population increased from 59,762 to 68,990 in ICE’s total office national detention network. That’s a total increase of 9,228 people or what we might refer to as 9,228 additional occupied beds.

It’s important to understand that this is not a static population. Theoretically, though unlikely in practice, if there was 100% turnover every two months, no single person in the first number (on September 21) would still be in detention for the second number (on January 7). If there was 0% turnover, all of the people in detention on September 21 would still be in detention on January 7—and there would be 9,228 new people. In reality, it’s a mix. Some people cycle through, some people stay for prolonged periods. This is why I think it’s reasonable to say “people,” what we’re really seeing is the growth in total occupancy.

Where does this growth come from?

People get into detention either through an arrest by CBP (usually Border Patrol) or an arrest by ICE. Although Border Patrol operates all over the country, not just at the U.S.-Mexico border, heavy border militarization and control lead to fewer border arrests, and therefore fewer detention numbers coming from CBP. Also, because people arrested by CBP have typically not been in the country very long, the likelihood that they have a criminal record is statistically very low.

This is why under President Biden—a time when border enforcement numbers were high but interior ICE enforcement was low—ICE’s detained population had even higher percentages of people with no criminal history. But to compare it this way would be an apples-to-oranges comparison to today, because today there is so much more interior enforcement and comparatively fewer people ending up in detention from border arrests. This is why I always break out the detention numbers from these two organizations and focus on detention numbers that come from ICE arrests, rather than ICE and CBP combined.

If we look at this breakdown over the three-month period, we see what we expect to see: a clear decline in people in detention from CBP and a clear increase in people in detention from ICE. The total detention increase of 9,228 sort of masks the fact that people who are in detention that came through ICE is up more than this: up 11,296 total people. We are going to now focus on this number to make sense of what’s driving this number.

Using ICE’s data, we can compare the total people detained on September 21 and January 7 by criminal history. There is total growth across the board: all three categories see an increase. However, the growth from people with criminal convictions is comparatively very small. Most of these convictions do not represent significant public safety risks or national security threats. A larger fraction represent people with pending criminal charges and the largest, by far, represents people with no criminal history. Immigrants do not have to have a criminal history to be deported. Immigration is largely civil law, and a civil administrative violation of any kind can make someone detainable and deportable. Based on what the Trump administration has been doing, even the absence of any violation can lead to detention.

When we combine the increase in people in detention based on ICE arrests only (ignoring CBP for good reason), we can clearly see that the previously observed increase of 11,296 is almost entirely driven by people with no criminal convictions (92%) and the large majority (72%) was driven by people with not even a criminal charge—nothing other than an alleged immigration violation.

If you re-read the Wall Street Journal’s accurate quote again, you’ll not only see how I got those numbers, but be able to reproduce the math for yourself if you are interested in checking the numbers:

“Syracuse professor Austin Kocher, who tracks official ICE data, finds that between Sept. 21, 2025, and Jan. 7, 2026, single-day ICE detentions increased 11,296. But only 902 of those were convicted criminals, 2,273 had pending criminal charges and 8,121 were other immigrant violators. ICE arrests have been trending upward since January 2025, but criminal arrests have plateaued.”

Thanks for following along. Feel free to circulate this figures with attribution and let me know if there are other immigration data points you’d like me to break down. If you liked this post, I would encourage you to check out my recent post that similarly breaks down the confusion surrounding deportation numbers: “Show Your Work: Visualizing the Deportation Data Problem.”