Should primaries be open?

To many people it seems self-evident that the purpose of a primary election (or caucus) is to let the members of a political party choose their party’s nominees. If so, it pretty directly follows that primaries ought to be closed to voters who aren’t registered party members. But if the case for closed primaries is that straightforward and obvious, why are so few of our primaries closed?

For more on that see this post from yesterday. In brief, the majority of party primaries in the U.S. allow at least independents to vote alongside party members, and in many cases voters can pick which party's primary to vote in when they cast a ballot. (They can't legally vote in more than one party's primary in a given election cycle, however.) A few states have non-partisan primaries for state-level offices.

Primary elections were introduced in the early 1900s with the intent of making elections more democratic by taking the choice of candidates away from political party organizations and turning it over to the public. One reason is that our system of government leads to a two-party duopoly (we’ve never had more than two viable major parties in our history), and as with similar concentrations of power in business, it makes sense to regulate matters in the public interest.

Primaries are paid for, regulated, and run by the states as part of our multistage process of choosing elected officials. (Presidential nominating caucuses, however, are generally paid for by the parties.) In the minority of states with closed primaries, unaffiliated voters are effectively disenfranchised in the first round of voting, even though their taxes pay for the primary elections they’re not allowed to vote in.

A pragmatic argument against closed primaries is that primary turnout is often low and tilted toward voters with the most partisan views. Semi-closed or open primaries can bring in a somewhat broader spectrum of people, potentially increasing the likelihood of nominating candidates with broader appeal and electability. (Of course, whether that’s good or bad is a matter of opinion. There’s currently a movement in the Republican Party in favor of closed primaries in the hope of nominating more solid conservatives as opposed to so-called RINOs -- Republicans In Name Only.)

The fact that the not-very-moderate Bernie Sanders did rather better in open primaries than closed ones is an obvious exception, but it no more disproves the general tendency than a snowstorm disproves global warming. (And Sanders also did better in caucuses, which as a rule have even lower and more partisan turnouts.)

A concern with open primaries is the risk of “party crashing” -- adherents of one party crossing over to vote in the other party's primary in a strategic effort to damage the other party’s general election chances. The most famous example of that was Rush Limbaugh's "Operation Chaos" in the spring of 2008, in which the right-wing talk-radio host urged his listeners to vote for Hillary Clinton in the later primaries in an effort to prolong the Democratic nomination battle. But available data indicate that only a small percentage of Republicans crossed over to vote in Democratic primaries, and it's not clear that very many of them were motivated by Limbaugh. McCain had become the presumptive GOP nominee by March 4, so they might simply have wanted to cast a presidential primary vote that had some effect on the outcome, and those who did vote for Clinton had other reasons besides shenanigans for doing so, including an impression that she was more moderate than Obama. The effect of Operation Chaos was probably greater than zero, but not much greater. In any case, Limbaugh called it off after the May 6 primaries, saying he wanted Obama, whom he viewed as less electable, to win the nomination. Of course, by then Obama has a large delegate lead and was generally expected to take the nomination anyway. Perhaps for an encore Limbaugh could ask his listeners to throw ice cubes into their yards at the start of winter and then take credit for falling temperatures.

Other real-world examples of attempted party crashing are hard to find, one likely reason being that voting in the other party’s primary means giving up a chance to vote in one’s own.

What about non-partisan primaries? A lot of people would like expand the use of non-partisan top-two jungle primaries like those in California, Washington, Louisiana, and Nebraska (at least for electing members of the legislature). But there’s a problem with top-two primaries.

Suppose the great majority of voters in a district prefer Party A and hate Party B. A lot of good Party A candidates enter the primary while much-less-popular Party B draws only two. If Party A voters divide their support across their large field, Party B’s candidates might well finish in the top two spots in the first round of voting, even if their combined percentage of the total vote is much smaller than for the larger number of Party A candidates. This could force voters in the general election to choose between two candidates most of them hate.

There’s fortunately a simple solution that’s trivial to implement: Use what’s called approval voting in the primary, meaning voters can vote for every candidates they deem acceptable, however many that might be. The winners of the primary would be the candidates deemed acceptable by the greatest number of voters. Oddly, the only place I’ve ever seen a proposal to use approval voting in a top-two primary is a previous post on this blog. I don't mean to suggest I'm the first person to think of it, only that I've never encountered a discussion of it. There's more on approval voting in general here. For more on top-two voting from advocates of the system see

(Another option is ranked-choice voting, which typically eliminates the need for a runoff. It's usually implemented via an instant-runoff system, though a possibly better alternative is the Borda count, which assigns more points to higher-ranked choices. A problem is that relative to approval voting, ranked-choice voting is in practice complicated to implement.)

However the voting system is designed, I support the right of voters to change their political affiliation at any time, including just before they cast a ballot. For that reason I'll support open primaries until I see some reason to think that they are harmful to democracy or lead in actual practice to a worse choice of candidates in the general election.

I do sympathize with the notion that a party’s members should be able to choose their party’s nominees without outside interference, and in a multiparty democracy I think that’s probably how things ought to work. But given our two-party duopoly, I think other concerns are of greater importance. And in practice, the small number of closed-primary states suggests that even the parties themselves don’t demand closed primaries.

[Updated 2017 March 8 to fix some typing errors and maybe improve clarity.]

Facebooktwittergoogle_plusredditpinterestlinkedintumblrmailby feather


Should primaries be open? — 1 Comment

  1. Pingback: Open versus closed primaries in the United States | D Gary Grady

Leave a Reply

Your email address will not be published. Required fields are marked *

Comments are moderated, which can take up to a day (rarely even two), so please be patient. I welcome agreement, disagreement, and corrections on anything from substance to spelling. I try to weed out spam and anything defamatory or pointlessly insulting (to anybody), unless of course I think it's really funny.