Democrats’ losses among Hispanic voters may be larger than they look
Conservative Hispanic Americans are ditching their Hispanic identity, surveys show.
CAROLINE SOLER AND BRIAN SCHAFFNER
One storyline of the 2024 election was evidence of a clear shift toward Donald Trump among Hispanic voters. In 2020, Joe Biden won about 63% of the vote among Hispanic voters. But by 2024, Kamala Harris’ support among this group dropped to just 55%. Cooperative Election Study data reveal that Democrats have lost support among Hispanic voters in every election since 2016. This drop has been especially pronounced among moderate and conservative Hispanics.
But there may be more to this story. What if some of the most conservative Hispanics aren’t just voting more Republican – they’re actually no longer even identifying as Hispanic? If that’s the case, then these polls may be understating the inroads Republicans are making among Hispanic voters.
Political scientist Patrick Egan has demonstrated that a small but significant share of Americans change their racial, ethnic, and other identities over time. These changes sometimes appear to be motivated by political attitudes – for example, liberal Democrats were more likely to shift into identifying as Latino, gay, or nonreligious, while conservative Republicans were more likely to adopt Protestant or born-again Christian identities. Our research finds that Hispanic voters were not only voting more Republican in 2024 than they were in 2020. In fact, some voters who identified as Hispanic in 2020 were no longer adopting that identity at all in 2024.
Hispanic identity is fluid
In 2024, the Cooperative Election Study re-interviewed more than 6,000 Americans whom we first polled during the 2020 campaign. On both occasions, we asked respondents whether they identified as Hispanic or Latino. The survey gives respondents two chances to identify as Hispanic: They can answer “Hispanic” to a question asking about their race; or they can do so through a follow-up question asking “Are you of Hispanic, Latino, or Spanish origin?” for those who didn’t select Hispanic as their race.
Our data reveals that 6% of people who identified as Hispanic in 2020 no longer did so by 2024, completely dropping the identity when answering either survey question. On the other side of the ledger, just 0.2% of respondents who did not identify as Hispanic in 2020 started identifying as such in 2024. That means we observed a net decrease in the size of the Hispanic group in our sample from 2020 to 2024. The Hispanic share of our survey sample decreased from 9.1% to 8.7% during those four years, solely due to changes in how people identified themselves.
One noteworthy point about those deidentifying as Hispanic is that most of this group only identified as Hispanic in the secondary question in 2020. This suggests that those who identify as Hispanic in the race question have a stronger attachment to that identity compared to those who only identify as Hispanic in the follow-up. Given this pattern, it is also worth highlighting that 16.5% of 2020 Hispanics showed signs of a weakening Hispanic identity – moving from selecting “Hispanic” as their primary racial identity in 2020 to acknowledging their Hispanic heritage only in the follow-up question by 2024.
Are Hispanic Americans less attached to their pan-ethnic identity?
To put this in context, Black Americans show much lower deidentification rates – just 2.1% of those who identified as Black in 2020 did not do so in 2024. In sum, the Hispanic deidentification rate was roughly triple that rate. That’s a striking finding, considering that Hispanic respondents actually have two different ways to identify in our survey.
This suggests that many Hispanic people may be less attached to their pan-ethnic identity than other groups – they are much less likely to feel that what happens to Hispanics as a group will affect their own lives, in contrast to how Blacks, whites, or Asian Americans feel about their respective groups. For those who do drop their Hispanic identity, two-thirds move to identifying as white. This shift toward white identification reflects the underlying flexibility that makes Hispanic deidentification possible in the first place. When you have mixed heritage or when your identity feels less central to your life, dropping it becomes more psychologically feasible.
Are these identity changes related to politics?
The political connection to these identity shifts become clear when we examine voting patterns and candidate preferences in 2024. To capture the complete picture of political support, we analyzed both actual vote choice (for those who reported voting) and stated candidate preference (for those who didn’t vote).
Among those who maintained a strong Hispanic identity – selecting “Hispanic” as their race in both 2020 and 2024 – Trump had a slight disadvantage, trailing by 1.8 percentage points. Those who weakened but didn’t completely abandon their Hispanic identity showed similar patterns, with Trump trailing by about 1 percentage point among those who moved from primary to secondary Hispanic identification.
However, the pattern shifts dramatically among those who have completely abandoned their Hispanic identity – as shown in the top bar of the figure. Voters who went from identifying as Hispanic in 2020 to not identifying as Hispanic at all in 2024 showed overwhelming support for Trump, giving him a 30-point advantage. This massive gap suggests that completely stepping away from Hispanic identity is strongly associated with conservative political views.
The ideological patterns behind these identity shifts support this political narrative. Hispanic respondents who said that they were conservative in 2020 showed a much higher deidentification rate (6%) compared to Hispanics who said that they were liberal (4%). Although moderates comprised the largest share of “deidentifiers” overall, conservatives were disproportionately represented among those who completely abandoned their Hispanic identity.
The overall trend is clear: The most Republican or conservative-leaning Hispanic Americans appear to be gradually removing themselves from the Hispanic category altogether. And voters who maintain any form of Hispanic identification lean more Democratic or liberal.
Not just a Trump effect
But this pattern isn’t just about Trump. We see remarkably similar patterns in our 2010-2014 panel study. About 6.7% of Hispanics stopped identifying as such during that period – nearly identical to the 6% rate we observed from 2020 to 2024.
The voting patterns across identity trajectories were even more pronounced in the earlier period. In 2012, Sen. Mitt Romney (Utah), the Republican presidential nominee, received 59.9% support among those who completely abandoned their Hispanic identity, compared to 45.2% among those who maintained a strong Hispanic identity. That’s a 14.7-point gap.
This suggests we’re looking at something more fundamental than Trump-driven polarization – for more than a decade, conservative and moderate political identity has been associated with weaker attachment to Hispanic ethnic identity. This pattern was even stronger during the Obama era than during these recent presidential elections when Trump was on the ballot.
What does this mean for understanding Hispanic voters?
These findings suggest that polls may be systematically undercounting Republican support among Hispanic voters, but not because pollsters are missing people. Instead, it’s because some moderate and conservative Hispanics are gradually removing themselves from the Hispanic category altogether. When we sample “Hispanic voters” during a given election cycle, we are likely missing some conservative and moderate voters who previously would have identified as Hispanic but no longer do so.
The fact that this pattern has persisted across different presidential administrations – through Obama’s presidency, Trump’s anti-immigrant rhetoric, and into the Biden era – suggests we’re witnessing something structural about how some Hispanic Americans navigate questions of identity, assimilation, and political belonging.
This raises important questions about how we interpret demographic trends in American politics. The story of Hispanic voters is not longer just about how this group votes – it’s about who gets counted as Hispanic in the first place.
Caroline Soler is a recent graduate of Tufts University, majoring in political science and mathematics, and is currently a research associate for the Cooperative Election Study.
Brian Schaffner is the Newhouse Professor of Civic Studies in the Department of Political Science and Tisch College at Tufts University. He also serves as a co-director for the Cooperative Election Study.