← All 148 executive actions

The Trump Administration's Mass Deportation Campaign and Its Effects

Immigration May 4, 2026
Our Analysis: Big Deal

The Trump administration mounted the largest interior immigration enforcement campaign in modern U.S. history, quadrupling deportations, raising street arrests roughly elevenfold, and booking around 400,000 people into detention—the majority with no criminal conviction. It tightened the border in the process: illegal crossings fell to multi-decade lows and net migration turned negative for the first time in over half a century. The administration also credits the campaign with a record drop in violent crime, but an October 2026 Atlantic analysis of FBI data found several of the steepest declines in sanctuary jurisdictions and slower declines in some states with the most ICE arrests, and researchers say the decline, which began under Biden, is not explained by enforcement.

But the campaign's central economic promise, that removing immigrants would open jobs for U.S.-born workers, is now contradicted by two independent national studies. An NBER analysis found employment among U.S.-born men without a college degree fell in high-enforcement areas with no offsetting wage gains, with roughly six immigrants leaving the workforce for every ICE arrest; a Brookings study of 86 "surge cities" put the toll at 668,000 jobs destroyed in nine months—roughly 13 to 30 jobs per arrest, including as many as 297,000 jobs that would have been held by American-born workers.

The campaign also separated an estimated 145,000 U.S.-citizen children from a parent, coincided with a two-decade high in deaths in custody and a pattern of fatal shootings by immigration agents that killed two U.S. citizens during the Minneapolis surge and, six months later, an unarmed Mexican worker who was not the target in Houston (part of more than 20 people the Times counted shot at since September 2025), and produced a roughly $1 billion detention-warehouse buying spree under DHS Inspector General audit—an initiative ICE has now largely abandoned, moving to offload seven of the eleven warehouses, bought for about $707 million, which a broker began marketing in October 2026 with bids expected around $600 million, after the GAO found more than $20 million already spent that cannot be recovered. The executive branch also shut down the Office of the Immigration Detention Ombudsman, the watchdog Congress created to investigate detention abuses and deaths. ICE has said detainees receive comprehensive medical care, but its own internal mortality reviews, briefly posted online in September 2026, found that staff broke with safe standards of care in six of eight 2025 deaths reviewed. In October 2026, ProPublica reported that immigration agents had detained 506 U.S. citizens from January 2025 through September 2026. Then-Homeland Security Secretary Kristi Noem had said in October 2025 that no citizens had been arrested or detained, a claim records show is false.

Courts have begun pushing back: a federal judge in June 2026 blocked ICE's courthouse-arrest policy nationwide as an APA violation, even as the Supreme Court ruled 6-3 that the administration could strip Temporary Protected Status from roughly 350,000 Haitians and 6,000 Syrians; by August, ICE was detaining Haitians and fitting others with ankle monitors in Springfield, Ohio. Meanwhile the deportation machinery's foreign-policy dimension has come under scrutiny: roughly 19,000 people have been sent to two dozen third countries under 27 mostly undisclosed agreements lubricated by cash payments and facilitated in part by the UN's migration agency, and a July 2026 lawsuit alleges ICE shared confidential asylum-application data with Iran's government to obtain travel documents for deportees—allegations DHS denies.

Details

Over the first sixteen months of President Trump's second term, the federal government built and ran the largest interior immigration enforcement campaign in modern American history. Deportations of people arrested inside the United States more than quadrupled, street arrests rose roughly elevenfold, and ICE booked around 400,000 people into detention. The Trump administration justified the effort on two grounds: that it would secure the border, and that removing immigrants would create jobs and raise wages for U.S.-born workers. On the first claim, the data is clear. Crossings collapsed. On the second, two independent national studies now point the other way, and that evidence is arriving alongside a mounting record of family separations, deaths in custody, a pattern of fatal shootings by immigration agents that has spread well beyond its Minneapolis flashpoint, adverse court rulings, and a detention build-out the administration itself is now partially unwinding.

The Economic Promise and the Studies That Undercut It

The Trump administration repeatedly argued that deporting unauthorized workers would expand opportunity for Americans. A May 2026 working paper published by the National Bureau of Economic Research, written by economists Elizabeth Cox and Chloe East of the University of Colorado Boulder, is the first national, causal analysis of how the enforcement surge actually affected the labor market, and it found no support for that claim.

The authors compared regions where ICE arrests spiked sharply between January and October 2025 against regions where they did not, using the Current Population Survey and detailed ICE arrest records. In high-enforcement areas, employment among likely undocumented immigrants who remained fell by about 4%, with the effect concentrated among men, whose employment dropped roughly 5% and weekly hours fell by about two hours. Crucially, U.S.-born workers did not fill the gap. The authors found "no evidence of positive spillover effects to U.S.-born workers". U.S.-born men with a high-school education or less, working in immigrant-heavy sectors such as construction, manufacturing, and agriculture, saw their employment decline rather than rise. In a statement accompanying the paper, East said heightened ICE activity "is harming the labor market overall," and "if anything, job opportunities for U.S.-born workers are going down as a result."

The researchers also found no sign that employers raised wages to lure Americans into the vacated jobs, which they read as evidence of falling overall labor demand rather than a tightening market. One widely cited estimate from the analysis, reported by Fortune, is that roughly six immigrant workers left the workforce for every ICE arrest. Many left out of fear rather than removal, as enforcement chilled entire communities. The authors attribute the unusually large chilling effect to a shift in tactics toward indiscriminate arrests in public places, including streets, schools, and churches.

A second national study, published by the Brookings Institution on May 29, 2026 by Marcela Escobari, Ian Seyal, and Paul Beach, reached the same conclusion by a different route and attached a larger number to the damage. Comparing the 86 metropolitan areas that saw the sharpest ICE arrest increases in the first half of 2025 (the "surge cities") against 255 cities that did not, using payroll data from the Quarterly Census of Employment and Wages and ICE arrest records from the Deportation Data Project, the researchers found roughly 668,000 jobs lost across the surge cities. Employment fell 0.73% below expected levels on average three to eight months after a surge, and 1.48% in the 51 cities that could be observed at least six months out—meaning the losses deepened over time. With about 52,000 excess arrests across the surge cities, that works out to roughly 13 jobs lost per arrest, rising to about 30 in the cities tracked longest, figures widely reported by Fortune. The study estimates that between 51,000 and 297,000 of the lost jobs would have been held by American-born workers—one to six U.S.-born jobs lost per arrest. Construction was hit hardest, down 2.2% (about 102,000 jobs), followed by accommodation and food services; even arts and entertainment, which employs few undocumented workers, declined sharply, which the authors read as evidence that suppressed consumer demand, not just missing workers, drove the losses.

