Skip to main content
ABC News
American Politics Now Has Two Big Racial Divides

There’s been a recent flurry of studies and analyses that take a deeper look at the results of the 2020 election. These examinations don’t contradict our early interpretation of the results from the days and weeks immediately following Election Day: The overwhelming majority of voters backed the candidate from the party that they normally lean toward, though then-President Trump did slightly better with voters of color and slightly worse with white voters than he did in 2016. But the new examinations and other data tell a nuanced story about the role of race in the 2020 contest.

What Biden’s speech can – and can’t – accomplish | FiveThirtyEight Politics Podcast

American voters …  

  1. Remain deeply polarized based on ethnicity and racial identity; 
  2. Were less polarized by racial identity in 2020 compared to 2016; and 
  3. Are very polarized by attitudes about racial and cultural issues. 

Racial differences in vote choice are still huge.

Since American presidential elections are so close, fairly small shifts in the electorate really matter in affecting who wins. But I worry that the media’s understandable emphasis on those shifts often overshadows longstanding patterns in American politics that include the overwhelming majority of voters, who aren’t swinging between the two parties. Despite the news coverage that sometimes implies that non-Hispanic white voters with college degrees are all flocking to the Democrats, about 42 percent of that group backed Trump in 2020, according to the recently released Cooperative Election Study. About 64 percent of Hispanic Americans backed Biden, per CES, which might be hard to remember amid the intense (and accurate) coverage of Trump’s gains among that voting bloc.

2 Protestors standing with signs that read “Stop Asian Hate”

related: Why The Recent Violence Against Asian Americans May Solidify Their Support Of Democrats Read more. »

In many ways, the 2020 election was basically like every recent American presidential election: The Republican candidate won the white vote (54 percent to 44 percent, per CES), and the Democratic candidate won the overwhelming majority of the Black (90 percent to 8 percent), Asian American (66 percent to 31 percent) and Hispanic (64 percent to 33 percent) vote. Like in 2016, there was a huge difference among non-Hispanic white voters by education, as those with at least a four-year college degree favored Biden (55 percent to 42 percent), while those without degrees (63 to 35) favored Trump. (There wasn’t a huge education split among voters of color.)1

Other surveys tell the same general story: Trump won white voters overall by a margin in the double digits and won whites without four-year degrees by even more; Trump lost among whites with at least a four-year college degree, lost by a big margin with Asian American and Latino voters and lost by an enormous margin among African Americans.

So the main reason that Trump nearly won a second term was not his increased support among Latinos, who are only about 10 percent of American voters and are a group he lost by more than 20 points. Trump’s main strength was his huge advantage among non-Hispanic white voters without college degrees, who are about 42 percent of American voters. His second biggest bloc of support was among non-Hispanic white Americans with degrees, who are about 30 percent of all voters. According to the CES, over 80 percent of Trump’s voters were non-Hispanic white voters, with or without a college degree. In contrast, around 70 percent of nonwhite voters supported Biden, and they made up close to 40 percent of his supporters. So it is very much still the case that the Republicans are an overwhelmingly white party and that the Democratic coalition is much more racially diverse.

With all that said, however … 

Racial differences in vote choice were smaller in 2020 than in 2016.

Trump did 7 percentage points better among Asian American voters in 2020 compared to 2016, 4 points better among Hispanic voters and 1 point better among both white and Black voters, per the CES. Biden did 4 percentage points worse among Asian American voters and 1 points worse among Hispanic voters compared to Hillary Clinton, while doing 1 point better among Black voters and 3 points stronger among white voters compared to Clinton.2

Republicans can govern without winning a majority. That threatens our democracy.

Other surveys and precinct-level data suggest that the Trump swing among Hispanics could have been larger than CES found, with Trump gaining in the upper-single digits and winning the support of over 35 percent of Latino voters. (Ultimately, we will never know exactly how different racial and ethnic blocs voted, since people aren’t required to state their race or ethnicity when they cast ballots.) But generally, the story of 2020 is that Trump did better with Asian American and Hispanic voters than in 2016, while Biden did better than Hillary Clinton among non-Hispanic white voters. 

And these shifts had electoral consequences. Republicans flipped two U.S. House seats in California with Asian American candidates running in those districts, which have relatively high shares of Asian American voters. Gains by Trump and GOP congressional candidates among Miami-area Latino voters helped flip two more House seats to the GOP, according to a new analysis of Latino voters by Equis Research. Arguably the most important shift in the electorate was Biden’s gain with white voters, since white voters are both the largest bloc in the electorate and make up a disproportionately large share of the vote in swing states (and often tipping-point states in the Electoral College) like Michigan and Wisconsin. 

Racial attitudes, not racial identity, may be driving voting behavior 

At first glance, it might seem surprising that Trump gained among voters of color, since he often demonized Asian Americans, Black Americans and Latinos and invoked racist tropes. I will admit that I did not expect Trump to make those gains before polls started showing them in the run-up to the election.

But the data and research in the wake of the 2020 election suggests that many voters of color who backed Trump either already held GOP views on some racial issues or adopted those views to align with their decision to back Trump. So their views on racial issues are often closer to those of white Republicans than people of color who are Democrats. Meanwhile, white Democrats tend to have racial views much closer to people of color who are Democrats than white Republicans. When you put those two things together (white Democrats getting more racially liberal and many people of color who are Republicans not being liberal on racial issues), the results are that Republicans and Democrats are very divided about views on racial issues, even as they are becoming less divided in terms of racial identity (more white people are Democrats, more people of color are Republicans).

