The Democrats hemorrhaged Hispanic votes in 2020. Initial hot takes tended to elevate shifts among HIspanic men above shifts among Hispanic women (Hispanic men liked Trump's macho persona, etc). This turns out to have been completely wrong. Pretty much all the most serious data analyses of the Hispanic vote in 2020 vote have found that the shift among Hispanic women was much larger than the shift among Hispanic men. Eric Levitz on the New York magazine site reports on one of these analyses, a slide show released by Equis Research. Levitz does a nice job of summarizing the more interesting findings from the slides and putting them in context, one of which is shown below, with shifts among Hispanic women far outpacing those among Hispanic men.
The Equis data also confirm the now widely-understood story that the pro-Trump shifts weren't confined to south Florida and Texas, but instead were all over the country.
The debate continues on the precise breakdown of pro-Trump shifts between defections and new voters who showed up and voted for Trump. As it does on the relative roles of pro-Trump views on the economy and the activation of social conservatism by increasing Democratic liberalism on race and other issues.
The Equis people seem determined to preserve immigration as an uncomplicatedly positive issue for Democrats among these voters, suggesting the only problem was lack of salience in 2020. I'm not so sure about this.
Kevin Drum suggests a different interpretation of how the immigration issue has evolved among Hispanic voters (see chart below):
"Roughly speaking, I think Trump suckered Democrats into becoming extremists on border policy. It's fine for Democrats to oppose the wall—most Hispanics oppose it too—but progressives, goaded by Trump, have staked out a position that's often only a finger's width away from not having any border security at all. And that's something that most Hispanics don't support.
To put it simply, Hispanics are like the rest of us: They care about other people, but they care about themselves more. When it comes to immigrants who are already in the country, legally or not, they're in favor of giving them a road to citizenship. But they're not especially keen on allowing lots of new immigrants in who will compete with them for jobs and housing. Trump may be a buffoon, but at least some Hispanics have decided that they can put up with that if the alternative is risking a big increase in the rate of illegal immigration.
Added to that, of course, is the well-known fact that many Hispanics are fundamentally conservative to begin with: Family oriented, religious, against abortion, in favor of low taxes, opposed to defunding the police, etc. It's been common knowledge forever that Republicans could win a big share of the Hispanic vote if they'd just moderate their base, but it turns out there's another way: Get the Democrats to radicalize their base instead. It seems to have worked."
Other interpretations of the immigration issue and the shifts are retailed in the Levitz piece. Read it and see what you think. I continue to find David Shor's interpretation most persuasive (also from the Levitz article):
"In an interview with Intelligencer last month, Democratic data scientist David Shor outlined one account of the 2020 results. His story goes like this: Democrats have long relied on the backing of Hispanic voters who self-identify as “conservative.” Such conservatives often have left-of-center views on economic policy, but lean right on issues of crime and policing. In 2020, Hispanic conservatives began voting more like white ones, and this was especially true of those who espoused the most right-wing views on “law and order” in policy polls. Therefore, Shor surmises that the heightened salience of policing issues following the George Floyd protests — and the Democratic Party’s association with “defund the police” in some media coverage of those protests — played a major role in driving Hispanic defections....
David Shor told me that he has observed the same trend in his own data. And he argues that the concentration of Trump’s gains among Latinas is consistent with his hypothesis that “defund the police” influenced Hispanic voting behavior since, in his polling, women rank crime as a more important issue than men do.....
Shor has long encouraged Democrats to do everything in their power to heighten the salience of the party’s most popular economic positions, while reducing the electoral relevance of immigration. This advice is grounded in a concern with the party’s steadily eroding support among non-college-educated white voters, who often lean left on matters of economics but right on those of borders. He often notes that in 2012, when Barack Obama and Mitt Romney framed their race as a referendum on whether the wealthy had “built” all they owned, the Democrat won 60 percent of the “pro–universal health care, anti-amnesty for the undocumented” vote. Four years later, when Hillary Clinton and Donald Trump both centered their messaging on conceptions of American identity — with the former championing multicultural diversity (“stronger together”), and the latter, white ethno-nationalism — the Democratic nominee’s support from pro-socialized medicine, anti-amnesty voters dropped by 19 points. Since Americans of this ideological bent are overrepresented in Rust Belt battlegrounds, that decline was sufficient to cost Clinton the presidency."
The story will continue to be debated as it should. But I think everyone can (or should) agree the Democrats have a very, very big problem here.