Escobari carried the findings into a June 26, 2026 New York Times guest essay, "ICE Is a Job Killer Everywhere It Goes"—an opinion piece, but one resting on the peer-visible Brookings data. She argued the core mechanism is that immigrant and native workers are complements, not substitutes: immigrants take on physically demanding tasks that let U.S.-born workers shift into supervisory and coordination roles, and they are about 80% more likely to start businesses that then hire Americans. The essay also punctured the assumption that enforcement fell mainly on Democratic strongholds: 56 of the 86 surge metros were in red states, primarily Texas, Florida, and Georgia, which therefore bore the brunt of the job losses. The administration's self-described "shock and awe" strategy—workplace raids, flooding immigrant neighborhoods with agents, holding children as leverage—was designed to spread fear and induce self-deportation, and the fear itself did most of the economic damage. After a major ICE operation was announced in Los Angeles in May 2025, spending in immigrant-heavy neighborhoods fell 20 to 25 percent over two months, costing the region an estimated $625 million in lost sales. How many people actually self-deported is contested: the administration claims, with little evidence, more than two million, while independent researchers put the figure closer to 200,000. One Rio Grande Valley developer told the Brookings team he had stopped building homes and started selling empty lots because retaining a crew had become impossible; another builder who once employed about 100 people finished his last home and began liquidating.

The direction of these findings is not novel in the economics literature. Prior research on enforcement episodes in the 1930s and 2010s, including East's own earlier work, generally found that when unauthorized workers leave, U.S.-born workers do not simply replace them. What was striking in 2026 was the messenger that amplified it. The Wall Street Journal editorial board, generally sympathetic to immigration enforcement, used the NBER study to argue that deporting noncriminal workers is, in its words, economically counterproductive. That was an unusual rebuke of the policy's stated rationale from the conservative editorial page.

Scale and Targeting of the Crackdown

The enforcement surge was real and historically large. According to an analysis by the Deportation Data Project, based at UCLA and UC Berkeley Law, interior deportations rose 4.6-fold in the Trump administration's first nine months compared with the last half of 2024, and street arrests in communities rose by a factor of eleven. ICE arrests resulting in detention quadrupled. The number of people booked into ICE detention from interior arrests reached roughly 400,000. DHS itself has claimed more than 605,000 total deportations in 2025, a figure that includes border removals.

The Trump administration described its targets as "the worst of the worst." The data tells a more complicated story. A Cato Institute analysis of detention records found that 73% of people in ICE custody in early FY2026 had no criminal conviction at all, and only about 5% had a violent conviction. The American Immigration Council found that more than one in three people deported from detention in 2025 had no criminal record at all, meaning no pending charges and no prior convictions, and that among those who did, roughly two-thirds had nothing more serious than a misdemeanor. Just 2% were tagged as suspected gang members. DHS, for its part, maintains that 70% of those arrested by ICE have been charged with or convicted of a crime, a figure that counts immigration-related offenses and pending charges differently than the independent analyses do.

Despite the scale, the campaign fell short of its own benchmark. The Trump administration promised one million deportations a year, which would require roughly 2,700 removals a day. Through the first half of FY2026, ICE was averaging about 1,266 removals per day, putting full-year removals on track for somewhere above 460,000. That is unprecedented for this century, but well under half the stated target.

A Nationwide Injunction Against Courthouse Arrests

One of the campaign's signature tactics—stationing agents at immigration courts to arrest people, many with pending asylum claims, the moment their hearings ended—was struck down in June 2026. Beginning in May and June 2025, ICE agents made arrests at courthouses nationwide, often immediately after DHS lawyers moved to dismiss a person's case so they could be funneled into fast-track expedited removal. On June 23, 2026, U.S. District Judge P. Casey Pitts of the Northern District of California blocked the policy nationwide in the certified class action Pablo Sequen v. United States, extending relief he had granted across Northern California in December 2025. Pitts found ICE's May 2025 directive permitting courthouse arrests violated the Administrative Procedure Act, calling it "an irrational departure" from the 2021 restrictions that had limited such arrests to national-security threats or evidence-destruction risks and writing, per Courthouse News, that the agency displayed "a complete lack of decision-making." He separately vacated ICE's waiver of the 12-hour detention limit in short-term holding facilities, restoring bond and parole consideration for people swept up under the policy. One immigration judge reported "a dramatic decline in attendance at master calendar hearings" while the arrest policy was in force. The ACLU of Northern California, which brought the case, called the ruling a "crushing blow to Trump's mass deportation agenda." DHS criticized the decision; the government's appeal of the earlier preliminary injunction remains pending at the Ninth Circuit. The courthouse-arrest tactic had driven a substantial share of the record wave of habeas challenges filed by detained immigrants, making the ruling one of the most significant judicial checks on the campaign to date.

Ending Temporary Protected Status

The campaign's reach expanded sharply on June 25, 2026, when the Supreme Court cleared the way for the administration to strip Temporary Protected Status from roughly 350,000 Haitians and 6,000 Syrians, exposing hundreds of thousands more people to potential deportation. In a 6-3 decision in Mullin v. Doe, the Court overturned lower-court orders that had postponed the terminations. Writing for the majority, Justice Samuel Alito held that the TPS statute's bar on judicial review "is clear, and its plain meaning is very broad," and rejected arguments that Trump's derogatory comments about Haitians showed the decision was tainted by prejudice, calling them "insufficient to show that the termination of Haiti's TPS designation was based on the race of the Haitian people." In a dissent joined by Justices Sotomayor and Jackson, Justice Elena Kagan wrote that Trump's statements "fairly shout, in their racial undertones and overtones alike, that race entered into the president's resolve to remove Haitians from this country," noting he had said Haitians "probably have AIDS" and had amplified false rumors that Haitian immigrants in Ohio were eating pets.

