latino-voters-trump-republicans-midterms

Buyer's remorse hits Trump's Latino voters

Russell Contreras

Data: UnidosUS/BSP Research/Shaw & Company; Chart: Sara Wise/Axios

A new UnidosUS poll finds that 1 in 4 Hispanic Trump voters say they would not vote for him again if given the choice.

Why it matters: The erosion of Latino support for President Trump, combined with dissatisfaction with the economy, signals danger for competitive GOP-held seats in the 2026 midterm elections.

Trump's gains with Latinos were one of the biggest political stories of 2024.
Now, a new UnidosUS/BSP Research/Shaw & Co. poll suggests those voters remain highly movable — a problem for Republicans in Latino-heavy battleground districts where both parties are watching for signs of a post-2024 snapback.

By the numbers: Two-thirds of Latino voters disapprove of Trump's job performance, compared with 30% who approve. Disapproval is a majority in every region tested, including 51% in Florida.

25% of Latino Trump 2024 voters say they probably or certainly would not vote for him again. That compares with 5% of Harris 2024 voters who say they would not vote for her again.
On the 2026 House ballot, Latino voters back the Democratic candidate over the Republican candidate 54% to 27%, with 19% undecided.

State of play: Latino voters are emerging as "the swingiest of the swing voters" ahead of the midterms, with economic anxiety and immigration concerns making them highly fluid in competitive districts.

The GOP had hoped Trump's 2024 Latino gains represented a lasting realignment.
Democrats see inflation, tariffs and aggressive immigration enforcement reopening Latino-heavy seats they feared were lost.

Zoom in: The political damage to Republicans is highly apparent in critical battlefield states containing high-density Latino populations, the UnidosUS poll found.

In Texas, where significant rightward shifts occurred in 2024, Trump's disapproval has skyrocketed to 67% among Latinos, and Democrats have opened a massive 54% to 28% generic House ballot lead.
In California, Latino disapproval of Trump stands at 71%, paving the way for a 59% to 22% Democratic advantage on the generic congressional ballot.
In Arizona, Trump faces a 67% disapproval rating, with Democrats leading the generic House ballot 53% to 25%.

Zoom out: Voter dissatisfaction extends into deep anxieties regarding governance, military stability and civil rights.

84% of Hispanic voters are deeply concerned that Congress is failing its constitutional duties by ceding too much authority to the executive branch and failing to act as a proper check and balance.
58% believe that their civil rights and civil liberties have become noticeably less secure under the Trump administration.

79% assert the president should be legally required to obtain congressional approval before executing military action.
64% explicitly oppose the war in Iran.

The intrigue: Latino voters say cost of living, not immigration, is the top issue shaping their views of Trump.

Immigration enforcement ranks second, ahead of jobs and the economy, the Iran war, tariffs and cuts to Medicaid, ACA/Obamacare and food assistance programs.

The other side: Florida remains the glaring exception.

Latino voters there back the Republican House candidate over the Democrat 42% to 38%, and Florida GOP candidates also lead in Senate and governor matchups tested in the poll.

The bottom line: Trump's Latino gains are no longer looking like a locked-in GOP advantage.

The poll gives Democrats a clear opening in Latino-heavy battlegrounds — but not a guarantee.

Methodology: This UnidosUS Bipartisan Poll of Hispanic Voters: Road to the 2026 Midterms was conducted April 27–May 14 by BSP Research and Shaw & Company Research on behalf of UnidosUS. The poll is based on a sample of 3,000 Latino registered voters across the U.S.

The margin of sampling error is ±1.8 percentage points for results based on the entire sample.
The survey was offered in English or Spanish, based on respondent preference, and conducted via live telephone interviews, text message invitations and online panels. All participants were confirmed registered voters.