A new analysis uncovers the demographic shifts that led to Joe Biden’s victory

MUCH ATTENTION has been paid to Donald Trump’s surprise performance in last year’s presidential election. The former president performed much better than the pre-election polls expected, especially in states with high concentrations of Hispanics or white voters without college degrees. But given that he won in spite of the misses, what did Mr Biden’s winning coalition actually look like?

A new analysis by Catalist, a firm that examines political data, reveals many different angles to Mr Biden’s victory. Catalist generates its estimates of how groups voted by combining polling interviews with tens of thousands of voters with administrative records of who turned out in the previous election. Pollsters call their cumulative databases storing these records the “voter file”. Because Catalist knows who American voters are—most importantly where they live, their demographics (such as age and race) and what their political partisanship is likely to be—it can generate estimates of whom they voted for. Catalist trains predictive models on the polling data, generates predictions for people in the voter file, then forces these estimates to match the actual results of the election in each precinct.

After a lot of modelling, the firm found that among the most important explanations for Mr Biden’s victory was that African-American voters helped the president clinch extra electoral votes in key sunbelt states. In addition, he made significant gains with college-educated and suburban whites, which helped propel him to victory in the north.

Democrats’ most devoted supporters are black Americans. In 2016, 93% of black voters who cast ballots for a major party voted for Hillary Clinton. In 2020, 90% voted for Mr Biden. But although black Americans as a whole became slightly more Republican, they also increased their turnout by 14 percentage points. This is helpful for the dominant group as its number of raw votes still grows. According to Catalist’s estimates, higher turnout among black voters provided nearly 1m extra votes for Mr Biden nationwide, relative to Ms Clinton’s numbers.

At the state level, higher black turnout alone could explain Mr Biden’s victory in Georgia and Arizona, the two closest battleground states. Of course, since the president’s margins there were so small, any number of demographic changes since 2016 could have put him over the top. Suburban white voters in Atlanta, for example, became much more Democratic over the past few years. But, put another way, if black turnout had stayed the same as in 2016, Mr Biden would have won just 279 electoral college votes instead of the 306 he actually racked up. Catalist’s numbers reveal the centrality of black voters to the Democratic coalition.

The biggest shift of any group came among Latino voters. They became much more Republican. Mr Trump’s vote share among the group increased from 28% in 2016 to 36% in 2020. Mr Biden’s share, meanwhile, was 61%—a six-point drop from Ms Clinton’s 67% four years earlier. Expressed as a proportion of ballots cast only for the major-party candidates, the Democrats’ vote share among Latinos fell by eight percentage points from 71% to 63% over the course of Mr Trump’s presidency. The shift among Cuban-American votes was 13 points, helping to explain why Mr Trump beat the polls by six points in Florida, but only by two or three in Texas and Arizona, where Latino immigrants are more likely to be of Mexican heritage.

Though Mr Biden did not stage the comeback with rural whites that many polls predicted, he did improve on Ms Clinton’s performance with the group. Though only 27% voted for Ms Clinton, 29% voted for Mr Biden. A two-percentage-point change might seem paltry, but in a close election small changes can make big differences. Mr Biden won by only 21,000 votes in Wisconsin, for example; rural whites provided a good portion of that margin.

Mr Biden’s biggest gains came in the suburbs and among college-educated white voters. Degree-holding whites sprung 54% for the Democratic candidate this year, up from 50% in 2016 (and 46% in 2012). This helps explain why suburban voters, the vast majority of whom are white, also drifted to the left. Mr Biden improved over Ms Clinton’s share there by three points, from 51 to 54%, according to Catalist’s estimates. These voting patterns match almost identically with the Democratic Party’s performance among suburban whites in the 2018 mid-terms. That indicates that their attitudes towards Mr Trump largely did not change over the second half of his presidency, despite the economic tumble in 2020 and his rough handling of the covid-19 pandemic. White women voted 48% for Mr Biden, two points higher than the share that supported Hillary Clinton. Mr Biden’s margin increased by four points among white men.

Demographics aren’t destiny… but they are prophecy

Overall, Catalist’s estimates confirm that the electorate is getting less white and much younger. Because of higher turnout among non-whites—particularly Latino- and Asian-Americans—in 2020 only 72% of voters identified as white. That compares with 74% the last time and 77% in 2008, the first year for which Catalist produced estimates. Since America’s younger generations are less white than its older ones, the country will continue to get more diverse for the foreseeable future.

Turnout also rocketed among millennials and Generation-Z voters (most of whom were not eligible to vote last time), a very Democratic group. The share of votes coming from these youngest generations increased from 14% to 31%. Boomers fell from 61% of the electorate to just 44%. The 2020 election was the first time since the late 1900s that they represented a minority of voters. Older millennials, aged 30-44, also became two points more Democratic (so did seniors).

These twin trends present a problem for the future of the Republican Party. Mr Trump may have stumbled upon a marginally winning message for Latino voters—to paraphrase, “the left is too ‘woke’ and socialist for you, and Republicans are good for economic growth”—but it was not enough to win over a majority of them, or of other non-white groups. The new voters who enter the electorate in the coming years will be less white and of younger, more progressive generations. This will push the country towards Democrats unless the Republicans change their policy agenda and figure out how to win young, diverse voters. They will also have to win more left-leaning suburban whites. So far, they have given no public signs they are evaluating such questions.