Explaining political polarization in legislatures

A new statistical model developed by Mattias Polborn, professor of economics and political science at Vanderbilt, and Harvard political scientist Jim Snyder sheds light on some causes of polarization in American politics.

“Today, the parties seem barely able to talk to each other, and that’s actually very difficult to understand given one of the fundamental models that would suggest that competition drives the parties together,” Polborn said. “It holds that candidates compete for the swing voters in the middle, and so even if the parties want to pursue more extreme policies, the competition makes both parties choose middle-of-the-road policies. Clearly, this is not happening, and the question is why.”

Polborn and Snyder’s paper, “Party Polarization in Legislatures with Office-Motivated Candidates,” appears online at the Quarterly Journal of Economics.Safe seats drive polarization

One of the big reasons competition doesn’t drive candidates toward the middle is because most elections simply aren’t that competitive to begin with, Polborn says.

Polborn says that’s because during legislative elections, voters recognize that not only are they voting for someone to represent their interests, they’re voting which party’s legislative caucus to increase – which voters understand has a greater influence on national policy than any individual legislator.

Polborn gives the example of a Tennessee Democrat running for Senate who breaks with her party on a few issues important to conservatives, like gun control or lower taxes. Because Tennessee generally skews to the right of center, she should theoretically have broad appeal because she shares many points of agreement with the state’s moderate independents and Republicans while remaining more competitive among Democrats than a traditional Republican.

However, even if this particular conservative Democrat more closely mirrors the values of most Tennessee voters, Polborn said, “They think, ‘If we elect more Democrats, then not only does our Democrat go to the Senate, but all the other Democrats become more powerful.’ This is something the Democratic candidate in Tennessee cannot get rid of, and it makes the election less competitive” said Polborn.

But party isn’t everything

Of course, in some districts, voters may elect a legislator who doesn’t align with the district’s predominant political leaning because those voters place greater value on a positive non-ideological quality, like competence, experience or even just likeability. For example, Polborn said, a majority-Republican district that places a high value on the pork-barrel spending in their district may prefer a Democratic candidate if they believe he or she is better able to protect or increase that spending.

The more voters care about these non-ideological qualities, the more likely they are to vote across party lines. And because those legislators hailing from ideologically unfriendly territory need to maintain the good will of the swing voters from the other party to keep their seats, they tend to be more moderate voices within their party in the statehouses and Congress. “Wave” elections moderate the winners, radicalize the losers

In a Republican or Democratic “wave” election, many districts shift party allegiances at once. Polborn and Snyder find that the incoming legislators in the newly dominant party tend to adopt more moderate positions than the average legislator from their party, for much the same reason that legislators in the previous scenario do: They depend on crossover votes to stay in office. Meanwhile, the newly minority party tends to reconfigure in a more polarized position because their moderates were the ones who lost their seats to the other party. The researchers also find that this effect is considerably more profound in Democrats than Republicans.

“One would hope that losing an election would induce soul-searching and moderation in the losing party, in order to be more competitive next time,” Polborn says. “But, instead, the extremists are more likely to survive and to assume leadership positions. It’s individually optimal for legislators, but very bad for voters.” A snowball effect

Moderation begets moderation, Polborn said. When most districts are moderate--and therefore competitive--their legislators reinforce moderation in their respective parties. And when both parties are moderate, it increases the likelihood that voters from even more districts would consider voting across party lines, which in turn raises the competitiveness of even more elections.

Unfortunately, the reverse is also true. “As parties become more polarized, most people feel more strongly about which party's position is better, and are therefore less likely to cross over and vote for a candidate of the other party,” Polborn said. “So, in a way, the polarization of the parties makes voters behave as if they were becoming more extreme, even if that is not true—that is, even if most voters fundamentally have the same moderate preferences as three decades ago.”

This is why gerrymandering can be so destructive, Polborn explained. “It directly creates less competitive districts in which more extreme candidates will win, but it may indirectly affect even districts that were not gerrymandered because it generates more polarization in the system, and makes voters less likely to go against their ideological preferences—which again leads to more polarization. It’s a vicious cycle.”

Journal Link: Quarterly Journal of Economics