Skip to main content

Why Open Primaries Won’t Change Our Politics Much

Legislators elected from closed primary systems are no more or less extreme than those from open primary systems.

By Seth Masket


(Photo: Jonathan Gibby/Getty Images)

One of the reforms Senator Bernie Sanders and his supporters have been pushing in recent months is open primaries. Many argue that more open primaries, allowing independents to participate, can produce more moderate nominees that are more representative of the electorate as a whole. Evidence suggests, however, that this doesn’t really happen. Here’s why.

The logic of the open primary is pretty straightforward. Under a closed primary, only people who are registered party members (usually for some time) are permitted to vote. Those party registrants tend to be die-hard partisans, and the candidates they pick will tend to be from the ideological extremes. Independent voters, who might legitimately want a more moderate set of nominees, are forbidden from participating. Allow them in, and you end up not only with more moderate nominees, but nominees who recognize it’s in their interests to keep moderate independent voters happy while they serve in office.

Eric McGhee, Boris Shor, Nolan McCarty, Steve Rogers, and I tested this assumption in a large-scale study a few years ago. We looked at two decades of voting behavior by state legislators across all 50 states, and we compared legislators based on the type of primary system that nominated them. Quite a few state parties have changed their primary rules one way or another over this time period, allowing us a good deal of leverage on the question.

People unaffiliated with a party tend, on average, to be less interested in politics and less likely to vote.

What we found was somewhat surprising. Legislators elected from closed primary systems are no more or less extreme than those from open primary systems. There are a few very modest effects California’s experience with a blanket primary system in the late 1990s actually may have moderated legislators slightly, for example but the overwhelming finding is one of no effect at all.

Why might this be the case? Why would the ability of independents to participate in a primary have no effect on the sorts of candidates the primary produces? There are a number of possibilities.

For one thing, Donald Trump’s nomination notwithstanding, parties tend to be pretty good at picking the sorts of candidates they like and giving them advantages in primary contests. Party leaders’ strategic allocation of funds, endorsements, expertise, etc. can help them choose the types of candidates they prefer and help those candidates win in the primary.

Second, even if independents are allowed to participate in primaries, that doesn’t mean they will. People unaffiliated with a party tend, on average, to be less interested in politics and less likely to vote. There are rarely enough of them who actually show up to tip the balance in favor of a different set of candidates.

Finally, those with weak party attachments who nonetheless still wish to participate in the selection of party nominees will actually react strategically to primary laws, as Barbara Norrander and others have found. Imagine, for example, a center-left voter who is not registered as a Democrat but still wishes to vote in her state’s primary. If she’s in an open primary state, she can participate without difficulty. If her state closes the primary, but she still wants to vote in it, she’ll need to register as a Democrat. By doing so, she makes the state’s Democratic voters marginally less extreme; she’s adding her moderate voice to those of many hardcore liberals within the party’s primary electorate. So somewhat perversely, the closing of primaries can lead to party registrants on the whole becoming more moderate. And you still get roughly the same outcome in the primary election.

Now, to be sure, drastic changes in primary laws can have some effects on what sorts of candidates get nominated. Top-two style elections in California and Washington, for example, are occasionally producing run-off races between candidates of the same party (such as in California’s United States Senate race), and sometimes the more moderate candidate wins. And it was no secret that Sanders did better this year in open primaries, although that’s hardly an argument for open primaries’ moderating influence.

But even if many states come to embrace open primaries in the next few years, we shouldn’t expect much of an impact on voter turnout, election results, or the type of candidates who get nominated as a result.