TPS, created by Congress in 1990, lets people already in the United States stay with work permits when their home countries are deemed too dangerous to return to; Haitians were first covered after the 2010 earthquake and Syrians during the civil war that began in 2012. The ruling was part of a broader push by DHS to end protections covering people from 13 countries, and it followed an earlier Supreme Court order that let the administration terminate TPS for Venezuelans. Immigration lawyers argued the terminations were carried out through an improperly fast process even though both countries remain dangerous, noting in court filings that four Haitian women deported in February 2026 were later found beheaded. The House had passed an extension of deportation protections for Haitians on a bipartisan vote in April, but the Senate had not acted. (The same day, in a separate decision, the Court also cleared the way for the administration to revive a policy restricting immigrants seeking asylum; on the two-track pattern of the Court's treatment of Trump's agenda, see the companion article on his second-term Supreme Court record.)

Enforcement then reached Springfield, Ohio, where roughly 15,000 Haitians had settled in recent years and where false rumors that Haitian immigrants were eating pets had spread in 2024. According to The Atlantic, ICE operations in Springfield began in early August 2026. Agents in unmarked vehicles detained people at apartment complexes, parking lots, gas stations and traffic stops, and ICE ordered other Haitians to check in at field offices, where they were fitted with ankle monitors. Lawyers and local activists told the magazine that at least 74 people had been detained by early October, nearly all of them Haitians and almost none with criminal convictions, and that dozens more had been fitted with monitors, including pregnant women and elderly people. On August 31, Pierre Damas Bel, a 20-year-old Haitian who had arrived legally in March 2024, held TPS and had just enrolled at Wright State University, walked into traffic on Interstate 70 near Springfield and was struck and killed by a truck. Authorities investigated the death as a suicide. His father, Ronal Bel, said in a statement that after his son was fitted with an ankle monitor, "he felt like he was treated like an animal," and that classmates had bullied him over it. Carl Ruby, a Springfield pastor who supports the Haitian community, told reporters, "It's our immigration policy that took his life." The truck driver, who voted for Trump in 2024, told The Atlantic that Bel "was doing what he was supposed to do" and "We did not do right by him." Ohio Governor Mike DeWine, a Republican who has called the effort to deport Haitians "a mistake," attended Bel's funeral in Springfield in early October.

Third-Country Deportations and the UN Agency Helping to Run Them

A dimension of the campaign that operated largely out of public view is its third-country deportation program: removing people not to their home countries but to nations with which they often have no connection. A PassBlue investigation published June 30, 2026 reported that roughly 19,000 people have been deported to 24 countries under 27 third-country removal agreements, many of which have never been publicly disclosed. Receiving countries include El Salvador (whose CECOT mega-prison has documented torture), South Sudan, Cameroon, the Democratic Republic of Congo (where the administration proposed sending 1,100 Afghan deportees), Equatorial Guinea, Ghana, Eswatini, Rwanda, Panama, Costa Rica, Ecuador, Libya, Mexico, and the Central African Republic—a country the State Department tells Americans not to visit "for any reason," and whose first planeload of deportees included Iranian activists, asylum seekers, and citizens of Jordan, Armenia, and Türkiye. Human Rights Watch reported in May 2026 that Mexico alone had received almost 13,000 third-country nationals, mostly Cubans.

The agreements were secured with a mix of cash and coercion, according to the investigation: payments of $7.5 million to Rwanda, $5.1 million to Eswatini, and $7.5 million to Equatorial Guinea; $30 million in UNHCR funds withheld from Cameroon until it agreed to accept deportees; and tariff threats, visa revocations, and aid withdrawal deployed as leverage elsewhere. The most striking finding is the role of the United Nations' own International Organization for Migration, led since 2023 by American Amy Pope. IOM provides "post-arrival humanitarian assistance" for deportees in the Central African Republic—work for which the U.S. awarded it $85 million in 2026—and collaborates in Libya with border guards described as notorious for torturing, sexually abusing, and extorting detainees. Critics quoted in the investigation call this "blue washing": a nominally humanitarian UN agency lending legitimacy to migration control for its largest donor. IOM stopped publishing detailed return data in 2020, lacks the treaty-based protection obligations that bind the UN refugee agency, and a 2022 UN Human Rights Office report found migrants in comparable programs were "virtually compelled" to accept returns. The investigation documented individual harms: IOM rapidly returned a gay Moroccan woman from Cameroon to Morocco, where homosexuality is illegal, without letting her meet legal counsel—she is now in hiding—and 14 deportees to Equatorial Guinea, who believed they were being transferred to another U.S. facility, were instead detained in a hotel with rotten food and limited sanitation, only one of them allowed a lawyer. The African Coalition Against Externalization filed suit on their behalf in June 2026. UN High Commissioner for Human Rights Volker Türk has said some of the deportations "have been conducted hastily, without properly considering the risks of torture."

A Lawsuit Alleges Asylum Seekers' Files Were Shared With Iran

The deportation machinery's diplomatic dealings drew a sharper legal challenge on July 7, 2026, when the Iranian American Legal Defense Fund and Public Citizen Litigation Group filed suit in federal court in Washington alleging that the administration unlawfully gave Iran's government confidential information from the asylum applications of Iranians it planned to deport. As the Washington Post reported, asylum-application information is confidential by federal regulation precisely because disclosure to an applicant's home government can expose the applicant and their family to persecution; the complaint contends the shared materials could jeopardize the lives of pro-democracy protesters, religious minorities, and LGBTQ Iranians. "The law couldn't be more clear that information in asylum applications is protected," said Public Citizen attorney Michael Kirkpatrick, calling the case "potentially a matter of life and death."

