Senator Elizabeth Warren, who ended her presidential campaign last week, was an unusually popular candidate among political journalists and white professionals with college degrees, so perhaps it’s no surprise that I, too, like her a lot, despite our significant political and ideological differences.
I like her intelligence, her range of experience, her energetic comportment, her ability to grasp and grapple with complicated matters of public policy, and the incisiveness of her best writing and remarks. Among her rivals, I found her less likely to start a dumb war than Joe Biden, more perceptive about capitalism’s vital upsides than Bernie Sanders, more possessed of virtues like strength, toughness, and honor than Donald Trump, more heterodoxly attuned to some of America’s most pressing challenges, as Will Wilkinson of the Niskanen Center argues, and personally likable. I’d rather have a beer with her than any of the other present or former Democratic candidates this cycle.
The failure of her campaign warrants a public conversation about its causes that is as nuanced, rigorous, and comfortable with complexity as Warren herself is at her best. But so far, the narrow question of whether sexism was to blame for her loss has so dominated public discourse as to miscast a formidable political actor as a hapless victim.
Sexism could well be one factor among many that cost Warren votes. But taking the sexism hypothesis seriously requires subjecting evidence to scrutiny like with any other argument. When asserting that the Warren campaign was disadvantaged by sexism, does the speaker simply mean that there is a tiny number of sexist voters who refuse to support women? Or are they making the narrow claim that Warren lost more votes of people averse to a female candidate than she gained votes of people excited by a female candidate? Or the more sweeping claim that Warren would have won the nomination but for sexism? Or the counterfactual claim that she would have won as a man, holding all else equal?
Specificity is vital, not only because no variation of sexism can be remedied until it is addressed in particular, but also because overstating the sexism that female candidates face may significantly erode women’s support among voters concerned about electability. Put simply, overstating sexism harms female candidates.
When significant male candidates fail to prevail in a campaign, they inspire multifaceted postmortems that grapple with their actions and ideas. In what follows I show Warren that same respect.
No Lane for Warren
In one telling, this cycle’s Democratic field was composed of progressives, led by Bernie Sanders, and moderates, led by Joe Biden, with Elizabeth Warren offering herself up as an alternative to both while positioning herself well to the left of Biden and slightly to the right of Sanders. This strategy arguably wound up making her less acceptable to progressives and moderates alike. Jill Lawrence offered a variation on this theory in USA Today, arguing that Warren lost voters who were tired of disruptiveness to Biden while voters who wanted disruption flocked to Sanders.
Similarly, Matt Yglesias split up the Democratic coalition between working-class, “beer track” Democrats, and richer, “wine track” Democrats. He argued that the candidates with the most working-class appeal, Biden and Sanders, competed for the beer-track vote, while the candidates who drew wine-track voters––Klobuchar, Buttigieg, and Warren––were competing three ways for a much smaller cohort. Note that there is a long history of liberal Massachusetts senators running for president and failing to connect with the working class. Warren has joined a club with John Kerry, Paul Tsongas, and Edward Kennedy in it.
Medicare for All, Revolution, and Authenticity
“I have a plan for that.” That’s the phrase that many associated with Warren early in her campaign. Her pitch was less exciting to some than promises of revolution. But it held together better. She seemed smart, knowledgeable about details, and rhetorically scrupulous. Many believed that as policy issues arose, she would dig into the substance and make decisions based on what really made sense with atypical seriousness and rigor. But her shifting positions on health care hurt her severely.
First, it was suspect when the woman with a plan for everything hadn’t articulated a health-care plan or how she would pay for it. Then she raised her hand in a debate and said she was in favor of eliminating private insurance, a position that seemed calculated to compete for Sanders voters, but that likely cost her some support from those most concerned about electability against Trump.
Warren hurt herself more when she stood on a debate stage and refused to give straight answers to questions about how she would pay for Medicare for All. The college-educated white professionals who tend to like Warren also tend to like Klobuchar and Buttigieg. And those rivals accused Warren of being evasive and inauthentic. “At least Bernie’s being honest and saying how he’s going to pay for this and that taxes are going to go up,” Klobuchar said. “I’m sorry, Elizabeth, but you have not said that, and I think we owe it to the American people to tell them where we’re going to send the invoice.” Suddenly, Warren seemed like a triangulator.
In an astute postmortem for NBC News, Benjy Sarlin explained why Medicare for All was a less risky position for Sanders, given his base, than for Warren. If you’re a college student or twentysomething who isn’t yet established in your career, or if your work doesn’t get you health care, or if your job is insecure, then the end of employer-provided health insurance is no great loss. But Warren relied more on older, college-educated voters who are wealthier than Bernie’s base––people likely to have secure jobs with insurance they like and that their kids rely on.
Later, when Warren declared that she would not even pursue Medicare for All until year three of her presidency, voters who opposed it were still worried, while those who favored it doubted her commitment.
