An earlier version of this article first appeared on Austin Kocher’s Substack. We’ve edited it for style and republished it with permission.

Even while staying busy with my own research, I try to read as much of other people’s work as I can.

Keeping up with the news is important, but I think reading deeply invested work by academic and policy experts will give you a less sensational and less emotional understanding of the immigration enforcement system than the news cycle alone.

These reports and articles take months or years to produce. They draw on data that most people never see.

And they tend to ask better questions than the ones that dominate cable news.

The problem is that a lot of this work is hard to find. Unlike books or journal articles, reports don’t have a central repository. They circulate online, and if you happen to be in the right networks you see them — and if you’re not, you don’t.

My effort here is to highlight a few pieces that you might have missed, all of which have come out recently and all of which I think represent really important work.

I’ve gone over my notes and marginalia for each of these and pulled out three key observations. Most of the pieces below are policy reports. One is a peer-reviewed academic article. Thanks to everyone named and unnamed for the massive aggregation of intellectual labor that went into these five pieces.

ICE Arrests across Trump’s First and Second Terms: Variation in Targeting, Method, and Geography
Chloe N. East, Elizabeth Cox and Caitlin Patler. NBER Working Paper, February 2026.

A lot of people have written over the past year about how immigration enforcement under Donald Trump’s second presidential administration has shifted away from people with criminal convictions.

I’ve written about this extensively myself, and this observation remains important. But it is also important to reproduce that analysis with greater academic rigor — which is exactly what this paper does — and to surface patterns in the data that have been missed, which it also does.

Using administrative data covering the complete universe of U.S. Immigration and Customs Enforcement arrests from 2015 through October 2025, this team compares the two Trump administrations directly to identify what is different between them statistically and what is driving the shift toward arresting people with less criminal history. Here are three key findings.

• The share of ICE arrestees with criminal convictions dropped sharply after the second inauguration, from about 52% to 37%. This decline was far steeper than at the start of the first Trump administration, when the share also fell but by a smaller margin.
• A major driver is the shift toward “community arrests,” which are arrests on the street, at workplaces, courthouses, and other community sites. These more than doubled as a share of all arrests, rising from about 19% to 44%. Community-based operations are far less likely to pick up people with criminal records compared to arrests conducted through law enforcement partnerships.
• Areas of responsibility containing major Democratic-controlled cities saw the largest spikes in community arrests during the second term. Across virtually all regions, as arrests increased, the share of people with criminal convictions declined.

Some in the news media have touched on this topic, but what this team brings is the kind of rigorous, comprehensive data analysis that deepens our understanding of what is driving enforcement. Digging into the full universe of arrest data and comparing the two administrations side-by-side is exactly the kind of work we need more of right now.

Mass Deportation and the Intensity of Policing in the United States’ 100-Mile Border Zone
Geoff Boyce. Law & Policy, October 2023.

This paper was published before the current administration, but it may be one of the most important pieces you can read to understand what is happening right now. Boyce’s argument is that the familiar binary between “border” and “interior” enforcement obscures far more than it reveals.

Drawing on an expansive archive of internal government records obtained through Freedom of Information Act requests and court litigation, primarily I-213 arrest forms from Border Patrol sectors in Tucson, Arizona, Buffalo, New York and Detroit, Michigan, Boyce documents what enforcement looks like across parts of the 100-mile border zone.

• In the Detroit sector, 86% of arrested noncitizens were of Latin American origin, far exceeding their share of the local foreign-born population. Agents cited contradictory justifications for stops: Slowing down was suspicious in 33% of records, but speeding up was suspicious in 17%. Avoiding eye contact was flagged in 39% of cases, while making eye contact was flagged in 25%.
• Local and state police function as critical force multipliers. In the Buffalo sector, nearly half of all arrests of long-term residents were initiated by other law enforcement agencies that channeled individuals into Border Patrol custody. Despite a nationwide directive prohibiting Border Patrol agents from acting as interpreters for local police, nearly 30% of “other agency” arrests in the Detroit sector involved agents being summoned for that purpose.
• The intensity of enforcement can be reduced through policy. When New York in 2017 banned state employees from inquiring about immigration status, “other agency” arrests in the Buffalo sector dropped from nearly 50% to 31%. When Tucson curtailed local cooperation with Border Patrol, arrests of long-term residents fell 53%.