The lawsuit, which names Secretary of State Marco Rubio, DHS Secretary Markwayne Mullin, and acting ICE director David Venturella, describes secret meetings beginning in March 2025 between administration officials and representatives of the Iranian Interests Section at Pakistan's embassy (Tehran's proxy in Washington, since the two countries lack diplomatic relations), at which U.S. officials allegedly handed over a list of roughly 150 names and agreed to monthly information-sharing meetings about Iranians in U.S. custody. In-person meetings allegedly halted when the U.S.-Israel war with Iran began in late February 2026, but ICE continued sending documentation. Advocacy groups believe roughly 115 Iranians have been returned to Iran on three deportation flights—the most recent on January 26, 2026, weeks after anti-government protests in Iran were met with a crackdown estimated to have killed thousands, and days after Trump told the protesters "help is on the way." The State Department declined to comment on "private diplomatic discussions," while DHS spokeswoman Lauren Bis said allegations that ICE shared asylum records with Iran "are FALSE," adding that "ICE meets and works to get travel documents for detainees with every country." The suit, also covered by NPR and CNN, seeks an end to the alleged practice and notification of everyone whose information was shared. The allegations echo the pattern in the tracked IRS-ICE data-sharing episode, in which confidential government data was repurposed for deportation enforcement.

What the Trump Administration Counts as Success

The strongest part of the Trump administration's case is the border. By its own accounting, illegal crossings fell to their lowest levels since the 1970s, with Border Patrol apprehensions on the Southwest border averaging roughly 245 a day, about 95% below the prior administration's average, and the United States recording negative net migration in 2025 for the first time in over fifty years. CBP reported multiple consecutive months of "zero releases" into the interior. Supporters argue these outcomes vindicate a deterrence-first strategy and a restoration of the rule of law.

That picture comes with two caveats. First, much of the decline predates 2025: monthly border apprehensions fell from roughly a quarter-million in December 2023 to about 50,000 by December 2024, before Trump took office. Academic work by Adam Cox, David Hausman, and Mary Hoopes finds that border policy generally has smaller effects on migrant flows than commonly assumed, and that "most of the decline happened on Biden's watch." Second, the border result is largely separable from the interior deportation campaign that is the subject of the economic and humanitarian critiques below; one can credit the former without endorsing the latter.

Public opinion is genuinely divided, and supporters can point to real majorities. The White House cites polling showing 73% say criminal undocumented immigrants should be deported and 61% support deporting unauthorized immigrants generally. A New York Times/Siena poll found 54% support for deporting immigrants in the country illegally. The pattern across most surveys, however, is that support holds for removing serious criminals but erodes sharply when pollsters specify long-term residents or people with no criminal record, which is exactly the population that has made up the majority of recent arrests.

The Claim That Deportations Lowered Crime

Violent crime fell sharply in Trump's second term. The national homicide rate dropped to its lowest level since at least 1900, and violent crime fell at the fastest pace on record, according to The Atlantic. Trump and White House adviser Stephen Miller have credited the immigration crackdown. In his September 22, 2026, address to the United Nations General Assembly, Trump described the border changes and said, "As a result, last year we achieved the largest drop in violent crime ever recorded." At an August 2026 rally in South Carolina, he said, "If you wanna get the migrant criminals out of your communities, you have to vote Republican."

On October 1, 2026, The Atlantic's Jack Rodriquez-Vars and Nick Miroff published an analysis of FBI and census crime data alongside ICE arrest data from the Deportation Data Project. It found that the declines did not track enforcement. Three of the five jurisdictions with the largest change in violent-crime rates were Connecticut, Illinois and the District of Columbia, all of which the Justice Department labels sanctuary jurisdictions. Six of the 11 states where the homicide rate fell fastest were on the department's sanctuary list. The states with the most ICE arrests did not have the largest declines. Georgia, consistently among the top 10 states for ICE arrests, ranked 33rd in its drop in the violent-crime rate. Virginia, also in the top 10, ranked 34th. Florida, second or third in monthly ICE arrests, ranked 18th. Some cities with many ICE arrests, including Houston, Memphis and Nashville, did report large declines. The authors wrote that the data "do not show cause and effect, let alone a clear correlation."

The analysis also noted that violent crime had been falling since the 1990s, spiked during the pandemic, and resumed falling under President Biden. Homicides also fell in countries without a mass-deportation campaign. Canada's homicide rate fell 16% from 2024 to 2025, and homicides in England and Wales reached a nearly 50-year low. Researchers interviewed by the magazine said enforcement was not a significant factor. A study led by UC Irvine criminologist Charis Kubrin, published August 10, 2026 in the Journal of Urban Affairs, tracked 11,500 neighborhoods from 2010 to 2018. It found that neighborhoods where the undocumented share grew saw larger drops in property crime and no significant change in violent crime. Charles Fain Lehman of the Manhattan Institute, a center-right think tank, told The Atlantic that researchers generally do not regard immigration enforcement as a significant cause of the decline, although removing a serious offender could reduce crime in a narrow, mechanical way.

The Department of Homeland Security told The Atlantic that comparing state ICE arrest totals with year-over-year changes in violent crime was improper. It did not provide evidence for the causal claim. "It's common sense: when you remove illegal aliens off our streets, the violent U.S. crime rate DROPS," Homeland Security Secretary Markwayne Mullin said in a statement.

The Macroeconomic Picture

Before 2025, economists warned that mass deportation could badly damage the economy. The Peterson Institute for International Economics had estimated that deporting 8.3 million people would cut GDP 7.4% by 2028, with even a 1.3 million removal scenario shaving 1.2% off GDP. The Senate Joint Economic Committee's Democratic staff warned of up to 44,000 lost U.S.-born jobs for every 500,000 immigrants removed from the labor force. Those worst-case scenarios have not materialized, largely because actual removals stayed far below the levels they assumed.

What actually happened in 2025–26 was a more modest drag. Brookings economists estimated that the combined effect of reduced inflows and increased outflows lowered 2025 GDP growth by roughly 0.2 to 0.3 percentage points, with a similar or smaller effect projected for 2026, and little net effect on inflation, because immigrants subtract from both labor supply and consumer demand. The Penn Wharton Budget Model projected that a sustained permanent deportation policy would still reduce GDP and, notably, lower wages for the 63% of workers who are high-skilled, while delivering wage gains to authorized low-skilled workers only if the policy is maintained for years. As Escobari noted in her Times essay, the localized damage is easy to miss in national statistics: the unemployment rate has hovered between 4.0 and 4.5 percent since 2025, a range too coarse to register concentrated metro-level shocks.

