Memo: A Latino Reset

A Latino Reset

By: Carlos Odio and Maria Di Franco Quiñonez

‍

The very early polling on a Harris-Trump election suggests a reset in the fight over Latino voters. The Hispanic electorate now looks more in line with the other Trump-era elections of 2020 and 2022 than with a wide partisan realignment.

As of a few weeks ago, a wide range of scenarios appeared plausible, including one in which Republicans built significantly on 2020 gains among Latinos. Now, early signs are that the Kamala Harris campaign is on track to hold off a GOP surge.

In our polling this cycle, as in public polls, we saw results that occasionally made us question what was actual movement in the electorate, what were merely expressions of discontent, and what were polling errors immune to adjustment. ("Can Donald Trump win a majority of Latinos under 30?" was one such hall of mirrors. Needless to say, that scenario can now be ruled out with a high degree of confidence.)

This uncertainty in measurement reflected the uncertainty in the real world: Latino voters were feeling beat down by a parade of crises. Their top issue priorities reflected insecurity that was both physical (gun violence, crime) and economic (rising costs, including housing and healthcare). And key voters distrusted both candidates for different reasons. The Biden-Trump rematch was scrambling signals and loosening traditional group norms.

With the entrance of Kamala Harris, we are seeing results that are back in a historically normal range. Relative to Biden, she sees rebounds across Latino subgroups, with the largest among young people. As one swing Latina woman put it in a recent Pennsylvania focus group, Harris is “a light at the end of the tunnel” for many who were torn about their vote choice.  

In short: Democrats are back in business.

At the same time, Harris still has more work to do: our polling suggests she is tracking short of Biden’s 2020 support levels among Latinos by a few points. Her consolidation of discontented Latinos has been quick and dramatic, and our research suggests she has room for more movement. But neither side can afford to take these voters for granted over the last 80 days.

The kind of Latino voters who could tip the outcome in not only Arizona and Nevada but also Wisconsin and Pennsylvania do not closely follow election news. They don't have hard partisan identities. They are experts in their lives but not in politics. And if history is any guide, many of them won’t make up their minds — about whether to vote, or who to vote for — until closer to the election.

In important ways, this race has just begun.Full toplines are available here.

Polling Takeaways

Here’s what we are seeing in the polling:

1. Where does Harris start with Latinos?  

An Equis battleground poll fielded July 22 - August 4th shows Kamala Harris leading Donald Trump 56-37 (+19) among registered Hispanic voters in the seven most competitive states (Arizona, Georgia, North Carolina, Nevada, Michigan, Pennsylvania, and Wisconsin).

In a previous wave of polling, fielded May 16 - June 6, before the fateful Trump/Biden debate, Biden led by a much slimmer margin, 46-41 (+5).

2. Which Latinos are moving?

Latinos are not a monolith, as we have been known to say, but Harris’s gains cut across most subgroups. Her most notable pick-up is among younger Latinos.

Harris' support among Latinos under 40 is 17 points higher than Biden's was. She is at 60% with the young Hispanics, compared to Biden's 43%.

Her 59% support among Latina women is up from 50% for Biden in early June. Among men, she saw a similarly sized bump, moving her from 41% to 51%.

• Is it just liberal Democrats coming home? Seemingly not. While liberals have moved the most, with Harris doing 16 points better with them than Biden was, she is also doing 12 points better among moderate Latinos and 7 points better among conservative Latinos.

• Another way to understand the movement: about a third of Harris' gains come from Latinos who have an unfavorable opinion of both Trump and Biden (the so-called "double haters"). Harris has the support of 65% of double-haters. Only 11% would vote for Trump. The other 1-in-4 are still up for grabs, and/or inclined to abstain.

• Is the movement different by state? Some variation is evident, but it is relatively small: Harris does 9 points better in Pennsylvania than Biden was doing, 10 points better in Arizona, and 12 points in Nevada.  

• Note: while the analysis in this poll focuses on the most competitive states in the presidential battleground, the full poll also included Colorado, Florida, New Mexico, Ohio and Texas. The trends discussed here consistently appear in these states as well. Case in point: Harris is doing 9 points better than Biden was in Florida and 13 points better in Texas. (We don't have enough sample size across both polls to break out the other states.)

3. Could this movement be more polling malarkey?

Could these shifts in polling be driven by changes in who is responding to polls, in a way unaddressable by our weighting scheme? Fortunately, as a check, we are able to compare support levels for Biden and Harris among the same individuals, by asking respondents from this most recent poll who they were thinking of voting for before the news of the president dropping out.

The exact same group of respondents who support Harris by a 19 point margin say they would have previously supported Biden by only 8 points (46-38).

• Notably, this in-poll movement reveals that Harris pulls nearly as much from the Trump column as she does from the previously uncommitted. This early data appears to reflect more than just a boost in enthusiasm. Harris is winning over some voters who had been supporting her opponent.

4. Who is planning to vote?

Back in June, Republican Latinos were more likely to say they would vote this year. With Harris as nominee, Democrats seem to be closing the gap with Republicans on self-reported vote intent.

In the latest poll, 78% of Democratic Latinos said they were "almost certain" to turn out – up from 73% in June. By comparison 79% of Republican Latinos said they were certain to vote (77% in previous poll).

