What Donald Trump can learn from George W. Bush about winning Hispanic voters
Bush's Hispanic surge in 2004 was further reaching and helped him win states Donald Trump lost
We’re now in the middle of one of those way-too-online primary season spats with Donald Trump trying to disqualify Ron DeSantis as a “Bush Republican.” The inciting event here is that Jeb Bush was a previous Florida governor who made an appearance at DeSantis’s inauguration and has said nice things about his successor. The missing context here is that before he was the purported avatar of a squishy establishment open-borders consensus, Jeb! had a reputation as the more conservative of the brothers, who probably lost his first gubernatorial bid in 1994 because he ran too far to the right. Had Jeb Bush emerged victorious in 1994, many believe that he, not George W. Bush, would have been the heir apparent for the 2000 presidential nomination.
The current spat is representative of all the ways that the term “Bush Republican” is now an epithet in the modern MAGA-ified GOP. Yet for all the political success Donald Trump has achieved—busting up the Blue Wall, coming within less than a point of re-election after being counted out for all of 2020—George W. Bush was the better vote-getter, actually winning re-election, which Trump failed to do, along the way becoming last Republican to win the popular vote.
Trump’s near-win in 2020 was significant because it showed the resilience of a Republican coalition based on working-class voters. (I’ve even got a book coming out that focuses heavily on this topic.) Trump hasn’t needed to be a popular politician in the classic sense because of the unique advantages such a coalition brings to the table in the Electoral College, with a 3.8 point pro-Republican bias in 2020.
A big part of Trump’s near-win in 2020 was his surprising strength among Hispanics—the same group he ranted about “bringing drugs, bringing crime” in his 2016 announcement speech. He surged in places like Miami-Dade County, giving him the advantage in Florida, and in the Rio Grande Valley in Texas.
But Trump’s showing, impressive though it was, was not the high-water mark for Republicans and the Hispanic vote in the modern era. George W. Bush earned an eye-popping 44 percent of the Hispanic vote according to the 2004 exit poll, topping Donald Trump’s 36 percent in 2020 in AP VoteCast survey, the more reliable of the two modern exit polls.
At some level, Trump’s Hispanic showing is being graded on a curve. Because Trump was supposed to be a uniquely awful politician for Hispanics, demonizing Mexicans and shutting down new Hispanic immigration, the fact that he did better than previous “kinder, gentler” Republicans is an eye-opener. Of course, this view relies on a fundamental misunderstanding of Hispanic voters, as Ruy Teixeira points out: that they’re activist “voters of color” who view immigration through same lens as Black voters view civil rights, instead of normie voters with middle of the road views on most issues.
When viewed on a longer time horizon, Trump’s 2020 performance seems more a bit less impressive and more of a return to the normal trendline—perhaps one that will presage Republicans performing above-trend for the next few cycles, perhaps not. George W. Bush likely did better than Trump among Hispanics in both of his elections. But Trump gets credit for lifting the party out of its 2008-16 trough, which seems to have been caused at least in part by immigration politics. The key ingredient was Trump downplaying immigration in 2020, while also elevating a brand of all-out culture war politics elsewhere that was more appealing to Hispanics than the buttoned-down establishment politics of John McCain or Mitt Romney. But, nonetheless, the all-time Republican champion of winning over Hispanic voters remains Trump’s party nemesis: George W. Bush.
Looking Back on Bush’s Hispanic Vote Performance
When George W. Bush was making the case that he could be the Republicans’ savior after two terms of being thoroughly outmatched politically by Bill Clinton, he pointed to his success with Hispanic voters in Texas, whom he split evenly as part of a rousing 1998 re-election victory with 69 percent of the vote. In a party less diverse than it is today, “Viva Bush” signs sprouted up at Republican rallies. Bush made a point of speaking Spanish and famously said that “family values don’t stop at the Rio Grande,” a striking reminder of a time when advocates of a softer immigration politics were not shouted down in Republican primaries. His first event upon arriving at the 2000 Republican convention was aimed squarely at Hispanics. An ad made for Bush’s 2004 campaign—copied directly from an ad Jeb Bush had run in Florida—showed a sequence of flags for various Latin American countries blended into an American flag. I couldn’t find the ad online, but this Los Angeles Times account of Bush’s advertising is emblematic of the kind of messaging his campaign ran to Latinos:
Over a fast-moving shot of a city skyline and an American flag rippling in front of the Texas state banner, a deep-voiced man reassures: “En nuestro pais ha llegado un nuevo dia.” (In our country, a new day has arrived).