It was admirable for Warren to affirm the dignity of historically marginalized racial and religious groups as well as gays, lesbians, and transgender people. But Warren went further than articulating the importance of diversity, equal treatment, and rhetorical inclusiveness into what looked, to people outside the most progressive bubbles in America, like radical performative pandering, akin to when Trump hugs and kisses the American flag like it’s a teddy bear.
Warren declared, during an appearance in Cedar Rapids, Iowa, that a particular transgender youth she met on the campaign trail would interview her eventual choice for secretary of education and exercise veto power over the selection. “Only if this person believes our secretary of education nominee is committed to creating a welcoming environment, a safe environment, and a full educational curriculum for everyone,” Warren pledged, “will that person actually be advanced.” That’s a patronizing stunt, not respectful outreach. We’re talking about a 9-year-old! Imagine the typical Democrat’s reaction if Trump plucked a Boy Scout from the crowd at a rally and declared to applause that the child would have veto power over the next secretary of defense.
Many of Warren’s language choices baffled me. For example, she used the term Latinx, despite the fact that, according to surveys, most people of Latin American descent feel ambivalent or even hostile to that neologism. “Warren is distinctively beholden to a hermetic academic-progressive world,” Ross Douthat hypothesized, “to a point where she doesn’t know how to talk to the less-ideological, less-woke, maybe-even-somewhat-conservative Hispanics whose votes her party needs.”
Warren listing her pronouns as she/her may have suggested politeness to trans people inside the elite progressive bubble, but outside of it, it looked like a famous public figure in her 70s faux-clarifying what everyone obviously knew to signal adherence to woke orthodoxies.
Warren’s woke virtue-signaling apparently did her no favors with historically marginalized groups. The New York Times reported on Warren’s success winning endorsements from black progressive activists––and her failure to gain traction among black voters. “The progressive activists who have showered her candidacy with validation have a different electoral lens than the black electorate at large,” the newspaper noted.
Even in her home state, Warren lost African Americans—in fact, she came in fourth, behind Biden, Sanders, and Bloomberg. She came in third among Latinos and a distant second among LGBTQ voters.
The research on whether sexism hurts female candidates is mixed. In a meta-analysis of studies about gender and politics at 538, Amelia Thomson Deveaux noted that “studies have found voters may be more biased against women when they run for executive offices,” but also that if you’re a female candidate Democratic voters will probably be especially friendly. “A recent experiment conducted by CBS and YouGov looked at the qualities prioritized by Democratic voters in a series of matchups between hypothetical candidates, and it found that Democrats showed a clear preference for female candidates,” she wrote. “One of the studies about executive leadership indicated that Democrats are more likely than Republicans and independents to see women as viable leaders, and a meta-analysis of multiple studies also found that Democrats are more likely than Republicans to support female candidates.”
Lefteris Anastasopoulos found that in House elections between 1982 and 2012, female candidates, while far less common than male candidates, did not suffer disparities in fundraising, vote share, or probability of victory. And in another study of congressional races, Kathleen Dolan of the University of Wisconsin found “no evidence of any direct, consistent, or substantial impact for gender stereotypes on evaluations of, or voting for, women candidates.”
Some empirical evidence suggests that sexism may have cost Warren, in particular, some votes. Using the hostile sexism battery, a scale developed by social psychologists, the researchers Brian Schaffner and Sam Luks found last July that “among the least sexist voters, Biden and Warren are neck-and-neck; among the most sexist Democratic primary voters, Biden is preferred by as much as a four-to-one margin.”
In a study conducted by the progressive think tank Data for Progress, Warren received little to no support from the roughly one-third of the Democratic primary electorate that does not reject premises like women are too easily offended, most women fail to appreciate fully all that men do for them, and women seek to gain power by getting control over men. (The studies would tell us more about the net effect on electability if they included analogous questions about male candidates. For example, what percentage of Democratic primary voters agree that men fail to appreciate all that women do for them? And are they much less likely to support male candidates?)
So far, the best data we have about actual votes comes from exit polls, including the Super Tuesday outcome that most surprised me: Warren’s third-place finish in her home state. Massachusetts voters have repeatedly demonstrated a willingness to elect a woman president: They preferred Hillary Clinton to Barack Obama in 2008, they preferred Hillary Clinton to Bernie Sanders in 2016, and Clinton bested Trump in Massachusetts by 30 points. Massachusetts voters have also preferred Warren in two statewide Senate races against men: She beat a male incumbent in 2012 and a male challenger in 2018.
How did Warren do among the voters who know her best? According to The New York Times, women were 56 percent of the electorate and Warren lost women to both Biden and Sanders. In her best demographic, white female college graduates, she barely bested Biden, winning 33 percent to his 31 percent. Among white females who hadn’t graduated from college, she won 15 percent, just four points better than Mike Bloomberg. Biden won 44 percent. At least in Massachusetts, the hypothesis that sexism cost Warren a win seems very weak.
