Ezra Klein, the editor-at-large at Vox and the host of The Ezra Klein Show podcast, has a new book out called Why We’re Polarized. He agreed to chat with us about the book’s arguments and the role that technology plays in it.

Nicholas Thompson: Welcome, Ezra! The book is marvelous. I read the galleys pretty much cover to cover and was captivated. You take readers through American history and explain persuasively that we’re vastly more polarized now than in the past. And your publisher craftily timed your release exactly to the impeachment hearings, which were perhaps the most partisan moment in history.

Ezra Klein: Thank you!

NT: When I started, I expected to find a chapter on filter bubbles and Facebook. Or about how Twitter pushes political conversations to extremes. Or how YouTube takes people from cookie-baking tips to jihadism. There are a few bits about tech, many of which are intriguing. But why don’t we start with this: Explain your general thesis and why the social platforms play a relatively small role in it.

EK: In a way, these answers are the same. The core story of the book is that over the past 50 years, the country’s dominant political coalitions have sorted by ideology, race, religion, geography, psychology, consumer behavior, and cultural preferences. This has, in turn, kicked off a series of feedback loops in which political institutions (the media, Congress) and actors (candidates, individual journalists) adopt more polarized strategies to both respond and appeal to a more polarized audience, which further polarizes the audience, which further polarizes the institutions, which further polarizes the audience, and so on.

Social media is one of those institutions, and in my view, is clearly a polarization accelerant. In the coming years it may prove a primary driver. But the bulk of the run-up in American party polarization predates social media, which means social media isn’t core to the story.

Courtesy of Simon & Schuster, Inc. 

NT: That makes sense. Both things can be true: Technology can be polarizing us but also be a relatively small part of why we’re polarized. And the feedback loop you describe is one of the things I fear for the most when I think about the press. A newspaper that is relentlessly critical of Trump will have Facebook and Twitter accounts likely to primarily attract critics of Trump, which means that stories critical of Trump will get extra page views, which creates incentives to do more of the same. Or the trend can work in the opposite direction, with stories praising Trump. Our industry is governed by all kinds of incentives and norms. But I do worry about this kind of baleful feedback loop.

EK: There are two basic problems caused by this kind of polarization. One is everyone is constantly angry and agitated, and politics becomes something only those with a high taste for constant conflict can stomach. But the bigger problem is how polarization interacts with our political institutions, which require high levels of compromise to function. My concern isn’t that politics is argumentative or uncivil or divided. It’s that we are trapped in arguments we cannot resolve, because in our political system, winning an election rarely gives you the power to actually govern. In our political system, bipartisanship is necessary. With parties this polarized, bipartisanship, as I show in the book, is irrational for the out-of-power party to offer. So what we’ve done, in effect, is escalated the intensity of our political fights, but made it almost impossible for one side or the other to win, at least in terms of policymaking. That leaves the public trapped in a system where everyone is fighting but their problems aren’t getting solved.


NT: Your book has a section explaining that one of the key reasons for our partisanship is the decline of parties. When parties are strong, pragmatists tend to be nominated; when parties are weak, purists tend to be nominated. And social media has helped gradually pull power from the parties and to the candidates. (How many tweets have you read from @DNC and how many from @AOC?) The classic example of course is Trump, who couldn’t have been more anathema to the actual Republican Party, at least at the start. Is that a fair reading of the argument?

EK: Yes and no. Social media basically focuses attention on content that generates the most intense emotional responses. That definitely advantages louder, more provocative, more inspiring, more outrageous candidates. It disadvantages quieter voices who may be great at getting things done but, for exactly that reason, don’t want to root themselves in the conflict-oriented politics that dominate social media.

That said, a lot of pragmatists still get elected! It’s important, in the Twitter-is-not-the-real-world way, not to overstate how much social media has taken over nominating processes. So I’m fully on board with the thesis that social media is pulling power away from parties, but parties, and all kinds of other mediating institutions, still have a fair amount of power.

NT: That sounds right. And I don’t want to overstate the power of social media. If Twitter was the same thing as the country, Joe Biden would have dropped out in June. And the only possible way Trump could lose would be if he ended up having to face off against Baby Yoda.

You have a section on group identity in the book in which you point out that one of the ways that groups develop strong identities is by demonizing their opponents. I’ve always thought that this is one of the worst features of Twitter. It makes it very easy to find (or amplify) something stupid that someone who disagrees with you has said—and then to make readers believe that this particular bad thing defines your opponent, or even their entire political tribe.

EK: People keep saying Twitter isn’t real life. True enough. But it shapes real life. It is, as you say, a platform that brings out the absolute worst in people politically. It’s got the incentive structure of a high school cafeteria, coupled with algorithmic virality. It’s trite to say that, on Twitter, we see the worst of the other side, not the best. But that’s fractally true, too: The incentive to have a group and be cheered on by that group seems to create pressure to form smaller and smaller groups. So soon enough, we break into factions, and see the worst of the people we used to think of as on our side, too. Look at Democratic primary Twitter if you need an example.

If Twitter were really walled off from the real world, it’d be fine. But when people say it’s not real life, what they mean is it isn’t representative of mass opinion. But politics isn’t representative of mass opinion, either. Political elites have an outsized effect on what actually happens in politics, and they’re constantly on Twitter, living in its controversies and resentments and feedback loops, and they (we!) create a politics that looks more like Twitter even if that’s not what the country wants.

And to be clear: I include media in my definition of political elites. As I argue in the media chapter of the book, one of the ways Twitter is most powerful is that it sets the agenda for much of the media, because journalists, as a class, are horribly addicted to the platform, and as our coverage tilts toward the kind of politics Twitter rewards, politicians increasingly try to act that way to get coverage. Marc Andreessen once said that the amazing thing about his Twitter account was that it was as if he had set up a bullhorn broadcasting his thoughts into every newsroom in the country. I remember reading that and thinking he was right, and it was grim.