The labor-force effects are real but their precise size is disputed. Federal Reserve Chair Jerome Powell noted that immigration policy had markedly slowed the flow into the labor force, and Bureau of Labor Statistics survey data showed the foreign-born labor force declining over 2025. But economists including those at PIIE caution that the widely circulated figure of a 1–2 million drop is likely overstated, in part because immigrants have grown wary of answering government surveys amid the crackdown. By spring 2026, Indeed Hiring Lab data showed the foreign-born share of the labor force had largely rebounded to near record highs.

The composition of what job growth remains tells its own story about the policy's premises. In a July 2026 essay, "The Pink-Collar Economy Is Here," University of Michigan economist Justin Wolfers—writing commentary, but from Bureau of Labor Statistics payroll data—calculated that 86% of the net new payroll jobs created during Trump's second term (403,000 of 468,000 at his data vintage; roughly 90% of about 716,000 by early July) went to women, pushing women past 50% of nonfarm payroll employment for the first time. Female-dominated industries added 828,000 jobs while male-dominated industries—the construction and manufacturing sectors that both mass deportation and the tariff agenda purport to champion—lost 218,000. Nearly all of the growth came from a single sector, private education and health services. Wolfers' argument is that the administration is chasing an outdated vision of the economy, trying to revive male-dominated goods production while the actual growth is in care and service work it undervalues. The deportation campaign compounds the pattern: the immigrant-heavy sectors it disrupted most—construction, manufacturing, agriculture—are precisely the male-dominated industries that shed jobs, consistent with the NBER and Brookings findings that removals hurt rather than helped the U.S.-born men working alongside immigrants. Overall job growth, meanwhile, was anemic: CEPR's analysis of the June 2026 jobs report counted just 57,000 jobs added that month with unemployment at 4.2%. The macro story remains fluid, even as the localized harms documented by the NBER and Brookings studies are clear.

Family Separations on an Unprecedented Scale

The human cost falls heavily on children, most of them American citizens. A May 2026 Brookings Institution analysis by Tara Watson and Maria Cancian estimated that about 205,000 children have had a parent detained since the campaign began, including roughly 145,000 U.S.-citizen children. More than 22,000 citizen children, the researchers estimated, have had every co-resident parent detained, leaving them without a parent in the home. By contrast, the first Trump administration's "zero tolerance" policy separated about 5,500 children at the border in 2018, an episode that ended after national outrage.

The Brookings figure is more than double the roughly 60,000 U.S.-born children that DHS data would imply. The authors argue official statistics undercount because ICE agents inconsistently ask about children and detainees fear disclosing them, worried about exposing caregivers. ProPublica, which the Brookings analysis drew on, documented at least 11,000 affected American children in just the first seven months using a more conservative methodology, and found that roughly 60% of detained mothers were ultimately deported.

These separations cascade through families and institutions. Most children stay with friends or relatives rather than entering foster care, but those arrangements are often fragile: older siblings raising younger ones, neighbors without legal custody, parents unable to manage young children alone. As reported by The New York Times, some detained parents face an impossible choice between being deported without their U.S.-citizen children or arranging passports and travel for them under threat of imminent removal. DHS responds that parents are given a choice of being removed with their children or designating a caregiver, and that "being in detention is a choice." Brookings counters that there is no systematic government approach to safeguarding the well-being of detainees' children. ICE involves itself in child protection only when children are present at an arrest with no immediate alternative care.

Deaths in ICE Custody

The campaign's enforcement tactics produced a mounting toll inside detention. By mid-July 2026, roughly 50 to 52 people had died in ICE custody since January 2025—about 31 in 2025 alone, the deadliest year in immigration detention in roughly two decades, and another 19 to 21 by mid-2026. Human Rights Watch and Physicians for Human Rights, in a June 2026 review titled "Dying in Detention," counted at least 52 deaths in the administration's first 500 days and found the in-custody mortality rate had risen roughly 140 percent—faster than the detained population grew. ICE's own official count stood at 51 in June 2026, a figure independent trackers call an undercount: on June 5, 2026, the agency stopped counting deaths that occur within 30 days of a detainee's release.

The dead were disproportionately people with no criminal record. Roughly three-quarters of those who died in 2025 had no criminal conviction, and a KFF analysis of 46 deaths through mid-March 2026 found about 32 stemmed from medical causes or worsening chronic conditions, nine were suicides—up from a single suicide in all of 2024—and 36 of the 46 died within three months of being detained. The population itself grew older and sicker after the administration rescinded Biden-era directives that had limited the detention of elderly and ill immigrants; the detained population climbed from about 39,000 in December 2024 to more than 68,000 by February 2026. A CNN investigation documented "crisis-level" health-care understaffing—finding, for example, that California's Adelanto facility had fewer physicians in mid-2025, when it held more than 2,000 people, than it had in early 2021 when it held fewer than 100. The dismantling of detention oversight that accompanied these deaths is documented in the next section.

ICE has publicly rejected the idea that poor care contributed to these deaths. The agency has said its facilities "provide comprehensive medical, dental and mental health care" and offer "the best healthcare that many aliens have received in their entire lives," and one of its death notices said that "at no time during detention is a detained illegal alien denied emergent care," according to an April 2026 San Francisco Chronicle investigation. That investigation sent the records of 32 deaths to 14 physicians, who concluded that at least 17 people died after medical staff delayed or failed to provide care that might have saved them. People were also dying faster: after a median of 27 days in custody in 2025, compared with 74 days from 2018 to 2024. ICE did not respond to the Chronicle's questions.

ICE's own reviewers reached similar conclusions in several cases. On September 25, 2026, the agency posted to its online records library eight mortality reviews, the internal assessments its Health Services Corps prepares after an in-custody death, covering deaths in 2025. Six of the eight found that staff had broken with safe standards of care, The Intercept reported on September 30: staff mismanaged medication, skipped physical exams, ignored lab orders, and delayed treatment. The review of Marie Ange Blaise found that staff "failed to initiate" CPR after she collapsed at the Broward Transitional Center. The review of Johnny Noviello, a 49-year-old Canadian who died at the Federal Detention Center in Miami of what his autopsy called preventable kidney failure, found that labs ordered at intake were never completed and that staff did not provide effective CPR. Reviews of two other men, Nhon Nguyen and Oscar Duarte Rascon, found their care met standards, though ICE medical staff had asked for Nguyen's release because they lacked the resources to care for him. About four days after posting the reviews, and shortly after a lawyer representing the Intercept reporter in a Freedom of Information Act lawsuit noticed them, ICE took them offline; copies remain in web archives. ICE did not respond to The Intercept's request for comment. By The Intercept's count, at least 56 people had died in ICE custody under the second Trump administration by late September 2026.