As I recently argued on the Project Censored podcast, rather than thinking about enforcement in terms of where the border is, we should think about what the border does. Boyce’s work is essential to that reframing. He proposes an “intensity” framework: the volume, diversity, and networked interconnectivity of law enforcement institutions operating in a given area at a given time.

This helps explain why the arrival of Border Patrol agents in cities like Minneapolis and Chicago should not be understood as an aberration, but as an extension of a long-standing enforcement logic. The enforcement authority, racial profiling, and web of inter-agency cooperation that Boyce documents in border communities travel with those agents wherever they go.

Immigration Enforcement in the First Nine Months of the Second Trump Administration
Graeme Blair and David Hausman. Deportation Data Project, January 2026.

The Deportation Data Project, based at the University of California, Berkeley, School of Law — in collaboration with the University of California, Los Angeles — has been one of the most important initiatives in the immigration space over the past year. Their core contribution has been making the data available in the first place.

Through repeated FOIA requests and litigation — they sued ICE when the agency failed to respond — they have obtained and published individual-level enforcement data that the administration has otherwise refused to share transparently. Without this project, we would have very little reliable information about what the Trump administration is doing on immigration enforcement.

It’s really valuable that the team fought to get this data and took the time to analyze it. Given their deep knowledge of the datasets — and the limitations of the data — this report provides an authoritative summary and analysis that will be useful to researchers, journalists, and anyone trying to understand the full picture of what has happened over the first nine months of the second Trump administration.

• Interior deportations increased by a factor of 4.6. Street arrests, meaning arrests on sidewalks, at workplaces, and in communities — rather than transfers from jails and prisons — increased by a factor of eleven. For the two decades prior to 2025, ICE had relied overwhelmingly on custodial transfers for its interior enforcement. Street arrests at this scale are, as Blair and Hausman put it, “a new phenomenon.”
• Arrests of people without any criminal conviction increased sevenfold. Arrests of people with violent crime convictions increased by only about 30%. The shift away from targeting people with convictions was evident in both street arrests and custodial transfers.
• Once detained, virtually no one was released. Release within 60 days of arrest dropped from 16% to 3%. Voluntary departures increased by a factor of 21, a pattern the authors attribute to the coercive pressure of indefinite detention with no prospect of release. In July 2025, ICE issued guidance asserting that anyone who had entered between ports of entry was ineligible for bond regardless of how long they had lived in the U.S. Despite hundreds of federal court opinions finding this policy illegal, ICE and immigration courts have continued to apply it.

Still, the administration is not close to its stated goal of deporting one million people per year. At the most recent rate, the government would deport under 300,000 people annually. That is unprecedented in this century, but well short of the political rhetoric.

Immigration Detention Expansion in Trump’s Second Term
American Immigration Council. January 2026.

If the Deportation Data Project report gives you the numbers, this report from the nonprofit American Immigration Council — which advocates for immigrant inclusion in the U.S. — gives you the full picture: the policy architecture, funding pipeline, infrastructure buildout, conditions on the ground, and the human stories of people trapped inside the system.

It is, in my view, the definitive overview of what has happened to immigration detention during the first year of the second Trump administration.

While the overall numbers have been covered by many people at this point, some of the most valuable parts of this report are its observations about what kinds of detention facilities are being built, what is happening to people once they enter the system, and how bad conditions have gotten.

