THE DEFINITIVE REASON HARRIS LOST

Just kidding. Here's how we can use data to start to answer this question.

When I woke up the day after, the New York Times had already published a video called “Why Trump Won” in which Times columnists each gave their take on why Harris lost to Trump. Ezra Klein talked about foreign policy.¹ Thomas Friedman blamed it on “cultural angst.” Nicholas Kristof blamed it on our inability to deal with inequality. Lydia Polgreen attributed it to “luck.”

Other pundits gave their takes via Twitter. David French tweeted that Harris’s loss was due to “defeat in Afghanistan, a porous border, inflation, and…Biden’s refusal to…step aside in time.” Some think that DEI lost the Democrats the election. Brianna Wu thinks that the Democrats lost the Jewish vote by being too antisemitic. Some think the Democrats were too liberal. Others think they weren’t liberal enough. Matt Yglesias tweeted that anyone who thought the “Democrats could have won by being more left-wing” was not engaging with reality.

These proclamations are silly and largely not based on data. They reveal more about the political beliefs of those who make them than anything real about the world. French is a neocon, so he overestimates the importance of the Afghanistan War. Yglesias is Twitter’s designated neoliberal shill, so he thinks Democrats need to be more moderate. Kristof is a progressive, so he immediately jumps to inequality as the source of the problem. I actually think these three pundits are very smart, but it seems to me they’re attributing electoral loss to their pet issue.

It’s irresponsible to attribute Harris’s loss to whatever issue you disagree with her on. If I did this, we’d say she lost because she didn’t promise to ban factory farms, she didn’t support appeasement with China and Russia, and she didn’t call for extreme AI regulation.

Obviously, this is flawed analysis. These are probably not the issues that lost her the election. We need to see the data! We don’t even know which party won the House. We need to get the exit polls, look at how different counties voted, crunch the numbers, and do some linear regression analysis. But I fear that too many elite liberals—DNC staffers, Times reporters, Brookings people, etc.—will make the same mistake they made in 2016: blaming it all on anything but themselves.

Blaming “irrational” or “racist” voters or indicting all of America is not useful. It doesn’t provide Democrats with any plausible future strategy to pursue (because, while Democrats can change their messaging, candidates, and policies, they presumably cannot make voters be less racist or irrational).² It’s a way of tossing up one’s hands, admitting no responsibility, and claiming moral purity. David Brooks put it quite eloquently:

There will be some on the left who will say Trump won because of the inherent racism, sexism and authoritarianism of the American people. Apparently, those people love losing and want to do it again and again and again.

I fear that the impulse to blame the electorate for your loss instead of looking inward is ubiquitous among liberal elites. And this way of approaching political science—going off vibes instead of data—is equally ubiquitous and pernicious. If we let the people at the DNC develop their worldview based on their own instincts and what they want to believe instead of data, we’re just gonna keep losing elections.

How to actually figure out why we lost

First, voting is complicated. People do it for plenty of reasons. Sometimes they weigh policy issues and choose the candidate that aligns most with them. Sometimes they care a ton about one issue and ignore everything else. Sometimes they choose the candidate based on their personality. Sometimes they vote purely on vibes.

Electoral politics is a complex social phenomenon, dependent on human behaviour and psychology. It is not a hard science, and individual elections are not repeated events. So, making hard-and-fast claims about correlation and causality is really difficult, even if you have lots of good data.

With that out of the way, let’s take a sober look at what data we have before making any claims.

Class matters more than race and gender

Women, and especially young women, swung toward Trump compared to 2016 and 2020. Education-level became a stronger correlate for voting, and neighbourhood density declined as a correlate. Compared to Mitt Romney in 2012, Trump has consistently improved his margins with black, Hispanic, and non-college educated voters.³

This means that liberals really need to drop the theory that people vote for Trump because they are racist. I already showed in another essay why the studies that supposedly support this idea are methodologically flawed and misrepresent the data.

In 2012, Democrats had an advantage with non-college educated voters, and college-educated voters supported the GOP. Now this class divide has flipped. The Democratic Party is no longer the party of the working class. Harris lost with voters under $100,000 in household income, and she won with voters over $100,000. In fact, despite faring worse than Biden with nearly every demographic, with these voters, Harris improved over Biden in 2020.

It is empirically true that the GOP is now the party of the working class. Even if you think working-class voters should support the Democrats, that doesn’t mean they do.

From the Financial Times

After the election, Bernie Sanders tweeted that “it should come as no great surprise that a Democratic Party which has abandoned working class people would find that the working class has abandoned them.” Jaime Harrison, the chair of the DNC, indignantly responded by listing all of Biden’s accomplishments for working Americans and calling Bernie’s post-election take “BS.”

Though it’s true that Biden has been quite pro-labour, I agree with Bernie here (as I’ve previously written about), and I think there’s a lot of evidence to back up the view that the Democratic Party has drifted to the right on economic issues over the last three decades. Biden’s recent policy wins are good for workers, but they can’t undo the harm that has been wrought by the neoliberal takeover of the Democratic Party. This phenomenon is explained in detail in John Judis and Ruy Teixeira’s wonderful book Where Have All the Democrats Gone?

Is Dick Cheney brat?

Allow me to do the very thing that those NYT columnists did, but I’ll be less definitive in my conclusions and I’ll provide argumentation and evidence.