A selection of named cases illustrates the range—medical neglect, denied medication, and a sharp rise in suicides:

Name Died Age Nationality Facility Cause / disputed circumstance
Maksym Chernyak Feb 2025 44 Ukraine Krome, FL Stroke; staff allegedly delayed calling 911 for ~40 minutes
Marie Ange Blaise Apr 2025 44 Haiti Broward Transitional Center, FL Cardiovascular; family says staff refused help for chest pain
Jesus Molina-Veya Jun 2025 45 Mexico Stewart, GA Apparent suicide (one of several at Stewart)
Isidro Pérez Jun 2025 75 Cuba Krome, FL Undetermined; oldest ICE detainee death since 2006
Ismael Ayala-Uribe Sep 2025 39 Mexico (DACA) Adelanto, CA Cardiac arrest from septic shock; alleged mishandled infection
Mohammad Nazeer Paktiawal Mar 2026 41 Afghanistan (U.S. ally) Dallas, TX (hospital) Anaphylaxis/asthma; agents rejected his wife's attempt to hand over his inhaler
Aled Carbonell-Betancourt Apr 2026 27 Cuba FDC Miami, FL Apparent suicide; held with no criminal charges (recorded as the 47th death)
Adrian Andreas Florian Jun 2026 85 Germany Harlingen, TX (hospital) Detained ~10 months despite dementia and chronic illness

Several of these rulings—Chernyak's, Blaise's, and Paktiawal's—were officially "natural causes" or "accident" but are disputed by families and physicians. Physicians for Human Rights concluded that many of the deaths it reviewed "may have been preventable."

Fatal Shootings by Agents

Beyond the custody deaths, the campaign's field operations produced a series of fatal shootings by ICE and Border Patrol agents—at least eleven people killed since January 2025, roughly seven of them in enforcement stops where the person was not attacking agents, and three of them U.S. citizens. Because these shootings share a recurring template—a disputed "weaponized vehicle" self-defense claim, no body-camera footage, video and witnesses contradicting the official account, and evidence withheld from local investigators—they are documented in full in a dedicated entry, Fatal Shootings by ICE and Border Patrol Agents.

The most visible flashpoint was Minneapolis, where DHS's "Operation Metro Surge" deployed thousands of federal agents beginning in January 2026, heavily affecting the city's Somali and Southeast Asian communities. Two U.S. citizens were killed by federal agents during the operation: Renée Good, a 37-year-old woman shot by an ICE agent on January 7, and Alex Pretti, a 37-year-old ICU nurse shot during a protest on January 24, each documented in a separate entry in this tracker. Federal officials and the president defended the shootings as self-defense; witness videos and local officials disputed those accounts. UN human rights experts warned the killings may amount to extrajudicial killings and called for independent investigation. A federal judge found that ICE had violated at least 96 court orders in Minnesota since the start of 2026, and the operation cost Minneapolis more than $200 million in a single month, with schools shifting to remote learning. More than 60 Minnesota CEOs, including the heads of 3M, Cargill, Target, Best Buy, and UnitedHealth Group, signed an open letter calling for de-escalation. Border czar Tom Homan announced the operation would wind down in February, and the president said afterward he wanted a "softer touch" following the citizen deaths. The pattern outlasted Minneapolis: on July 7, 2026, an ICE agent in Houston fatally shot Lorenzo Salgado Araujo, a Mexican construction worker who was not the operation’s target, again on a disputed vehicle-ramming self-defense claim for which no video has emerged; by then The New York Times counted more than 20 people shot at by immigration agents since September 2025, nearly all of them in their cars.

October 2026: U.S. Citizens Detained by Immigration Agents

The federal government does not count how often immigration agents detain U.S. citizens. ProPublica keeps its own count. On October 6, 2026, it published a database of 506 citizens whom immigration agents held against their will from January 2025 through September 2026. Its October 2025 count had been more than 170. ProPublica counted anyone who was not free to leave, whether handcuffed for minutes or jailed for days. It drew on social media, DHS statements, court records, local news and its own reporting, and it required two sources for each incident. The count covers only warrantless detentions away from borders and airports. It includes only people ProPublica could name, so ProPublica calls the total an undercount. At one Idaho raid in fall 2025, witnesses estimated that agents detained more than 200 citizens, but ProPublica could name only 30.

ProPublica's main findings:

  • More than 100 citizens were held after agents questioned their citizenship. Nearly all were people of color.
  • More than 300 were detained after agents accused them of assaulting or obstructing officers. About 130 of them were at protests.
  • ProPublica found 137 federal criminal cases. Nearly half were dropped or dismissed. About three dozen ended in prison or probation, and about three dozen were pending as of September 15, 2026.
  • At least 70 were children or teens. About 20 children were handcuffed or zip-tied.
  • 36 were held for at least a day without being able to reach a lawyer or family.
  • About a dozen were deported. Most were children removed with their parents.
  • More than 120 were detained in Illinois during Operation Midway Blitz, the fall 2025 Chicago operation.

A detention in the database is not a finding that the detention was illegal. Agents may briefly hold a citizen they reasonably suspect is in the country illegally. They may also arrest a citizen accused of interfering with or assaulting officers. ProPublica reported that some protesters did assault officers.

ProPublica described cases such as these:

  • Brian Morales Garcia (Texas, spring 2026). Officers stopped the car he was in and did not believe he was a citizen. He was held for four days without a phone call and deported to Mexico. After he sued, he was allowed back into the country in September 2026. In court filings, officials had disputed his citizenship (ProPublica).
  • Jeyli Salguero (Oakdale, Minnesota, January 14, 2026). Agents entered her family's home without a judge's warrant while looking for her brother, who did not live there. Salguero was briefly held (ProPublica).
  • Three siblings (Brownsville, Texas, July 23, 2025). Agents detained three citizens, including a 14-year-old, after they stopped to film an immigration arrest (ProPublica).
  • George Retes (California, 2025). Agents detained the Army combat veteran during a raid on the farm where he worked as a security guard. He was held for three days without a phone call and released without charges (ProPublica).