For example, around 15 percent of Black adults and 38 percent of Latino adults either said they opposed the Black Lives Matter movement or were non-committal about it (they didn’t support or oppose it), according to polling conducted by Civiqs around last November’s election. That’s fairly similar to the percentage of those groups that voted for Trump in 2020.  Furthermore, about half of Black adults who said they opposed Black Lives Matters’s goals backed Trump in 2020, while only 7 percent of Black adults who agreed with the goals of Black Lives Matters backed Trump, according to polling from PRRI. Seventy-three percent of Latino voters who opposed Black Lives Matters’s goals backed Trump, similar to the 83 percent of white voters who opposed Black Lives Matters’s goals. 

Biden wearing a mask at a press conference. The left half of the page is all microphones and journalists taking notes. On the right, behind Biden, you can see the Washington Monument in the background.

related: Five Questions The Political Media Is Wrestling With In Covering America’s ‘Uncivil War’ Read more. »

We don’t know whether those voters were already suspicious of Black Lives Matter and that led them to vote for Republicans, or if they were conservative-leaning and therefore adopted the anti-Black Lives Matters posture of the GOP. Either way, that’s an important finding; it means that some people of color will vote for Republicans despite the GOP’s resistance to Black Lives Matters. In contrast, to be a Democrat is basically to support the Black Lives Matter movement, even if you are not Black. About 88 percent of white Democrats said they supported Black Lives Matter last November, similar to the number of Black (90 percent) and Hispanic Democrats (85 percent) who expressed the same sentiment.

Similarly, among voters of color who backed Trump, 54 percent said they had never heard Trump do or say anything that could be considered racist, according to a post-election survey from PerryUndem. That puts those voters closer to white men who backed Trump (74 percent of whom said that he didn’t say racist things) than white Biden voters, 93 percent of whom indicated that Trump had said racist things. About 21 percent of Hispanic Republicans said that white people in America face a lot of discrimination, according to a recent Pew Research Center poll. That’s similar to the number of white Republicans who have that view (28 percent) and much higher than the share of white Democrats (4 percent) or Hispanic Democrats (6 percent) who said that white Americans face a lot of discrimation. Most Latino and Asian American voters who strongly disagreed with the idea that Black Americans are particularly disadvantaged because of slavery and racial discrimination supported Trump, like white voters who held that view, according to an analysis of pre-election Democracy Fund + UCLA Nationscape polling by Robert Griffin of the Voter Study Group.

It’s not clear that voters of color backed Trump because of these racial views — so we can’t really say if the Democratic Party is too “woke” for these voters. It could be that they really like Trump and the Republicans’ stances on racial issues, or simply that those racial issues aren’t driving their voting choices. The Equis analysis argued that compared to 2016, immigration wasn’t as big of an issue in the election discourse in 2020, helping Trump win Latino voters who liked his economic policies but had been turned off by his racist rhetoric in 2016. 

I don’t mean to suggest that racial issues don’t matter to voters of color. It’s almost certainly the case that the Republicans’ bad reputation on racial issues is a major reason for the broad patterns that I noted at the start of this story — Asian American, Black and Hispanic voters all lean heavily Democratic. But the 2020 results suggest that as American politics are increasingly divided on these lines of racial attitudes, the Democratic Party may be wooing more swing white college graduates than people of color. They also indicate that the often racist rhetoric of Trump and Republicans like him doesn’t turn off swing voters of color as much as it turns off white college graduates.

Overall, racial identity remains a huge factor in U.S. elections. The most important problem that Democrats face is that they consistently lose among America’s biggest racial voting group (white voters), and one of the biggest cohorts in that racial group (white voters without college degrees). It is a huge problem for Republicans that the clear majority of people of color vote against them, since that’s a big and growing bloc of the electorate. It is unlikely those broad dynamics will change. 

That said, it’s possible that the Democrats keep making gains among white voters and Republicans among voters of color. And even if those voting patterns don’t change, it seems very likely that the two parties will remain fairly split on racial attitudes. American politics in 2021 may be best described as a majority-white but heavily nonwhite racially liberal Democratic Party against a racially conservative and very-but-not-totally white Republican Party, with people in each camp pretty unified around their views on major racial issues.

Why House Democrats probably won’t pass their reparations

A majority of Americans think climate change should be a political priority

CORRECTION (May 3, 2021, 11:53 a.m.): An earlier version of this article incorrectly stated that the U.S. electorate is about 45 percent non-Hispanic white voters without college degrees, according to “States of Change,” a demographic report on the electorate. According to “States of Change,” this share of the electorate is actually 42 percent.


  1. Surveys from 2020 suggest that Black voters with four-year college graduates overwhelmingly backed (90-8) for Biden (90 percent voted for him, versus 8 percent for Trump, per CES) and Black voters without four-year degrees graduates (90-9) voted similarly. Hispanic voters with at least a four-year degree voted similarly (66-31) to those without one (63-33). Asian Americans with college degrees were more Democratic-leaning (72-26) than those without degrees (57-40) according to the CES, but the survey had a fairly small number of Asian respondents (about 1,200), so I don’t want to overemphasize that particular finding.

  2. These numbers aren’t zero-sum because many fewer voters (2 percent) backed a third-party candidate in 2020 compared to 2016 (6 percent. So, for example, both Trump and Biden won a slightly higher share of Black voters in 2020 compared to 2016. CES will release a validated voter report later this year, in which they use voter files to determine exactly which of their respondents actually voted. Such validated results do not typically vary much from the initial survey results.

Perry Bacon Jr. was a senior writer for FiveThirtyEight.