5. Where is Harris relative to 2020?

Support for Kamala Harris among Latinos across the battleground states remains a few points shy of what Biden received in 2020, with some variation across states.

Adjusted to two-way vote, with undecideds allocated proportionately, Trump is 4 points ahead of where he ended up in 2020 in these toss-up states. (The allocated two-way vote today is Harris 60 - Trump 40. The Latinos in this poll who voted in 2020 say they went Biden 64 - Trump 36 in that election.)

The story varies by state. As a key example, Harris support in Wisconsin essentially matches estimates from 2020. Meanwhile, she is about 5 points shy of 2020 support levels in Nevada, one of the states where Biden as nominee had been slipping the most.

• Harris, of course, does not need to match Biden's 2020 Latino performance everywhere in order to win. Nevada is a state where our scenarios suggest she has some wiggle room, if non-Latino support levels remain relatively stable. Arizona, a state that Biden won by a fraction of a percentage point, is more of a tightrope: Harris would likely need Latino support in the same range as in 2020. She is currently a mere 2 points shy.

6. What other unknowns lurk in the polling?

• Third party. When offered as an option, some 9% of Latinos say they would support Robert F. Kennedy, Jr.. This is the lowest recorded support for Kennedy in Equis surveys since we started polling him by name. But his inclusion on the ballot remains another source of unpredictability.

• Swinginess. The number of undecideds dropped with the introduction of Harris as the Democratic nominee. And yet, by our measure, some 15% of registered Latinos still fall into a persuadable category — either their 2024 vote doesn’t match their 2020 vote (switchers), their presidential vote doesn’t match their down-ballot preferences (splitters), or they remain uncommitted. That leaves a great deal of possibility.

• New voters. Based on previous election cycles, it is reasonable to expect that some 30-40% of Latinos who vote in 2024 won't have voted in the last presidential election. Uncertainty is highest among this segment: not just about whether they'll vote but who they will vote for. Some of the least-partisan, most persuadable Latinos are "low-propensity" voters.

• Among the newly registered, Harris is leading 51-35 (+16), up from Biden's 47-41 (+6).

• Meanwhile, a majority (56%) of Latinos who didn't vote in 2020, which includes new registrants, are saying they are certain to vote this year. That is a 10 point jump in vote intent from the prior survey.
  ‍
• Among those who were registered to vote in 2020 but abstained, Harris has pulled even: 47-45 (+2). Biden had been losing to Trump among these 2020 non-voters back in June, 32-50 (-18). That is a 20 point swing, in margin, among 2020 non-voters.

• Special mention: Trump. This memo focuses on Harris because it is her arrival that has crystallized a view of the Latino electorate in 2024, and her early definition is likely to have the largest effect on public opinion. "Voters are more persuadable with information about candidates they know less about" (Broockman & Kalla). But we cannot write a list of possible curveballs without mentioning the most erratic element. If new information persuades, Trump and his allies manage to produce a lot of it – whether detailed plans for his second term or, less substantively, the return of his unfiltered thoughts via social media.

7. In sum:

An early look at the Harris-Trump election suggests a reset in the fight over Latino voters. The strong start from Harris appears to have forestalled some of the most outlandish scenarios for the Latino vote. Out the gate, the vice president has quickly amassed the support of a wide swath of discontented Hispanic voters, and she still has running room. What those last Latino voters do could determine the overall result in hotly contested states.

About the Polling

Equis is a set of organizations working to create a better understanding of Latinos, innovate new approaches to reach and engage them, and invest in the leadership and infrastructure for long-term change and increased engagement.

The data in this memo comes from the Equis State Series, ongoing polling of Latinos in 12 states, conducted by Equis Research in partnership with Searchlight Research. Other state-specific polls were conducted with GBAO, Normington Petts, and TargetSmart. The Equis State Series has interviewed some 16,700 Hispanic voters since July 2023. Additional insights come from extensive message testing as part of the Equis Playbook. Past vote estimates are from Catalist.

The memo specifically relies on movement between two Equis polls:

a July 22 - August 4th survey of 2,183 registered voters who identify as Hispanic or Latino, fielded in 12 states (AZ, CO, FL, GA, MI, NV, NM, NC, OH, PA, TX, and WI). Of those respondents, 1,242 were registered in the 7 battleground states (AZ, GA, NC, NV, MI, PA, WI).  

a May 16 - June 6 survey of 2,600 registered voters fielded in the same states, with 1,477 in the 7 battleground states.

Except where otherwise noted, the numbers reported in this memo are based on data from the 7 battleground states.  

Both surveys were conducted via live phone and text-to-web. The samples were weighted to match the composition of registered voters modeled as Latino by state, gender, age, education, and gender by modeled vote choice (Catalist's Vote Choice Index) at the battleground level across the 12 states. For states with at least 200 respondents (AZ, FL, NV, TX, PA, and WI for the August survey), the sample was also weighted by gender, age, education, 2020 vote recall, and either party affiliation on file (where available) or modeled vote choice to match the composition of registered voters modeled as Latino in each state.

Language of interview, country of origin, education and other variables were monitored during the fielding. The survey was conducted in English and Spanish, according to the respondent’s preference.

MoE for the August survey is ~2.9% for the full sample (12 states) and ~3.7% for the seven battleground states.

MoE for the June survey is ~2.6% for the full sample (12 states) and ~3.2% for the seven battleground states.