The top-2 “primary” in California and Washington is a democratic abomination that violates the right to free association

As we lurch toward California’s “primary,” hoping the collective action problem voters, in a fit of mindless antipartyism, needlessly created eight years ago doesn’t bite them in the ass and give them two Republicans they don’t particularly want to choose from in the general election, it seems like a good time to make the case against this model from a more theoretical perspective. We’ve posted about the practical problems with this nonsense, and to some extent about the small-d democratic problems with it. I thought a post being a bit more systematic about this might be in order. As I see it, there are three core reasons to dislike it from a democratic perspective. I doubt the first two will be particularly controversial here; the third might be.

Perverse incentives (1)

The top-2 primary potentially punishes the party that has enthusiasm and momentum, in two ways. The immediate threat is simply because such a party is likely to inspire greater participation, which includes attracting more quality candidates. Participatory energy and enthusiasm is generally better for democracy than the alternative and shouldn’t be punished by the electoral system . But even if this threat is averted, and the party with majority support manages to get the top two slots, this robs the party of the ability to choose a candidate again, this time in the general election. In the Senate race, were Feinstein and DeLeon both to move through to the general and end up with similar levels of support, it throws the choice of which Democrat to elect to California’s Republican voters.

Perverse incentives (2):

The top-2 primary incentivizes meddling by the central party organization/party elites. Generally speaking, I agree with Scott’s post almost entirely. I don’t think there’s anything presumptively illegitimate about party elites having a preference and using influence and resources to promote it. The main reason *I* generally oppose it is that a) I’m not convinced it’s likely to be particularly effective, and that party elites are as good at determining ‘electability’ as they think they are, and b) many on-the-ground activists find it far more objectionable and illegitimate than I do, so such interventions risk antagonizing and inflaming, and widening the perceived gap between “party elites” and activists/voters for, in most cases, dubious potential gains, and c) even if (a) and (b) weren’t the case, it still seems a questionable allocation of scarce resources, relative to trying to defeat Republicans. In a normal primary, the case for staying out of it isn’t difficult to make, but in a 4-6 Democrats, 2 Republicans environment, the case for intervention becomes much harder to resist, while still carrying the same backlash risks as unnecessary interventions in normal primaries. It’s one thing to trust the voters to pick a candidate, it’s another altogether to trust them to solve a collective action problem with limited information when they’ve been deprived of the best method to do so.

Freedom of Association

This one gets me a lot of blank stares. I’m going to try to break down my thinking into a series of premises. I don’t present this as a compelling defense of my conclusion, so much as a road-map to how I get there, so people who are sure I must be wrong can figure out where our premises diverge.
Throat-clearing caveat: I am not trying to make a constitutional argument here. I have no idea where this would fit in current doctrine. All my knowledge of relevant cases is second-hand, via political theorists who write about this stuff. (From what I don know, it seems like the court’s holding in Jones would strike down the current system, too, but I don’t know the case well enough to make that claim.) I’d be equally committed to everything I say below if freedom of association were not constitutionally protected, because I regard it as a basic human right.

Premise #1: Political parties are properly understood as sites of free association

This one seems pretty obvious, but is sometimes contested or denied. It certainly *is* true that political parties are very different creatures than most voluntary associations, and the kind of limits placed on them are going to be different, and in many ways stricter and more limiting. But it’s a mistake to think freedom of association applies to a specific kind of group, with a specific set of free association rights (and limits on those rights). Freedom of association properly understood is not homogenous in details. It applies to different kinds of groups differently, while still applying to all of them.

Premise #2: Political parties appear to be indispensable to functional political democracy in the modern world.

Reasons: Too obvious to belabor. Citation: the discipline of comparative politics (passim). This is rarely directly resisted or denied, but antipartyism tends to avoid, ignore or minimize this truth. I suspect much of this some combination of the impulse to get the politics out of politics, and misapplied and/or excessive individualism, but whatever the source, it’s a mistake.

Premise #3: The right to exclude may be curtailed quite severely for political parties, relative to other groups, while they remain a site of protected free association.

One of the central features of free association rights for most groups is that have they have a fairly robust right to exclude, even for deeply illiberal and widely-held-to-be-illegitimate reasons. Political parties’ right to exclude is, even on our most expansive plausible interpretation, very limited—a simple willingness to quasi-publicly declare one’s preference for the party is all it takes to be included in closed-primary states. In open primary states, the declaration isn’t necessary.

Political parties are limited in their capacity to exclude precisely because of the unique role they play in democracies; they ought to be governed by democratic values and norms in ways other groups are not. This doesn’t mean they’re not an association and don’t have a purpose, it just means that exclusion plays a different, lesser role for them. (Exclusion also makes little sense as a practical matter, for reasons associated with the next two premises.)