The narrator continues in Spanish about Gov. George W. Bush’s family values in a 30-second TV commercial that features good-looking students, scientists and cheerleaders who could presumably benefit from his presidency. The spot concludes with the Republican candidate, wearing a cotton work shirt and jeans, saying four words: “Es un nuevo dia.”
The exit polls in 2000 showed him surging from the dismal 21 percent earned by Bob Dole in 1996 to 35 percent, and precinct results at the time appear to confirm this shift. Take a look at California for example, which was vaguely considered in-play at the time. My early #ElectionTwitter-style map showing localized swings from 1996 to 2000 showed Bush surging more in Hispanic areas of Los Angeles, along the border outside San Diego, in heavily Hispanic Imperial county just inland of that, and in parts of the Central Valley. As you can probably surmise, the color scheme here is reversed from modern conventions—Republican gains are in blue.
The more electorally significant swing was in Florida, where very similarly to the shift between 2016 and 2020, Cuban American voters surged from a tepid Republican showing in 1996 to a three-to-one advantage in 2000. Bill Clinton had been viewed as tough on Cuba, imposing new sanctions on the island after the Cuban military shot down two civilian aircraft piloted by Cuban exiles. That was all undone by an incident in 2000 when the Clinton administration sent Federal agents to a home in Miami to forcibly return young Elián González to his father back in Cuba. The Miami Cuban vote surged towards Bush. Without the Elián González incident, Bush would have lost the state and the presidency.
As crucial as the burgeoning Hispanic vote would prove to be in 2000, Bush’s Hispanic support reached new heights in his re-election in 2004, losing Hispanic voters by just 9 points according to the exit polls. This finding triggered waves of apoplexy among Democrats and seemed a bit too-good-to-be-true for Republicans. Left-leaning analysts like Matt Barreto of Latino Decisions worked overtime to debunk it with surveys and precinct data showing a Democratic advantage remaining mostly intact. The exit poll trutherism campaign was so successful that eventually people just rounded Bush’s number down to 40 percent. Either way, this was a high water mark for Republicans in the fastest-growing slice of the electorate.
The controversy over the 2004 exit poll highlights just how much we don’t know about how different demographics actually vote. The exit polls themselves can be unreliable when it comes to small demographic subgroups (I encourage you to take a look at their estimates of Asian voters over the years). And they miss in both directions: 44 percent for Bush in 2000 may have been too high, but its finding of 32 percent for Trump in 2020 was almost certainly too low and far off the double-digit precinct level swings seen in heavily Hispanic precincts.
The alternative to exit polls—which are often less reliable than the pre-election polls—is to look simply at precinct results, which is what Barreto used in 2004 and again in 2016 to debunk the counterintuitive exit poll finding of Trump gains that year. The problem with this method—known as ecological inference—is that while they can tell you something about which direction voters are trending from cycle-to-cycle, people who are surrounded by people just like them vote differently than people who live in integrated communities. The Hispanic voter in a Hispanic-majority precinct tends to be a lot more Democratic than the average Hispanic.
With that being said, ecological inference can be useful for validating trends, and the results in counties where Hispanics were a majority of the vote in 2020 confirm that 2004 was indeed a high-water mark for Republicans with Hispanic voters.
Here, it’s best not to focus on the margins, as heavily Hispanic counties are probably more rural (and likely more Republican) than the Hispanic vote as a whole. The vast majority of Hispanic voters are concentrated in metro areas, and even those with a high Hispanic population like Los Angeles, Houston, and San Antonio tend to be heavily influenced by trends in other groups, such as the shift of college-educated voters to the left. Comparing precincts over time proves difficult over a period as long as 2004 to 2020 because the composition of precincts changes too much over time, precinct boundaries change, or the data simply isn’t available.
Hispanic-majority counties are a rough and unrepresentative measure, but have the virtue of being more stable in their composition over time than large metro counties. These tell a story roughly consistent with exit polls and contemporaneous accounts. Trump surged, doing 20 points better on net than he had in 2016. But 2016 was a historic nadir, where he lost such counties by 34 points. He lost them by 14 points in 2020, but this was not quite as good as Bush, who lost them by 9 points in 2004 and 12 points in 2000. This would be consistent with a Trump performance that was 5 points worse on the margin in 2020 than 2004, and 2.5 points in popular vote share. For sake of argument, Bush getting something like 40 percent nationally in 2004 and Trump getting something like 37 percent in 2020 would be a plausible result, though the true figures are probably unknowable.
