Amy Walter | May 10, 2021

We know that 2020 produced the largest voter turnout in modern history. But, for a detailed understanding of who voted—and how they voted—we had to wait until state voter files were updated and analyzed. This week, Catalist, a Democratic data analytics firm, released their first deep dive into the 2020 election using their database, which, they note “includes 15 years of voter registration records, supplemented by large-scale polling, modeling and precinct-level geographic analysis.” 

What they found was a national electorate more diverse than any in American history. Overall, they estimate that 72 percent of the electorate was white, a 2-point drop from 2016. As recently as 2008, 77 percent of the electorate was white. Turnout among Asian Americans was up 39 percent from 2016, while Latino turnout was up 31 percent. Even so, white voters continue to have the highest turnout rates of any group. Seventy-four percent of white voters turned out in 2020, compared to just 50 percent of eligible Latino voters, 63 percent of Black voters and 62 percent of Asian voters. In other words, even as turnout among voters of color increased, they are still punching below their weight. This was especially pronounced in fast-growing and diverse sunbelt states. A Catalist estimate of voters and non-voters across battleground states in the South and West — Arizona, Florida, Georgia, Nevada, North Carolina, and Texas found “nearly as many non-voting people of color (11 million) as there are voters of color (13 million), mostly concentrated in Latino communities. The numbers look quite different among white people in these states, with only 9 million non-voters and 24 million voters.”

Meanwhile, the Baby Boom generation’s reign as the dominant demographic group has ended as Millenials and GenZ have increased their political influence, a trend that Catalist calls “a demographic change that is functionally permanent.” Those under the age of 40 made up one-third of the electorate, a seven-point jump from 2016, while the share of the electorate of Baby Boomer age dropped to 44 percent. As recently as 2008, Boomers made up 61 percent of the electorate.

Yet, even as the 2020 electorate was more diverse than ever, Biden underperformed both Obama’s 2012 and Hillary Clinton’s 2016 performance with Latino and Black voters. How is that possible? After all, the conventional wisdom has long held that any surge in non-white turnout would benefit Democrats exclusively. 

Catalist estimates that “22% of Latino voters were first-time voters, who we haven’t seen in our entire database of general election voting going back to 2008. By contrast, just 14% of the national electorate was entirely new. When we take a more expansive view of new voters, including people who were new to voting in a given state compared to 2016, a full 40% of Latino voters were new presidential voters, while the national number is 29% of the overall electorate. With such a large number of new Latino voters in the electorate, it is plausible that they drove a big part of the change in Latino’s overall support numbers. As marginal voters enter into the electorate, their partisan preferences may move closer to a 50 / 50 split naturally.” In other words, these new voters were likely less ideological and less partisan than habitual voters, which helped boost Trump’s overall share of the Latino vote. 

However, the overall surge in Black turnout (Catalist estimates that turnout went up 14 points among Black voters from 2016) helped Biden yield “more net Democratic votes from Black voters as a whole in 2020 compared to 2016.” For example, in Georgia, “higher black turnout yielded an additional 200,000 net votes for Biden, above and beyond the margin that Clinton got from Black voters in 2016. This was key to winning the state, as Trump would have handily won the state without the extra 200,000 net votes.” In other words, despite the fact that Biden’s margin among Black voters was smaller than Obama or Clinton’s, the overall increase in Black votes more than overcame that shrinking margin. 

Overall, Catalist concludes that the unprecedented surge of new voters into the electorate — especially younger voters and voters of color, ultimately benefitted Biden more than Trump. They estimate that new voters “supported Biden at substantially higher rates overall (56%) than 2016 voters who returned in 2020 (51%).” 

Even so, some states with the highest percentages of new voters — like Texas and Florida — voted for Trump, while states with fewer new voters — like Pennsylvania and Minnesota- voted for Biden. For example, Catalist estimates that the total percent of new 2020 voters in Texas (defined as those who didn’t vote in that state in 2016), was 34 percent — five points above the national average. In Minnesota, just 22 percent of the electorate was new — seven points below the national average. 

When comparing election data from one year to the next we often assume that changes in voting behavior reflect a change in behavior of the same people. In other words, if candidate X took 40 percent of the Latino vote in one election, and 45 percent in the next, that candidate must have improved his/her standing with the same people who showed up the previous election. But, what can also be true is that the candidates attracted a very different group of Latino voters, including many that may not have voted in the previous election. It can also be the case that voters who showed up in the previous election stayed home in the next one. This tremendous churn means we have to be very cautious about making assumptions about the electorate’s composition in 2024 based on what we saw in 2020. It’s also why Catalist warns Democrats that despite Biden’s success with new voters in 2020, “we do not necessarily expect these new voters to signal a new, enduring Democratic majority, with ever-more Democratic-leaning new voters building larger and larger Democratic leads election after election. The flip side of the coin shown here involves people who dropped out of the electorate, who may also be more Democratic than the electorate as a whole. This is plausible if they are more likely to move around, face barriers to re-registering and voting, or are less likely to be consistent voters for any other reason.”

So, was the 2020 election an outlier or should we take the results to be part of a broader trend? The answer is both. 

Read the article on the Cook Political Report >