John Sides: I was struck by this statistic early in the book: Even if voters had picked the most moderate candidate in every U.S. House election between 1980 and 2014, 80 percent of the polarization between Democratic and Republican members would have occurred anyway. Why is that important to know?
Andrew Hall: The point of that 80 percent statistic — which is based on an analysis that Adam Bonica first developed — is that there just aren’t a lot of moderate choices for House voters. If we want to understand where polarization is coming from, we have to understand why so few moderate people run for office.
In other words, we need to ask the question “Who wants to run?”
JS: You argue that it’s not voters who are asking for more ideologically extreme representatives. In fact, they reward more moderate candidates, correct?
AH: Right. And this is why the fact that so few moderates run for the House is important. If they ran, they might actually win!
Correctly estimating this is hard. More extreme candidates tend to run in districts where voters prefer more extreme representatives, creating selection bias. To start, I study very close primary elections between a more moderate and a more extreme candidate, which is like a randomized experiment in which some districts randomly get more moderate nominees and other similar districts randomly get more extreme nominees. Analyzing this “natural experiment,” I find that more moderate nominees perform much better in the general election than more extreme nominees. I then replicate this finding using other techniques. The advantage of more moderate candidates is strikingly consistent across these various approaches.
JS: Here’s one thing that struck me in that result. You note that voters often don’t know much about candidates’ policy positions and seem to lack the necessary information to reward or punish candidates based on ideology. So why do you think that more moderate candidates do better?
AH: One reason, as Dan Thompson and I documented in a recent paper, is that extremist House candidates mobilize opposing party turnout more than their own party turnout, on average. This could happen even if most voters aren’t aware of specific candidate positions.
Second, voters may rely on candidate attributes that are easier to observe. For example, in a previous paper, I found that extremists were more likely to be men. Maybe attributes like gender help voters figure out who is more moderate. Or maybe voters vote for more-moderate candidates because they like their other attributes. Voters don’t have to know the positions of candidates and intentionally vote for more moderate candidates. They just have to prefer voting for them, in general, for whatever reason.
Finally, as the political scientist Robert Erikson has long argued, even if most voters are uninformed, a relatively small number of informed voters can still swing many elections. Recent estimates suggest that something like 15 percent of House voters are swing voters, so this is quite plausible.
JS: You argue that more ideologically extreme candidates are more willing to bear the costs and benefits of office. Why?
AH: My basic argument is that more extreme people lose more from staying out of the race. A moderate candidate might look at the opposition and say, gosh, I don’t particularly like what this person stands for, but running is super hard and I’d rather not quit my job. A more extreme person might look at the same opposition and say, wow, that person is anathema to me, I can’t bear the thought of being represented by that person. Even though it’s hard for me to run, I have to do it.
JS: How have the costs of running changed over time? Why might those changes help create polarization?
AH: Campaign finance is a big factor. Candidates are asked to spend an enormous amount of time — four or more hours a day! — “dialing for dollars,” and they really dislike it. Congressman Dan Lungren told a congressional hearing, “I hate raising money for campaigns. The only two people I know who enjoyed it both went to prison.” The problem is that they have to do it, because they know their opponents are doing it. It’s a race to the bottom that leaves everyone worse off.
When you see how relentlessly demanding campaign fundraising is, you start to ask yourself: Who in their right mind would want to quit their job and go through this? And that’s how these changes can help create polarization: A lot of people aren’t willing to go through it, and the people who are willing to tend to be more ideologically extreme.
JS: As the costs increase, you argue that the benefits have decreased. Maybe we can start with this graph of U.S. House salaries.
AH: Yes! Perhaps understandably, the public has not been enthusiastic about our elected officials raising their own salaries. So, after adjusting for inflation, salaries for legislators have fallen noticeably since the 1960s.
In the book, I find that when state legislatures raise their salaries, there is a decrease in how polarized the set of people who run for office is. So the decline in House salaries may be lowering the incentives for more moderate people to seek office.
Another big thing that has changed in the House is that party leadership controls more and more of the legislative process, which makes holding office less attractive for individual people who want to influence policy. Here I am leaning on excellent work by Danielle Thomsen.
JS: So if reducing polarization is the goal, what reforms might work?
AH: I’m fairly cautious in the book, because I think that reducing polarization carries with it a lot of trade-offs that make it hard to be sure it’s the right goal. Right now, polarization feels like a big problem, but previous eras with less polarization were no picnic. In fact, an American Political Science Association committee report in 1950 declared that the problem in American politics was too little polarization!
One idea is to make running for office easier by reducing the burdens of fundraising. But even this is tricky. Reforms like public financing or bans on corporate contributions might make running easier, but they might increase polarization by favoring small donors with extreme viewpoints. I suspect a limit on how much candidates can spend would be far more effective, but that would require a constitutional amendment.
Finally, there are actions that require no political reform. For example, With Honor is an organization that aims to help politically moderate veterans run for office in an effort to reduce polarization. In the language of my book, you can think of groups like this as subsidizing some of the costs of running for office to attract more moderate candidates who wouldn’t otherwise seek political office.