The overarching story of recent American elections is that 1) voters of color, who have long been Democratic-leaning, are a growing share of the electorate; 2) white voters with college degrees are increasingly shifting to the Democrats; and 3) white voters without degrees are aligning more with the GOP. But those trends, because they get so much focus, can warp our understanding of the electorate as it exists right now. Demographics are not yet destiny in American elections — millions of people don’t align with the party their race and ethnicity or education would predict. Case in point: In 2016, more than a third of President Trump’s support nationally came from non-Hispanic white Americans with college degrees (26 percent) and Asian, Black and Hispanic voters (12 percent), according to Pew Research Center data. On the flip side, about a quarter of Hillary Clinton’s supporters were non-Hispanic white Americans without degrees.
White Americans without degrees aren’t as likely to vote for Trump as in 2016, according to polls — which partly explains why Biden leads in national polls and key swing states like Pennsylvania. But a big reason Trump could still win the Electoral College, despite the poor marks Americans give him for his handling of COVID-19 and his job performance overall, is that the Black, Hispanic and college-educated white voters who backed him in 2016 are largely still with him, particularly in key swing states.
In other words, while Trump is a radical departure from previous GOP candidates in terms of personal style and his frequent racist comments, voters haven’t radically changed their voting patterns amid his rise in U.S. politics — the Americans who voted for Mitt Romney in 2012 overwhelmingly backed Trump in 2016 (about 90 percent) and those who backed Trump in 2016 are overwhelmingly behind him in 2020 (about 94 percent).
That fact helps explain how Trump won in 2016 and why he might still come back in 2020. Here’s a more detailed look at those three groups:
College-educated white voters
Polls suggest white people with college degrees nationally are likely to be more supportive of Biden than they were Clinton. (Pew data suggests that Clinton won whites with degrees by 17 percentage points, while Biden leads among them by 23 percentage points.)
But Biden’s advantage with white Americans with college degrees varies a lot by state, and that’s key. In Democratic-leaning swing states such as Maine, New Hampshire and Minnesota, only about a third of white voters with college degrees are supporting Trump, according to recent polls. That’s why he may not win any of those three states, even as whites without degrees — one of the more pro-Trump demographic groups — account for the majority of voters in all three.1 This represents a continuation of what we’ve seen over the last four years — polls showed college-educated white voters in Maine and Minnesota, in particular, were significantly more Democratic-leaning than those in the rest of the country in 2016 and 2018.2
The good news for Trump is that white voters with degrees in the South remain more conservative-leaning than those in other regions of the country. Polls suggest a majority of white college graduates in Georgia and Texas prefer Trump over Biden — as do more than 40 percent in Florida and North Carolina. Again, this isn’t too surprising: In 2018, white liberals nationally and in their states were enamored with Georgia Democratic gubernatorial candidate Stacey Abrams and Texas Democratic U.S. Senate candidate Beto O’Rourke. But white college graduates overall preferred Abrams and O’Rourke’s GOP opponents — by double digits in both races.
White voters with degrees in Georgia and Texas remaining fairly pro-Trump is hugely important. If he were to lose either state, it would severely complicate his path to winning 270 electoral votes. But he’s effectively tied in Georgia and Texas — even as white voters without degrees are a clear minority of the electorate in both states (just 32 percent in Texas and 36 percent in Georgia).
Outside of the South, a clear majority of white college graduates prefer Biden to Trump in most battleground states. In fact, it’s likely that white voters with degrees will vote Democratic at higher rates than in any recent reelection. That said, at least a third of white college graduates are likely to back Trump in basically every swing state.
What’s keeping this bloc with Trump? Many of these voters are simply longtime Republicans who hold conservative views. But particularly in Georgia and Texas, it’s worth thinking about religion and race. Being white and also an evangelical Protestant is strongly correlated with voting Republican — much more so than being white and not having a college degree. According to data provided to us from the Democracy Fund + UCLA Nationscape polling initiative, Georgia ranks seventh among states in terms of the percentage of its white registered voters who are evangelical Protestants (about a third), with North Carolina 10th and Texas 13th. For comparison, Virginia is the state that backed Hillary Clinton in 2016 with the highest percentage of its white registered voters who are evangelical Protestants — it was 17th among the 50 states by this measure.