NT: Andreessen may have decided that it was grim, too, because he’s put his bullhorn on the shelf and decreased his Twitter activity by roughly 99 percent over the last two years. (And if he starts up again, please let me know! As far as I can tell, he is the only person on the service who has blocked me.)

Anyhow, you have an interesting line related to this, roughly halfway through the book, where you say, “The more political media you consume, the more warped your perspective of the other side becomes.” And then, shortly after that, you point out that following a few people of different beliefs is actually poison, not an antidote. It makes one more partisan. Explain why that is.

EK: I just looked and I’m blocked by Marc, too. Weird! At any rate: In the section you’re quoting from, I describe a study called “The Parties In Our Heads,” which surveyed people about the composition of the political parties. What proportion of Democrats are union members, African Americans, LGBT, atheist? What percentage of Republicans are evangelical, over age 65, Southern, or earn more than $250,000 a year? What they show is that misperceptions of who makes up the other party are everywhere, but they actually get worse as you consume more political media.

Which makes some sense: Political media focuses on the most engaging voices, not the most representative. Fox News covers “the Squad” way out of proportion to their power in the Democratic primary because the Fox News viewership gets very upset about these young, diverse, democratic socialist women. Meanwhile, Rep. Richard Neal, the chair of the uber-powerful House Ways and Means Committee, is basically unknown to the Fox viewership.

But it’s not obvious that being exposed to even representative voices on the other side moderates partisanship. I run through a study in which Twitter users were paid to follow more people from the other side. The result of that exposure was that the conservatives became more conservative and the liberals, if anything happened at all, became more liberal (but that effect wasn’t statistically significant). Persuasion is very hard to do, and mostly what seeing arguments on the other side does is make us defensive, which pushes us back towards our side.

NT: This isn’t the subject of your book, but I often wonder whether there is any democratic country in the world that has not become more partisan in the past decade. Western Europe has gotten worse; the UK is a disaster; India is not doing well. The one countervailing example I can think of is perhaps Japan. Do you have a hypothesis about whether this is a global phenomenon?

EK: I’m not sure it is a global phenomenon. It’s really hard to do good data comparison here, but I recently wrote up a study that tried to create a comparable, historical data set of trends in partisan polarization across nine countries, including the US. They found that party polarization had actually declined in five of the nine. Again, the data here is a bit murky so I wouldn’t bet my life on these results, but at the very least, it’s not completely clear that rising polarization is the norm.

NT: Can you think of any part of the modern internet that makes people less partisan? At WIRED, we’ve written about a subreddit called Change My View. And we’ve profiled people who spend their days trying to calmly confront political trolls and partisans. But those are tiny islands of hope in a raging ocean of obloquy. What else is there? If I had to come up with one major invention of modern technology that counters the trends you write about, I reckon I would choose Wikipedia.


EK: I think those are reasonable examples. But in general, it’s not hard to make people less polarized. Take them out of contexts that encourage zero-sum group competition, and put them in a context that encourages cooperation, or just brings out another set of values entirely. The problem isn’t that we are somehow tuned to be partisan, it’s that we have a political system designed around irresolvable, zero-sum power struggles between political coalitions with dramatically different ideologies and demographic compositions. The big point of the book is that we are in a system that incentivizes this kind of politics, and so this is the politics we’re getting. There’s no small internet hack or utopia that’s going to change that.

NT: OK. Let’s get to solutions. At the end of the book, you cautiously propose some routes to salvation. You suggest breathing and thinking while you tweet, which is an eminently good idea. And you’re for electronic voting, which I fully support, too. But you don’t mention the possibility of overhauling the algorithms of the major tech platforms to deprioritize outrage and pop filter bubbles. Do you think that’s possible?

EK: It seems possible to me! I offer a bunch of ways to change the overall political system at the end of the book, but as you note, I don’t really discuss social media—that, again, is because I’m not convinced social media is the key driver here. That said, I think it’s bad we’re constructing the informational commons of the future atop algorithms that select for the stories and comments that generate the most intense emotional reactions. I don’t think politics, or really anything else, is well-served by cranking the volume to 11, forever. But I’m not very optimistic that any of the social media companies will decide it’s truly in their interest to calm the passions on their platform, because that would likely reduce time spent, and harm their bottom lines. Capitalism is a definite contributor to all this. And nor am I optimistic that the government will regulate these algorithms out of existence.

NT: What about starting either a new social media platform—or even just a news aggregator—that is devoted to countering polarization?

EK: Lots of people have tried this. All of them have failed. Polarization is a successful business strategy because it’s what people who are interested in politics choose, because it reflects the reality of politics as people experience it. As I’m at pains to say throughout the book, polarization itself isn’t necessarily a bad thing, and the truth doesn’t live in the middle. When the differences between the parties and candidates are large, people are going to feel passionately about who’s right.

The actual recommendation I make towards the end of the book is to take more of your politics offline and into your community. National politics is a shitty form of entertainment. It makes you feel bad. It’s disempowering. If you want to be politically involved, get involved in making your community better. Vote for your ideals nationally, but spend your time working locally. If you’re reading this in the Bay Area and you spend your time tweeting about how bad Trump is, hook up with organizations working to make housing more affordable, and spend time doing that, instead. Working with other people on real problems actually does cut polarization for a variety of reasons, but it’s also just a more nourishing, empowering, enjoyable, and effective way to be involved.

NT: OK! Thank you very much. It was a great pleasure to chat with you here. Everyone please go buy the book.

When you buy something using the retail links in our stories, we may earn a small affiliate commission. Read more about how this works.

Read more: