Deportations and affordability are both eroding Trump’s 2024 coalition.

Sixty-eight percent of Americans agree that the Trump administration is “focusing too much on deportations and not focusing enough on fixing the economy, inflation, and the rising cost of living” — one percentage point below the share who said the same in our February survey. Among the 2024 Trump voters who say they will not vote Republican in the 2026 House midterms, agreement reaches 69%; among his still-loyal voters it is 27%.

Defection from the 2024 coalition has not receded.

On a like-for-like basis, the share of Trump’s 2024 voters who would no longer vote Republican was 14% in February and 16% in May. The two-point difference is within the margin of sampling error, so the level is the story: roughly one in six of Trump’s own 2024 voters has left the fold, and the gap has not closed since February. Asked why their view of the Republican Party changed, defectors name the economy and cost of living first (52%), immigration second (38%, combining two questions), and Trump’s personal conduct third (35%). The immigration total splits into enforcement and deportation policy (23%) and the treatment of legal immigrants and U.S. citizens (15%).

Voters would redirect enforcement money to the cost of living.

Asked what the government should do with a hypothetical budget surplus resulting from reduced immigration-enforcement funding, Americans put reducing grocery and food prices first (29%) and lowering health care costs and protecting Medicaid second (24%). Twelve percent would keep the money with immigration enforcement and 10% would cut the budget instead of spending it. Hiring more local police drew 3% and border-security technology 6%.

Majorities call the everyday consequences of deportations going too far.

Fifty-three percent of U.S. adults say mass deportation of immigrant workers is hurting the economy and raising costs for American families, against 35% who say the deportations are necessary to protect American jobs and enforce the law. Judging concrete consequences of stepped-up enforcement, majorities call nearly every one “going too far”: higher grocery prices after farm and food workers are deported (60%), children disappearing from a classroom (57%), elderly people losing the home workers who care for them (55%), a coworker forced out after losing work authorization (55%), and restaurants closing for lack of staff (52%). Deporting immigrants who committed violent crimes is acceptable to 80% of Americans; nearly every other enforcement action tested reads as going too far to 53–68%.

A crime-and-border caveat does not raise support for enforcement.

A randomized experiment tested whether pairing pro-immigrant messaging with a crime-and-border concession — “we need secure borders” and support for deporting undocumented immigrants who have committed violent crimes — changes agreement. All three versions landed at 66–68% agreement, a spread within the margin of sampling error. If anything, the caveat reduced intensity: strong agreement fell from 48% with the argument alone to 42% and 40% when the caveat was added.

Working-class voters split from the nation on the economy and match it on immigration.

The survey included a 1,000-person oversample of adults with household incomes under $60,000. Working-class voters are markedly sourer on the economy than the nation as a whole — the working-class penalty on net approval of economic issues runs 4–7 points — but on immigration enforcement the gap is 1–2 points, and the same 68% share says the administration is too focused on deportations. Working-class Hispanic voters are sharply negative on the administration’s record and break Democratic on the 2026 generic ballot by roughly 20 points.

Selected measures · U.S. adults, May 2026

Measure	All	Working class	WC Hispanic
Administration too focused on deportations	68%	68%	71%
Economy is their most important voting issue	43%	45%	46%
Mass deportation is hurting the economy	53%	51%	62%
Net approval, handling of the economy	−34	−40	−54
Net approval, immigration enforcement	−13	−13	−44
Generic ballot, net Democratic margin	+9	+11	+22

Working class is defined as household income under $60,000 (unweighted n = 1,325). Working-class Hispanic respondents number 221, so those figures are directional. Net approval is the share who approve minus the share who disapprove.

Methodology

This survey was fielded May 18–20, 2026, with data collected by Verasight, a member of the AAPOR Transparency Initiative. The sample consists of 2,000 U.S. adults aged 18 and older: 1,000 in a nationally representative sample and 1,000 in an oversample of adults with a confirmed household income under $60,000. The margin of sampling error, which accounts for the design effect, is ±3.1 percentage points for the nationally representative sample and ±3.4 points for the low-income oversample. Data are weighted to the April 2026 Current Population Survey on age, race/ethnicity, sex, income (nationally representative sample only), education, region, and metropolitan status, as well as to a running three-year average of partisanship from the Pew Research Center NPORS benchmarking surveys and population benchmarks of 2024 vote; a combined weight integrates the two samples while holding low-income adults to their correct national share. February comparisons come from a separate national survey of 1,500 U.S. adults fielded February 6–10, 2026. Full question wording is in the topline files above.