In its surveys, Democracy Fund + UCLA Nationscape asks respondents their views on issues like whether they prefer their relatives to marry someone of the same race and whether they agree or disagree with statements like, “Generations of slavery and discrimination have created conditions that make it difficult for Blacks to work their way out of the lower class.” Robert Griffin, research director for the Democracy Fund Voter Study Group, analyzed responses to these questions to create an overall “score” for voters on their racial attitudes. Georgia’s white registered voters are more likely to express more negative sentiments about people of color than those in all but nine states. (So they are more likely than white registered voters in other states to reject the view that slavery and discrimination are holding back Black people and more likely to be unsupportive of their relatives dating people outside of their race.)
Florida (No. 11), Texas (No. 12) and North Carolina (No. 13) rank similarly on these measures of racial attitudes. Again, for comparison, the blue state where white registered voters hold the highest level of negative attitudes about minorities by this measure is Delaware, which is 17th highest of the 50 states.
So while white voters with college degrees are often portrayed as a monolith, and a Democratic monolith at that … they’re not. And that fact is crucial to Trump’s reelection chances.
Polls suggest between a quarter and a third of Hispanics nationally are backing Trump — fairly similar to 2016.3 And like in 2016, a sizable bloc of Hispanics in Arizona, Florida and Texas — three important swing states that have fairly large Hispanic electorates — are likely to back the president.4
Polls for the last two decades have shown about 30 percent of Hispanics identity as Republican. And despite his nasty rhetoric toward Mexico in particular, about as many Hispanics backed Trump as other recent Republican presidential nominees. As we wrote in 2018, there are plenty of explanations for that enduring pro-Trump Hispanic bloc: the anti-abortion views and Catholic and evangelical faiths of some Hispanic Americans ; the priorization of issues like jobs over immigration policy for many Hispanic voters, and the longtime courting of Hispanic voters by the GOP in Florida and Texas in particular.
But, again, the fact that the bottom hasn’t dropped out of Trump’s support among Hispanic voters is hugely significant to his continued competitiveness.
Polls suggest about 10 percent of black voters both nationally and in key swing states with large black electorates are supporting Trump. That is similar to 2016 as well and again reflects broader partisan dynamics — surveys over the last three decades have shown about one of every 10 Black Americans identifies as a Republican.
If the race is really tight, Trump’s maintaining this sliver of Black support could be critical. In an analysis of the 2016 election, Griffin, John Halpin and Ruy Teixeira argued that Clinton would have won Michigan and Pennsylvania if Black voters in those states had supported her at the levels they did Barack Obama. (About 95 percent of black voters in those states backed Obama, and about 90 percent supported Clinton.)
I don’t want to belabor that point too much — the Clinton-Trump matchup in those two states was so close that basically any increase in support among any demographic group would have won it for Clinton. And it’s not surprising that black voters supported Obama, the first-ever Black president, at slightly higher levels than they did Clinton. But again, the bottom hasn’t fallen out for Trump among Black voters — he is basically as popular as previous GOP candidates.
Last month’s Republican National Convention featured a lot of non-white speakers for a party whose voters are more than 80 percent white, and I think this analysis explains why. In theory, Trump could win this election by further growing the GOP’s advantage among white voters without degrees, who are likely to represent more than 40 percent of the electorate. That might happen to some extent, particularly among white men. But polls suggest that white women without degrees won’t be as GOP-leaning in 2020 as they were in 2016, either because they like Biden more than Clinton, have grown tired of Trump or some combination of both. So Trump probably needs the white college graduates and Black and Hispanic voters who backed him in 2016, and perhaps a few more, to back him again to win this election. And they might.
Make sure to check out FiveThirtyEight’s presidential forecast in full; you can also see all the 2020 polls we’ve collected, including national polls, Colorado polls, Florida polls, Michigan polls, Minnesota polls, North Carolina polls and … well … all the states, really.