Most political observers are familiar with this basic logic: to win elections, you run to the middle. There, in the ideological center, is where the swing voters are.
But, somehow, this doesn’t seem to happen. Candidates often appear to have views that are out of step with average voters. So the question is: Will this cost them votes?
Two new studies shed light on extremism and its consequences. Together, they show that presidential candidates are often much more extreme than swing voters — indeed, sometimes more extreme even than their own partisan base.
However, in American presidential elections at least, there is scarcely any penalty for being extreme. To put it bluntly: Candidates may be extreme because they can get away with it.
The first study, by Vanderbilt political scientist Larry Bartels, examines American National Election Study data from the nine most recent presidential elections (1980-2012). He compares swing voters — those who are likely to vote but are genuinely up for grabs — with the party bases, or those who are likely to support one party and also be active in politics.
Across a range of issues, the Republican and Democratic bases are more extreme than swing voters, and both parties’ bases have become more polarized over time. No surprise there. Bartels also finds that on most issues, the Republican base is more extreme (and has become more extreme), relative to swing voters, than has the Democratic base.
The surprises come in the behavior of the presidential candidates. Bartels measures the ideological positions of the candidates based on the perceptions of the most knowledgeable respondents in these surveys, after correcting for any biases in these perceptions due to respondents’ own partisanship.
Bartels finds that the positions of the presidential candidates are actually as extreme, if not more extreme, than their partisan bases — and have become more extreme over time. This tendency appears more pronounced among Democratic presidential candidates than among Republican presidential candidates.
Overall, then, Bartels finds little evidence that either Democratic or Republican presidential candidates mimic the views of swing voters or respond to trends in the opinions of swing voters.
So are presidential candidates punished for extremism? The second study, by Martin Cohen and three other political scientists, finds that they aren’t. Candidates that are arguably more ideologically extreme — such as Barry Goldwater or George McGovern — do not lose much vote share compared with more centrist candidates, once other factors are accounted for. Those other factors are important: Goldwater and McGovern did lose in landslides, but this had as much if not more to do with the fundamental conditions in the country, not with their own ideological positions.
The graph below is a good summary of how much more economic fundamentals (here, the percent growth in real disposable income) affect vote share than does candidate extremism.
At most, Cohen and colleagues find that extremism might cost a candidate 1-2 points of vote share, but even those estimates are not statistically significant. They conclude that there is “little evidence of an electorally important relationship between candidate extremism and vote outcomes.”
This does not mean, they note, that candidate policy views have no impact whatsoever — just ask former Republican congressman Todd Akin of Missouri. But it does mean that an American public that is not that ideological may not use ideology as a shortcut in voting for presidential candidates. And this in turn allows presidential candidates to have views well to the left or right of the average voter.