Open Primaries and Duverger's Law
In an amusing piece on DailyKos yesterday, Kos notes how the Clinton campaign has an account of why each of the states that Obama has won doesn't count. Caucuses (that Obama wins) don't count because they're not really representative. States where Obama lived don't count. And states with large African-American populations don't count (because, you know, uhm, you can't count the votes of people who are your own race). Anyway, this is obvious self-serving nonsense (as would be a similar effort to disqualify Clinton's victories), but it does raise an interesting question: What are we to make of victories in open primary states?
Both Obama and McCain have benefited from the participation of independents in some of their respective party primaries, which is more or less the point of having open primaries. The danger of a closed primary is that the party faithful will pick a candidate that they like but who lacks appeal to the swing voters needed to win the general election. Permitting independents to vote moderates this effect, as the independents are turned off by party stalwart ideologues. This has worked almost exactly as predicted on both sides, with independents helping each party to pick the most electable candidate (assuming, as opinion polls have shown, that McCain and Obama are, respectively, the two parties' strongest candidates).
One complaint about open primaries is the possibility of spoilers. A conservative independent who plans to vote for the Republican nominee in the general can vote in the Democratic primary of the weaker candidate---or whichever candidate is trailing, so as to prolong the Democratic contest---in the hope of sabotaging the Democratic ticket. There is some evidence of this sort of thing happening (to and for Republicans and Democrats alike), but it is hard to say that this is a serious threat to the integrity of party primaries.
The more serious question is one of party self-definition. Recent Supreme Court precedents have treated political parties as expressive associations that have a right to control who gets a say in determining the party's standard bearers. If one wants to run with this principle, one could say that open primaries, even though constitutionally permissible, nonetheless violate the spirit of the parties' expressive association rights. Given the low barriers to joining a political party, why should people who aren't even willing to call themselves Democrats or Republicans get a say in picking the parties' candidates?
The answer, in my view, is Duverger's Law, which states that first-past-the-post elections lead almost inexorably (though not, apparently in Canada) to a two-party system. If it were very easy to find or start a party to one's liking, then it would be a sufficient answer to independents to tell them to buzz off during candidate selection time. But independents can't start their own party (yes, I know, the Whigs replaced the Federalists, and the Republicans replaced the Whigs, but that was a LONG time ago), so giving them a say in determining the shape of the general election choice will typically require that they be given a role in the major-party nominating process. Accordingly, the argument that the parties have rights of expressive association just like any other civil society organization seems quite wrong.
Posted by Mike Dorf
Both Obama and McCain have benefited from the participation of independents in some of their respective party primaries, which is more or less the point of having open primaries. The danger of a closed primary is that the party faithful will pick a candidate that they like but who lacks appeal to the swing voters needed to win the general election. Permitting independents to vote moderates this effect, as the independents are turned off by party stalwart ideologues. This has worked almost exactly as predicted on both sides, with independents helping each party to pick the most electable candidate (assuming, as opinion polls have shown, that McCain and Obama are, respectively, the two parties' strongest candidates).
One complaint about open primaries is the possibility of spoilers. A conservative independent who plans to vote for the Republican nominee in the general can vote in the Democratic primary of the weaker candidate---or whichever candidate is trailing, so as to prolong the Democratic contest---in the hope of sabotaging the Democratic ticket. There is some evidence of this sort of thing happening (to and for Republicans and Democrats alike), but it is hard to say that this is a serious threat to the integrity of party primaries.
The more serious question is one of party self-definition. Recent Supreme Court precedents have treated political parties as expressive associations that have a right to control who gets a say in determining the party's standard bearers. If one wants to run with this principle, one could say that open primaries, even though constitutionally permissible, nonetheless violate the spirit of the parties' expressive association rights. Given the low barriers to joining a political party, why should people who aren't even willing to call themselves Democrats or Republicans get a say in picking the parties' candidates?
The answer, in my view, is Duverger's Law, which states that first-past-the-post elections lead almost inexorably (though not, apparently in Canada) to a two-party system. If it were very easy to find or start a party to one's liking, then it would be a sufficient answer to independents to tell them to buzz off during candidate selection time. But independents can't start their own party (yes, I know, the Whigs replaced the Federalists, and the Republicans replaced the Whigs, but that was a LONG time ago), so giving them a say in determining the shape of the general election choice will typically require that they be given a role in the major-party nominating process. Accordingly, the argument that the parties have rights of expressive association just like any other civil society organization seems quite wrong.
Posted by Mike Dorf