• The administration has created entirely new categories of detention infrastructure. Florida opened “Alligator Alcatraz,” a tent camp at the Dade-Collier Airport bordering the Everglades that is wholly owned and operated by the state under a 287(g) agreement, with no direct ICE involvement. No state had ever previously argued it could run its own immigration detention facility. Meanwhile, the military base tent camp at Fort Bliss — “Camp East Montana” — became the largest detention center in the country by November, holding over 2,700 people in soft-sided temporary structures, with plans for up to 5,000. The administration reportedly transferred $10 billion to the Navy to build tent facilities that could house up to 10,000 people each.
• People are disappearing inside the detention system. FOIA data show that transfers between facilities have become dramatically more common. In 2024, 47% of people taken into ICE custody were never transferred from their initial facility. In the first half of 2025, that dropped to just 23%, and the share of people transferred four or more times doubled. One person was transferred 15 times across facilities in Florida, Arizona, California, and Hawaii before being deported from Louisiana. ICE’s own detainee locator system has become unreliable, with people sometimes not appearing for weeks after arrest. One man’s family had to search a detention commissary app to find him.
• Conditions have deteriorated across the system. By April, nearly half of all detention centers were operating above contractual capacity. At the Krome North Service Processing Center in Miami, overcrowding reached nearly triple capacity, with 60 to 80 people crammed into rooms designed for 25 and women left in chains on buses for hours without access to bathrooms. At the newly reopened Delaney Hall facility in Newark, New Jersey, people were sent in while basic plumbing was not operational, food was limited to two meals a day, and a riot broke out after guards served only slices of bread for dinner. Thirty people died in ICE detention in 2025 as of December 18, more than during the COVID pandemic.

The report also documents the gutting of oversight. The Department of Homeland Security Office for Civil Rights and Civil Liberties was cut from 150 staff to 22, the Office of the Immigration Detention Ombudsman from 110 to 10, and Congressional surprise inspections have been effectively blocked.

This is an essential reference document.

Profiting Off Pain: Privatized Detention, Mass Surveillance and the Drive for Immigrant Prosecutions
American University Washington College of Law, Immigrant Justice Clinic and the National Immigration Project. January 2026.

This is the report to pair with the AIC detention report above, and the distinction matters. The AIC report documents what is happening.

This one asks why the system exists, who built it, and who profits from it.

• The report’s central argument is that criminal prosecution, privatized detention, and mass surveillance are three components of a single interlocking system. The legal engine at the center are Sections 1325 and 1326 of Title 8, which criminalize unauthorized entry and re-entry.
• The anatomy of who profits is detailed and specific. According to the report, GEO Group reported $2.42 billion in total revenue in 2024, with 41% coming from ICE contracts alone. CoreCivic drew 29% of its total revenue from ICE. GEO Group spent $1.4 million lobbying Congress and DHS in 2024, and its political action committee contributed more than $3.2 million to Republican candidates in the 2024 cycle, according to the report.
• Congressional bed quotas, first mandated in 2010 at 33,400 beds, created a built-in financial incentive to detain. ICE contracts include “tiered pricing” structures that give the agency a discount for each person detained above the guaranteed minimum, meaning ICE pays less per person the more people it locks up. The $45 billion allocation in the One Big Beautiful Bill Act caused private prison company stocks to jump between 50 and 70 percent, according to the report.

This is a report about a feedback loop: criminalization fills beds, filled beds generate profits, profits fund lobbying, lobbying produces more criminalization. The report also documents how the surveillance industry and the detention industry are converging, with companies like Anduril, Palantir and Amazon Web Services deeply embedded in the enforcement apparatus.

It flags the resuscitation of World War II-era provisions that criminalize failure to register with the federal government or to produce registration documents when stopped by a federal agent. Understanding this loop is essential to understanding why the detention system keeps growing regardless of which party is in power — and why the American Immigration Council’s projection of 135,000 beds is a business plan.

This report was produced with significant contributions from law students in American University’s Immigrant Justice Clinic, including Andrew Gamble, Kailey Kynast, Jack Murer, Junnah Mozaffar and Kimly Tran, supervised by Professors Jayesh Rathod and Chloe Sugino.

It is exactly the kind of clinic project that produces real impact. These students deserve recognition for the depth of research they helped pull together.

A Case for Reading Deeply

Thanks to all of the authors and contributors to these five reports for helping us understand the immigration enforcement apparatus from a systemic and theoretically robust perspective. This kind of work takes months or years to produce, and it matters enormously.

One of my favorite books is Cal Newport’s “Deep Work.” His argument is that the ability to focus without distraction on cognitively demanding tasks is becoming both rarer and more valuable. I think about that a lot in the context of immigration.

The news cycle can exhaust you. Instead, you need to be able to put other things aside and focus on high-quality resources like the ones I have highlighted here. Print them out if you can, sit with them, read them top to bottom and read the footnotes. That is the kind of work it takes to develop a real understanding of these systems — and it is part of my case for reading deeply.