What This Means for the Electoral Map
When it comes to the Hispanic political battleground, there are three regions: Florida, Texas (some extra nuance coming up), and the Southwest. Yes, there are Hispanic voters elsewhere, but they are either a small percentage of their respective states or are located in states (New York, Illinois) not in danger of flipping at the presidential level.
In Florida, Trump basically matched George W. Bush’s performance in Miami-Dade County, performing one point worse than he did in 2004. A look at precinct data reveals that Trump’s performance, though strong, did not necessarily outdo the post-Elián González surge to Bush. If we expand our aperture a bit to Puerto Rican voters in the Orlando region—where Hispanics make up more than 40 percent of the vote in Osceola County—Trump performed worse, but this is almost certainly due to the composition of the electorate. The majority of Florida Puerto Ricans are new to the state since 2000. Their Democratic lean has pushed the Orlando region left, but Republicans are surging, closing the gap with post-Hurricane Maria migrants. Ron DeSantis, for example, won Osceola, doing notably better than Trump in 2020:
In Texas, we’ve read a lot about Trump’s surge along the Rio Grande Valley. And in some counties, like rural Starr and Zapata Counties, the surge was an epic one, with a 55 point swing in Starr County, once the most Democratic and machine-driven in the country.
But, using a starting point in 2004, the story becomes one of urban-rural polarization. Trump did better than Bush in rural counties and worse in urban counties, and because the population here is more urban than rural, Trump did worse overall. While Bush lost the Hispanic majority counties in the state by 10 points, Trump lost them by 21.
Here we need to account for Bush’s favorite son status, though. Bush won his home state by 23 points in 2004 while Trump only won it by 6. Hispanic voters have become much more integral to the Republican coalition since then, leaning more to the right relative to the state as a whole.
We also need to consider El Paso as distinct from South Texas. The significant gains Trump made in South Texas largely eluded him El Paso, which swung by roughly 8 points compared to 20-plus point swings in urban South Texas. Removing El Paso entirely from the Texas tally would mean that Trump performed just 5 points worse than favorite-son Bush in Hispanic-majority counties. It’s also further from being a true political battleground—though the Democratic Congressional incumbent notably underperformed here in 2022, a trend which bears watching. Politically, El Paso is closer to New Mexico, part of our next region, than it is Texas.
In the Southwest — here I’m counting Arizona, Nevada, New Mexico, and Colorado — Trumpism has yielded more mixed results. Trump lost all four states in the region, while Bush won all of them at least once. Here, our analysis relies more on overall state lean relative to the country as the Hispanic vote in each is notable, but mostly scattered in non-Hispanic majority metro areas.
In Arizona and Colorado, the story is mostly one of a shift among college whites—but Trump’s swings in Hispanic precincts in Phoenix were less impressive than those in downtown Houston or Miami, a telling sign of the rot in the political culture of Arizona Republicans compared to their highly professionalized counterparts in Florida and Texas.
But Bush went even further than winning now-tossups like Arizona. He won New Mexico, the nation’s most Hispanic state. Trump has not really come close to matching this feat, losing the state by 10 points — though the state did have a two-term Hispanic Republican governor Susana Martinez from 2011 to 2019. While Trump has made gains in some rural Hispanic majority counties, he lost 20 points relative to Bush in the state’s largest population center, Bernallilo County, home to Albuquerque. If Republicans lived up to their full potential among Hispanic voters, a working-class state like New Mexico would be competitive.
The one relative bright spot for Republicans in the region is Nevada, where Republicans retook the governorship last fall—their only high-profile statewide defeat of a Democratic incumbent—and came achingly close to winning a Senate seat. The state shows the promise and peril of Trump as a candidate: realigning voters in the most working-class swing state, shifting them to the right as the nation as a whole, but underperforming personally so he does not actually win it. Though better than his showings in Phoenix, Trump’s gains in Hispanic precincts in Las Vegas were spottier than they were in Texas or Florida.
When it comes to the Hispanic impact in the fight for 270 electoral votes, only a few states matter. One of them, Florida, appears to have been taken off the table by Ron DeSantis, whether or not he becomes the Republican nominee. That leaves the Southwest—Arizona, Nevada, and if we’re feeling a bit punchy, New Mexico—as the last true political battleground where Republicans can capitalize on their recent gains among working class minorities. And relative to Florida or Texas, Trump has underperformed there.