In Texas, where 53 percent of the Democratic primary electorate was female, Biden got 34 percent of the female vote, Sanders got 30 percent, and Warren got just 13 percent, tying with Bloomberg. Although Warren lost every age cohort, she did much worse with the youngest voters, up to age 24, than with voters between 25 and 64, the opposite of what one might expect if traditional views about gender were behind her struggles. In California, which has had two female senators for more than three decades, and where Clinton handily beat Obama in 2008, Warren finished a distant third among women voters. In Virginia, where 57 percent of the primary electorate was female, just 11 percent of women voted for Warren.
Nothing in these numbers substantiates the conclusion that sexism played a conclusive or even a major role in Warren’s loss, and the fact that she performed so poorly among women, who were a majority of the electorate, clashes with intuitions about what a loss due to sexism would look like. As Matt Yglesias observed, Warren did better with college-educated white men than with working-class women. “While no candidate wins any demographic universally, Warren didn’t come close with women. The stronger predictor of who supported her and who didn’t was education, not gender.”
One retort is that women failed to vote for Warren because they didn’t believe that she could win. In this telling, the problem isn’t voter animus; it’s that women are seen by voters as less electable, including by women who fear that sexism makes it impossible for a woman to become president.
A more fleshed-out version of this argument appeared in Mother Jones in late January. “The electability of women remains in question because people keep questioning women’s electability,” Pema Levy reported, citing research showing that voter perceptions of whether other voters would cast a ballot for a woman matter more than voter self-perceptions and preferences.
This gets confusing quickly, as some people who complain about that dynamic simply call it sexism, using the same word for unequal treatment of women and fear of that unequal treatment. If people are happy to vote for and live under a female president, but refrain from voting for one because of their belief that others aren’t ready, is that “sexism”? Choosing a different, more precise word for the phenomenon would better serve clarity and female candidates, since the theory itself posits that assertions of sexism can hurt female candidates.
“Getting people to understand that sexism is a factor could inadvertently make them less willing to vote for a woman,” Thomson DeVeaux of 538 wrote, capturing the ouroboros-like quality of this hypothesis. “It’s a weird, self-defeating conundrum. You get people to accept that sexism in politics is real and it shakes their confidence in whether women can win.” After Warren dropped out, the advocacy group UltraViolet Action stated, “It is clear that there is a glass ceiling held firmly in place for women by a media who relentlessly shape voters’ perceptions of who is electable through a deeply sexist lens. In a year in which primary voters’ top concerns is electability—the media has had a massive impact on how voters perceived the candidates—and when Warren was on the top of the polls, the main narrative driven by the media was that she was not electable.”
I’m of two minds about this theory. Democratic voters this cycle said electability was very important to them, and it is plausible that many of those voters worried that it would be harder to beat Trump with a woman because, rightly or wrongly, they perceive the electorate as sexist.
But the description of the news media put forth by many Warren supporters, wherein she was treated worse than other candidates, questioned more about electability, and disadvantaged overall, seems dubious to me. As best I can tell, Warren was the favorite candidate of journalists, endorsed by The New York Times (along with Klobuchar), The Des Moines Register, The Austin Chronicle, and The Boston Globe. Commentators have expressed grave doubts about the electability of Sanders at least as frequently as Warren, and when I look at the voters Warren won, they overlap a great deal with the demographics of people who are the heaviest consumers of political media, whereas the voter demographics she lost overlap with demographics that consume much less political media. If constant questions about electability from the media caused her loss, how to explain that?
I remain open to evidence of sexism in any individual contest and believe it is worthwhile to face it squarely. Sexists walk this earth. Some of them vote in Democratic primaries. Without a doubt, some voted against Warren, just as some anti-Semites voted against Bernie Sanders. There’s no way to know exactly how many or to what degree they were offset by voters who preferred Warren because they want a woman president. It’s plausible that being a woman cost Warren votes on net and that she suffered more for her gender identity than any other candidate suffered for their identity traits. It is also plausible that if she were a man she’d have had no more success than Michael Bennet, Deval Patrick, Beto O’Rourke, John Delaney, Cory Booker, or Andrew Yang.
Given that Warren finished a distant third even among women in every significant Super Tuesday state, the notion that she would’ve won but for sexism is highly implausible, even if sexism hurt her candidacy on the margins.
Going forward, Democrats should stop worrying that nominating a woman would cost them elections. No definitive evidence shows that sexism is a significant drag on female presidential candidates. When Warren or a candidate like her runs again, her success is far more likely to turn on her ability to differentiate herself from rivals, the support that she attracts from the working class and people who didn’t attend college, and avoidance of the kind of pandering that alienates many Americans. Given that none of the remaining contenders for the White House will serve more than one term, focusing on those missteps rather than sexism gives Warren her best shot at winning in 2024.