If you agree with me that Democrats are losing ground among the working class because they’ve abandoned economic populism, you’ll see where the Harris campaign might have gone wrong. Harris ran to the right of Biden on economic issues, flirting with billionaire donors who wanted her to defang the NLRB and Federal Trade Commission. She spent more time campaigning with billionaire Mark Cuban than United Auto Workers leader Shawn Fain. On foreign policy, she embraced Dick Cheney and the neoconservative faction of the Democratic Party. In a patronizing way that is so reminiscent of Hillary Clinton’s doomed 2016 campaign, Harris clapped back “I’m speaking” to anti-war protesters who interrupted her event—even though a large majority of Americans want the U.S. to stop funding Israel’s war.

This implies that the DNC’s strategy for this election relied on a flawed model of electoral politics. They assumed that they could move hard to the centre, picking up moderate votes. They thought voters belonged to a right-left spectrum with social conservatives, fiscal conservatives, and war hawks on one end and social progressives, economic populists, and peaceniks on the other. But this isn’t the 1990s anymore. Things have changed a lot in the last 30 years.

Both major parties are basically to the right of the median voter on important economic issues. Vast majorities of Americans hold progressive views on campaign finance, labour, taxation, and healthcare. This is also true of foreign policy. Americans are largely restrainers, not interventionists. They are NATO-sceptical and they are not hawkish on Ukraine, Taiwan, or Israel.

I seriously doubt there were many Trump voters in Pennsylvania who switched their allegiance to Harris just because Liz Cheney told them to. My hunch is the true electoral spectrum has much more to do with being anti-establishment vs. being a member of the socioeconomic coastal elite. By my tentative model, the more Harris moved to the right on economics and foreign policy, the more she alienated swing state voters. I’ve been advocating this for a while now, since before Harris’s loss.

And it seems she lost on messaging. Trump succeeded, yet again, in cultivating a brash, anti-establishment image. Harris tried to run as the reasonable, small-c conservative, establishment candidate, and it didn’t work. This is perhaps because vast majorities of Americans have felt the country is on the wrong track and have been losing faith in government institutions for a decade.

This map, showing how counties swung from 2020 to 2024, is from the Financial Times.

But I’m not convinced of my theory. Like I said, we need more data and more time to really settle things. And there’s good reason to believe that Harris’s defeat was overdetermined due to other factors:

Other considerations

For the first time in history, every governing party in the world that faced elections this year lost vote share. As John Burn-Murdoch of the Financial Times wrote, “voters don’t distinguish between unpleasant things that their leaders and governments have direct control over.” They blamed the Democrats for inflation, even though it was a global phenomenon. And voters ignored the economists who said that Trump’s tariffs would cause even more inflation.

Also, Kamala Harris was a bad candidate. She did far worse than Democratic Senatorial candidates nationwide. Harris lost Montana by 20 points, whereas incumbent Democratic Senator John Tester lost the state by only 8 points. Sherrod Brown lost by 4 points in Ohio, whereas Harris lost by 11 points. This is perhaps because the Democrats had an extremely undemocratic selection process for their candidate. There was no legitimate primary, and the convention was not open. This is reminiscent of 2016 and 2020, when DNC bigwigs and major donors illicitly helped the establishment candidates (Clinton and Biden) defeat populist Bernie Sanders. A significant portion of Sanders voters split with the party after this, ultimately voting for Trump.

Also, voters overwhelmingly prefer Harris’s policies to Trump’s. This could be another point in favour of the theory that Harris was a terrible candidate. It also points to the idea that Democratic messaging is bad.

In a campaign, the important things a political party can influence are the candidate, the messaging, and the platform. (Obviously voters vote based on other things, like whether or not the economy is doing well, but parties can’t really affect that.) While Democrats had the right platform, they had a bad candidate and made many messaging errors. Harris refused an invitation to go on the most popular podcast in the world, which incidentally caters to millions of young American men, the exact demographic that swung so hard for Trump. Both Trump and Vance went on Joe Rogan’s podcast for three hours each. There were lots of other own-goals (e.g., Biden calling Trump supporters “garbage”; Harris not being able to differentiate herself from the unpopular Biden on The View).

Democrats really need to start winning

MAGA Republicans are terrifying, and they’re ready to make much more profound changes this time around. We need to get our act together on the left. Trump handily won the popular vote. Any successful movement on the left that seeks to significantly diminish the power of MAGA, not just chip away at its lead, will need to pursue paradigmatic changes.

Democrats need to do some focus groups, like Senator-elect Andy Kim did. When he won a previously red district in 2020, he organized listening sessions with Trump voters who also voted for him to understand what motivated them politically. They expressed extreme disgust with the status quo and politics more generally.

The DNC needs to fire the people who keep corrupting its primaries. Democrats need to cater to swing state voters, not billionaire donors. Liberals need to stop alienating people who don’t hold maximally leftist views on every single social issue. We need to ditch vibes-based political science. We need to take a long and hard look at the electoral data from the last few decades and figure out where it all went wrong. A tweet from political historian Brian Rosenwald summed up my thoughts quite well:

Here’s the challenge for the Democratic Party: Missouri voted for a $15 minimum wage, paid sick leave, and abortion rights and Democrats cannot compete there at all. Why? The answer to that is the road to a comeback.

1. Klein went on to publish a very good essay that should be required reading for all think-tankers and party staffers in DC.

2. Likewise, blaming Harris’s defeat on the political group you dislike the most—for example, the anti-war wing of the party—is not fair. Trump won by millions of votes. Third-party voters in Michigan, the state with the highest proportion of Arabs, did not swing this election. Perhaps they did swing the election by not voting. But if this is the case, that’s probably an indication Democrats should change their policy regarding Palestine and try to win them back.

3. Voting data from 2012, 2016, 2020, and 2024.