Premise #4: Whatever limits we might place on groups, in order for their FA rights to be protected, those limits must not thwart their purpose (as long as that purpose is lawful and legitimate).

Voluntary associations are, for the most part, purposeful. We come together for concrete reasons, to accomplish specific things. Some groups have illegitimate purposes. Their freedom of association is highly attenuated. (The society for eating babies can gather to discuss how fervently they wish they could eat babies, but most actions directed toward fulfilling their purpose can’t and shouldn’t be protected by freedom of association.) At the margin, of course, which purposes are legitimate and which are not is a matter of law, which means it’s a question for the people to answer. If we do hold that a purpose is legitimate, we can still craft laws that limit their association, but if they thwart any hope of achieving the purpose, we ought to either revisit those limits or our decision this is a legitimate purpose.

The legitimate purpose of political parties is to win elections and govern. Making an affirmative argument for this legitimacy isn’t something I’m going waste your time with, beyond “see premise 2.” In order to do this, they need a selection mechanism that solves the collective action problem associated with internal factions. (This is all attenuated by our political and electoral systems and Duverger’s pesky law, but I think it’s still more or less true in parliamentary systems with PR. There are still strong incentives to not split off at every significant disagreement; factions will still be part of parties.)

Premise #5: The days of the smoke-filled rooms are, for better or worse, over.

I can imagine a clever argument that a political party could hold some sort of meeting or have an internal process to decide who “their” candidate is prior to the “primary,” and broadcast this information far and wide (but not on the ballot itself, where they’ll be marked with “prefers Democratic party” along with the other aspiring Democrats). In other words, a right to formally endorse is all that is required for freedom of association’s demands to be met with respect to elections—there is no difference here for the Democratic party, the NRA, and SEIU local whatever. There are two serious problems here: first, that this ignores the actual core purpose for which parties exist. The purpose of political parties is to govern. The purpose of interest groups is to influence government. Endorsement (along with other mechanisms) is sufficient for the latter, but not the former. If parties are to have the right to free association, it must be consistent with their purpose.

Second: parties are expected to be democratic now, internally. This may have been a mistake and it may still be one—I don’t share this view but it’s pretty common in political science. (See ch. 3 of Democracy for Realists, which I plan to write about soon-ish here.) But it doesn’t matter; it’s become a democratic expectation. I don’t think the principle of democracy demanded the internal democratization of parties, strictly speaking, but it’s happened in a way that’s changed demands and expectations; if the parties were to fully abandon internal democracy now, they’d be viewed as illegitimate. In our political culture, going back to something other than a broad election (or something similar) would be not be a democratically legitimate option. For better or worse (better, I say, with a fair amount of respect for many who disagree) democracy is part of how parties associate. I’m framing this as a theoretical issue about legitimacy, but for related reasons its also a practical one; parties telling people someone other than their preferred nominee is the *real* democrat here might be pretty ineffective.

Interestingly, California’s current terrible system exists because the Supreme Court did rule that previous primary system violated political party’s free association rights—In California Democratic Party v Jones (2000), SCOTUS , held that California’s initiative-created “blanket primary” violated free association rights. It seems to me the current system violates them far more clearly and directly. Under the blanket primary, There was one ballot for each voter, and they could choose on a race by race basis which party’s primary to vote with—so one could participate in the D primary for governor and the R primary for Senate in the same election. I’m sympathetic to the Party objection here, and I wouldn’t have voted for the initiative myself, but this seems far less objectionable on FA grounds than what we’ve got now. Without the blanket primary it was already the case that voters could flip back and forth, from one party to another at will, between elections. The blanket primary merely increased the velocity of potential flipping (and associated concerns with strategic voting, etc). It’s a difference in degree, but not kind, from the standard American party primary system. But each party still had a primary—the Democrat, Republican, and anyone else with the most votes moved forward to the general election. With less control over their membership than they were accustomed, the parties still got to choose. Now, the parties don’t get to choose at all. I don’t have strong views about this case, and have some sympathy for the basic arguments in the District court’s holding and Stevens’ dissent. But the arguments that carried the day in Jones seem far more compelling against the current system, which does far more damage to the free association of political parties.

Reading recommendation: clever defenses of partisanship are all the rage in political theory today, but I don’t think any of the recent ones surpasses Nancy Rosenblum’s On the Side of the Angels: An Appreciation of Parties and Partisanship. It’s historically rich, drawing on the history of parties and antipartyism primarily in US History. The focus isn’t on freedom of association per se here, it’s broader, but reading it clarified, challenged, and shaped my own thinking on parties, partisanship and related matters. Good review and overview here.