Administration officials have repeatedly said agents do not detain citizens. DHS spokesperson Tricia McLaughlin told ProPublica in October 2025, "We don't arrest US citizens for immigration enforcement." Then-Secretary Kristi Noem said in late October 2025 that "no American citizens have been arrested or detained." That claim is false. An NPR fact check on November 5, 2025 found it "is not true," citing Retes and other documented cases. NPR also reported that DHS had acknowledged in other statements that it had held citizens. During Operation Midway Blitz, DHS said there were "no administrative arrests of U.S. citizens." ProPublica later counted more than 120 citizens detained in Illinois during that operation. In statements quoted by ProPublica in October 2026, DHS said "ICE does not zip tie or restrain children." ProPublica's records include about 20 children who were handcuffed or zip-tied. Asked about the findings, DHS said its operations are "highly targeted" and that anyone who obstructs agents will "face consequences which include arrest."

Congress and the courts have also addressed the issue. On December 9, 2025, Senate Democrats led by Sen. Richard Blumenthal (D-Conn.) released an investigation based on interviews with nearly two dozen detained citizens, ProPublica reported. The citizens described being dragged from cars, held for days and denied medical care. ICE told Congress that 131 citizens were arrested for assaulting, obstructing or interfering with officers during Operation Metro Surge in Minnesota, according to ProPublica. ProPublica could identify fewer than 100 of them by name. In fall 2025, the Supreme Court issued a temporary order letting agents in the Los Angeles area make stops in which "apparent ethnicity" may be a factor. Justice Brett Kavanaugh wrote in a concurrence that agents "promptly let the individual go" once they learn the person is a citizen, according to ProPublica. DHS has cited that opinion to defend its stops. Salguero is a plaintiff in a lawsuit over warrantless home entries. DHS has moved to dismiss it, saying it has "paused" the policy (ProPublica).

As of October 2026, the federal government still does not publish a count of citizens detained by immigration agents.

Dismantling of Detention Oversight

In early May 2026, DHS shut down the Office of the Immigration Detention Ombudsman, the internal watchdog charged with investigating misconduct, abuse, and deaths inside immigration detention. HuffPost first reported that the office was removing its public signage, ending inspections, and taking its complaint website offline, and DHS confirmed the closure days later. Congress created the office in 2019 to operate independently of ICE and CBP, giving it authority to receive detainee complaints, inspect facilities, and report systemic problems to department leadership and Congress.

DHS attributed the closure to a funding lapse. In a statement, it said "DHS did not shut down the Office of Immigration Detention Ombudsman—Congress did," pointing to the appropriations bill that had just ended the longest government shutdown in U.S. history. That bill funded most of the department but was silent on the office and did not require its closure. Immigration attorneys and Democratic lawmakers argued the executive branch cannot unilaterally eliminate a body established in statute, and the Robert F. Kennedy Human Rights organization is litigating to block the shutdown. Adam Isaacson of the Washington Office on Latin America, who told HuffPost he believed the closure was unlawful, described the broader erosion of oversight as "death by a thousand cuts."

The shutdown was the culmination of a longer drawdown. On March 21, 2025, the executive branch placed the staff of the detention ombudsman and two sister oversight bodies, the Office for Civil Rights and Civil Liberties and the Citizenship and Immigration Services Ombudsman, on administrative leave and moved to dissolve them. A DHS spokesperson said at the time the offices had "obstructed immigration enforcement by adding bureaucratic hurdles" and that the goal was to remove roadblocks to enforcement. A March 2026 report by the Washington Office on Latin America and the Kino Border Initiative found the ombudsman's staff had been cut by roughly 96%, from more than 100 to five, leaving those five to cover 220 active detention facilities, while congressional appropriators drafted a bill that would zero out the office's $28.6 million budget. The closure came as deaths in custody were running at their highest level in two decades, which critics said made independent monitoring more necessary rather than less. DHS has said the rising toll reflects the larger number of people held rather than worsening conditions.

The Detention Build-Out, a Procurement Scandal, and a $700 Million Retreat

The campaign was financed by a roughly $45 billion detention expansion in the 2025 reconciliation law, and the rush to spend it produced its own controversy. Under then-Secretary Kristi Noem, ICE pursued a "Detention Reengineering Initiative" envisioned at up to $38 billion to convert industrial warehouses into facilities holding more than 92,000 detainees—a plan that would also have shifted ownership of detention infrastructure from contractors to the federal government itself. ICE bought eleven vacant warehouses for roughly $1 billion before Noem was ousted.

Independent reviews found the government consistently overpaid. A CoStar analysis cited by the Washington Post found DHS paid an average of 13% above market. The individual deals were more extreme: ICE paid $123 million in Socorro, Texas, for a property last valued near $11 million, and $68 million in Flowery Branch, Georgia, for a site previously valued at $400,000; a single facility in Hamburg, Pennsylvania cost $87 million. Many warehouses lacked proper zoning and working plumbing, and most were purchased quickly and largely without competitive bidding. The DHS Office of Inspector General opened an audit into whether the purchases were cost-effective, layered atop a separate corruption inquiry into Noem's top adviser Corey Lewandowski. According to NBC News, Lewandowski allegedly sought multimillion-dollar payments from contractors. Lewandowski has denied the allegations. A George Washington University procurement-law expert called the contracting pattern "a blazing red flag," and even GEO Group, one of ICE's largest detention operators, had warned investors that warehouse conversions would be operationally fraught.

