Electability has been the buzzword of the 2020 Democratic primary for president.
Electability—that is, which candidate seems best positioned to take on Donald Trump—has fueled former Vice President Joe Biden’s rise in the polls to become an early front-runner. However, electability has hurt female candidates, many of whom answer questions from voters about whether a woman can win the presidency after Hillary Clinton’s 2016 loss.
Polls show that Democrats want a candidate who is electable above all other characteristics. And there’s a reason for that—President Donald Trump was viewed as unelectable and won narrowly, so now Democrats are eager to find the person who will not face the ignominy of losing to him a second time.
Voters and pundits have often interpreted electability as running a white male, who would be more typical of how past presidents have looked, against Trump. Moreover, much of the attention in the presidential race has been on who can win back the 8 million or so voters who switched from Barack Obama to Trump in the 2016 election, concentrated in the key swing states of Michigan, Wisconsin, and Pennsylvania, none of which had voted for a Republican before Trump since at least 1988.
Sen. Kamala Harris, speaking to an almost all-black audience, slammed this idea of electability, speaking in Detroit on May 6.
“There has been a conversation by pundits about ‘electability’ and ‘who can speak to the Midwest,’” she said in front of the NAACP. “But when they say that, they usually put the Midwest in a simplistic box and a narrow narrative. And too often their definition of the Midwest leaves people out. It leaves out people in this room, who helped build cities like Detroit.”
Harris’ comments pointed to another, less talked-about group of voters that Hillary Clinton missed in 2016 and led to Trump’s victory: disproportionately—but not exclusively—African-American voters, many of whom stayed home in 2016 after turning out in record numbers for Barack Obama in 2008 and 2012.
Research done by political scientists indicates that their staying out was crucial to Trump’s victory.
According to voter data compiled by political scientists Bernard Flaga, Brian Schaffner, Jesse Rhodes, and analyst Sean McElwee, African-American turnout dropped by about 12 points in Michigan and Wisconsin in 2016 from 2012, and about two points in Pennsylvania. Had voting patterns remained constant across racial and ethnic lines, they argue that Clinton would have won these three states. Simply put, fewer black people voted and more white people voted in key swing states.
One common explanation for this drop, cited by Clinton and others, is voter-identification laws, though political scientists John Sides, Lynn Vavreck, and Michael Tesler note in their book, Identity Crisis: The 2016 Presidential Campaign and the Battle for the Meaning of America, that voter identification laws did not correlate much with black turnout. Rather, they argue that Obama’s “extraordinary black support was concentrated among African-Americans with a strong sense of solidarity with other blacks,” and that Hillary Clinton did not enjoy such support, despite having high favorable ratings—but far lower percentages of very favorable ratings and therefore enthusiasm—among African-Americans.
The authors argue that even with the drop in white turnout, it was still possible for Clinton to win in 2016 with elevated African-American voter turnout to Obama levels. However, turnout levels remained at levels similar to Democratic losses in 2004 and 2000.
For its part, the Trump campaign, according to Bloomberg, worked to reduce black turnout by highlighting Clinton’s past comments on criminal justice and “super predators” on social media and radio.
Trump won the presidential race by about 78,000 votes in three states, meaning that any small change in voter turnout would affect who won.
Tesler, a political scientist at UC-Irvine, thinks that it’s unlikely that a white candidate could mobilize Obama-like levels of support among African-Americans.
“Joe Biden would probably have the best chance given his association with Obama,” he said in an email, adding that a “white candidate like Bill DeBlasio with progressive racial policies and close personal ties to the black community might also help mobilize African-Americans,” referring to the New York City mayor who is reportedly soon announcing a bid.
Moreover, the electability contest among voters may be altogether too hard to game out so early on in the race. Even putting Trump aside, people just aren’t very good at telling who is electable. Vavreck, a UCLA political scientist, noted that voters have a hard time judging how other people will vote.
“From a survey research point of view, it is harder to ask people to make judgments about what they think other people are going to do in the future. So—asking people how they will vote is one thing; asking people how they think everyone will vote is a different thing—and that is by definition harder for people to estimate,” she told Fortune.
Besides Trump, Obama was also once viewed as unelectable. Former 2008 Clinton campaign top strategist Mark Penn predicted in an infamous 2007 memo that Obama was “unelectable except perhaps against Attila the Hun.”
He would go on to win 365 electoral votes and the largest number of votes ever won by a presidential candidate.
More must-read stories from Fortune:
—Nancy Pelosi says President Trump is becoming “self-impeachable”
—This proposed legislation could light up the pot business
—What exactly is ranked-choice voting?
—Get up to speed on your morning commute with Fortune’s CEO Daily newsletter