By mid-2026 the initiative had collapsed into a costly retreat. The New York Times reported on June 18, based on internal documents, that ICE plans to offload seven of the eleven warehouses—purchased for more than $700 million—by transferring them to other federal agencies or selling them outright, a reversal of Noem's signature initiative under her successor Markwayne Mullin, who had privately expressed skepticism of the plan. The seven are in Romulus, Michigan; Social Circle and Flowery Branch, Georgia; Hamburg and Tremont, Pennsylvania; Salt Lake City; and Roxbury, New Jersey. Most were never converted or used to hold a single detainee. The agency appears to be proceeding with four sites, in San Antonio and Socorro, Texas; Surprise, Arizona; and Hagerstown, Maryland—though a federal judge has blocked all work at the roughly $100 million Maryland facility, and ICE told federal courts in New Jersey and Michigan it would halt work pending environmental review. The plan foundered on environmental lawsuits over skipped federal impact reviews, the inspector general's investigation, and local opposition that included conservative communities and Republican politicians worried about utilities, the economy, and protests. Even a former Trump administration immigration official, John Fabbricatore, conceded the warehouses' "scale and footprint" made them an easy target, while Claire Trickler-McNulty, a senior ICE official in the Biden administration, told the Times the plan "seemed questionable from inception," kept alive only by "the endless blank check ICE has for detention." DHS said in a statement that it is "moving swiftly to utilize EXISTING detention space with our state and county partners," and ICE now plans instead to buy detention facilities from the private prison companies it already contracts with. Days earlier, Mullin had acknowledged "some due diligence that maybe wasn't actually checked off." The retreat, covered in follow-ups by the AP and NBC News, also exposes a gap in the deportation machine itself: border czar Tom Homan had told the Times ICE hoped to have 100,000 detention beds by the end of 2025, but the agency topped out at holding around 70,000 people, and any resale of the overpaid warehouses will almost certainly come at a loss to taxpayers.

The disposal moved forward in the fall. On September 24, 2026, the Government Accountability Office reported that ICE had pursued the warehouse purchases and its other detention-expansion initiatives "without necessary analysis and planning," and that it had already spent more than $20 million on costs it cannot recover, such as security, on the warehouses it intends to sell. On September 30, Bisnow reported that the government had hired the brokerage Cushman & Wakefield to sell seven of the warehouses, for which it paid $706.5 million. According to Real Estate Alert, bids were expected to value them at about $600 million. The listing, called the "Federal Warehouse Portfolio," markets the buildings to logistics companies as vacant distribution centers and does not mention ICE, according to Project Salt Box, a watchdog group that tracks detention contracts. The list had changed since June. It now includes the Socorro, Texas, property, which ICE once planned as an 8,500-bed detention center. Meanwhile, DHS reversed its decision to sell the Roxbury, New Jersey, warehouse, which it bought for $129.3 million, and told a court it may convert that site to detention after all. An asking price is not a completed sale, but any sale below the purchase price would add to the losses the GAO identified.

Roxbury itself became a point of dispute. After New Jersey sued, DHS agreed or was ordered to halt converting the warehouse into a detention facility pending an environmental review. On October 8, NPR reported that neighbors, a Dover council member and a man later deported to Honduras said ICE had taken people arrested in a September 4 raid in nearby Dover to the site. The man said he and nine other men were held in handcuffs in the parking lot for several hours. On October 13, a DHS official confirmed to the office of Rep. Analilia Mejia (D-N.J.) that ICE officers had used the warehouse parking lot to initially process 14 detainees on two occasions in September. DHS said it did so to protect its officers and that no detainees were taken inside the building. Mejia said the use was "potentially in violation of a court order." The same NPR report quoted an anonymous official at a government contractor who said his company had proposed more humane, FEMA-style camps in Texas at a meeting with a DHS deputy director, and was told the idea was not what White House adviser Stephen Miller wanted. Miller was not at the meeting. The official said Miller "did not want facilities that were humane." White House spokesperson Lauren Bis called the account a "false hearsay allegation" from someone who had never met or talked with Miller, and said Miller "has never made any remarks to anyone about wanting illegal aliens to have worse detention facility conditions."

Public Opinion Has Shifted

As the campaign's interior tactics became more visible, approval eroded. An NPR/PBS News/Marist poll in late January 2026 found 65% of Americans said ICE had gone too far, up from 54% the previous June, and 62% said ICE's actions were making Americans less safe. PRRI found only 35% approved of the president's handling of immigration by early 2026, down from 48% a year earlier, and that support for separating parents from children had fallen to around 11%. Supporters note a partial rebound: a late-April Harvard/Harris poll showed a modest uptick in approval as the Trump administration toned down its enforcement rhetoric after Minneapolis, with ICE's favorability ticking up to 40%. The throughline across pollsters is that Americans broadly back removing serious criminals but recoil from indiscriminate raids, family separations, and the targeting of long-settled, law-abiding residents.

The Bottom Line

Judged against its own two-part promise, the campaign delivered on border security and failed on jobs. Crossings fell to historic lows, but the best national evidence—now two independent studies using different data and methods—shows that interior deportations reduced, rather than expanded, employment for the U.S.-born workers the policy was meant to help, with no wage gains to show for it and losses reaching an estimated 668,000 jobs across the hardest-hit cities, a majority of them in red states. That economic backfire sits alongside an estimated 145,000 U.S.-citizen children separated from a parent, a two-decade high in deaths in custody, a pattern of fatal shootings by immigration agents that killed two U.S. citizens in Minneapolis and, months later, an unarmed worker who was not the target in Houston, at least 506 U.S. citizens detained by immigration agents, repeated findings of indiscriminate arrests of people with no criminal record, a nationwide injunction against courthouse arrests, a shadow network of third-country removal deals paid for in cash and facilitated by a UN agency, allegations that asylum seekers' confidential files were handed to the government they fled, and a $1 billion detention-buying spree now under federal audit, with seven overpaid warehouses headed for disposal at a likely loss. The headline GDP damage has been contained, mainly because removals never reached the scale earlier forecasts assumed—and what job growth the economy produced flowed overwhelmingly to female-dominated service sectors, not the male-dominated industries the policy claimed to protect. But the institutional and human costs are large: broken families, eroded trust in communities, deaths, wasted infrastructure, and a procurement scandal. The administration's newer claim that deportations drove the drop in violent crime is not supported by the crime data, which show some of the steepest declines in sanctuary jurisdictions. These are the kind of costs that are slow and expensive to reverse, which places this among the most consequential and durable actions of the term.