Monday, March 21, 2011

The Power of Organizing, The Folly of Top-Two

A group of people working together can accomplish greater things than any one of them working alone. We see it all around us as people rise up against tyranny, or dig each other out from terrible disasters. Cooperation is a force-multiplier.

Which is why top-two primaries (or, as they've erroneously come to be called recently, "open primaries") are a doomed and useless gesture.

First, some vocabulary. If a few years ago you asked anyone what an "open primary" was, invariably they would have told you "A primary election that does not require voters to be affiliated with a political party in order to vote for partisan candidates," [Wikipedia] or something to that effect. But somehow, in the last year or two, the name "open primary" has been applied to what used to be known as a "top-two" or "jungle" primary. In this sort of primary, rather than each parties nominees appearing on separate ballots (and the winner of each party's vote continuing on to the general election), all the candidates of all parties appear on a single ballot, and only the first- and second-place candidates compete in the general election. While all "top-two" primaries are also "open primaries," it is not the case that all "open primaries" are "top-two."

The problem (make that "a problem," but we'll just focus on this one for today) with top-two is, if there are a large number of good candidates from one party (let's call them the "A" party), and relatively fewer good candidates in the other parties, then the A-party candidates could split the A-vote, potentially making it so that none of the A-party members get into the general election, even if one (or more of them) could have defeated candidates from the other parties in the general election. When I wrote about this last May in the lead-up to the vote on California's top-two primary proposal (Proposition 14), I said that "it would be really nice if we could have a party primary before the open primary." I was at least half-joking: such a response is absurd, meant to show that the original proposal is absurd.

Well, the absurdity passed, and now the absurd response has followed. The California GOP plans to hold a party primary in advance of the top-two primary in order to select their favored candidate. [Via Ballot Access News.]

Coordination is a force-multiplier. (So is repetition.) People will always seek to coordinate in order to maximize their effect on a situation. That's why we have political parties. That's why parties want to keep non-members out of the primaries. That the GOP has decided to go this route in California is a simple, logical, and inevitable response to the rules under which they operate.

Proponents of top-two try to claim that it would elect more moderate candidates, and would ease the stranglehold that the two major parties hold on the political system. It hasn't done so in Louisiana, nor in Washington, and it won't do so in California. But there is a change to the rules which we could make that would do those things: spoiler-free, consensus-seeking voting methods, like approval voting or score voting. These would actually change the incentives, for voters and for the parties they organize.


  1. What if Approval (or Score) Voting were used in the first stage to pick the two finalists? Perhaps that might make a "top two primary" more likely to produce more moderate candidates?

    If both parties were to choose their favored candidate prior to the "open" primary, it'd be a reversion to the state that existed prior to the use of primaries (w. the continued cost of a primary). And that wouldn't be all bad..., since all-in-all primaries haven't done that much to improve the US's democracy.

    But if the Republican party holds itself to the one candidate regimen, I don't think the same strategy would make sense for the Democratic party... So I guess there might be some uncertainty still as to who the Democratic party would nominate or maybe a third party might have a chance to be the 2nd highest in some cases.
    This might or might not produce candidates more fit for winning in the general election for the Democratic party.

    Fewer candidates probably would lower voter turnout in the general elections but I think it would depend on how close/exciting the elections were and how important the key differences between the two finalists were seen. No doubt the Main Stream Media won't be helping that much to make issues of substance more important in the election. But third party enthusiasts might be able to get their issues debated if they were willing to vote strategically on them.

    The damage from "top two" will depend on how we respond to it. It does appear to have kept the Palin and tea-bagger-supported senatorial candidate in WA off the general election ballot and thereby stymied the gains of the Republican party due to reducing the enthusiasm of tea-partiers in the last election.

  2. On further reflection if the Republican party did choose 1 candidate per election, it might behoove the Democratic party to choose 2 candidates per election...

    But you'd agree that these games would be radically changed if Approval Voting were used in the first stage of a top two primary?

    That might be a good target for getting folks used to Approval Voting.


  3. Yes, I think approval would be a useful thing for parties to implement in their primaries; or at least, would change the game. And I also think that it would be a good way to introduce voters to the idea of approval voting.

    The difficulty is that, in many states, the legal barriers to changing a party's primary rules (particularly of a major party) are almost as steep as changing the rules for general elections.

    On the other hand, the incentives are different: within a party primary, approval would get a better candidate for the party and *increase* its chances of winning the election; whereas in the general election, using approval voting involves the two major parties *giving up* part of their chance at victory. So that may make it easier to enact for primaries.

  4. All reformers need to ask themselves "Who'd benefit" and then get a significant subset of those folks to subsidize their campaigns...

    It also pays to pick our battles carefully with those in power.


  5. Interesting topic.

    Looking at the Wikipedia page differentiating the primaries, there are important distinctions.

    Traditional Open Primary: Pick any party's primary to participate in. Highest voted candidate within moves on.

    Blanket Open Primary: Vote among all candidates among all parties. Set number of candidates regardless of parties move on.

    Closed Primary: Vote within party. Party has candidate with most votes move on to general election.

    And then you have all sorts of laws spring up about whether candidates who lost in the primaries can run as independents (sore loser provisions). This often introduces the trickiness of a third candidate if there's not already a minor party or independent competing.

    But I see a major flaw with both the traditional open primary and closed primary. In both of those, the sample set of voters is from a section of the political spectrum not representing the voter population as a whole. Consequently, even with a good system like Approval Voting, you have the best candidate within a section moving on--not the best candidate of the entire voter population.

    On the other hand, a blanket open primary (such as used with top-two) samples from the entire voter population. And with Approval Voting, you get a good result for those moving forward.

    To me, I see the difference in these primary systems as the population or sub-population that actually represents the candidates moving on. To me, it seems more appealing for the candidates representing the entire population to move on. Of course, if we use Plurality to determine this, the results will be trash anyway. But I see this an appropriate role for Approval Voting. And if there were only two moving on, then voters would also get the appeal of a majority winner within the general election.

  6. Aaron,
    the rub is you gotta make the results of any primary indefinite enough to be interesting so as to get more voter turnout.

    Traditionally, the blanket open primaries have not been too reliable at drawing more voter turnout. Perhaps that might improve if Approval Voting were used so that the major parties no longer have the incentive to game the system by backing one or two candidates, but I think it's fair to say that the biggest problem with Top Two is how it denies voice to minor parties. For this reason, I would only possibly support a modified form of "top two" for "less local" elections that almost inevitably tend to get won by the candidates from the two major parties anyways...


  7. Third parties are going to get screwed every time Plurality is used. The issue is Plurality, not top-two. If a third party candidate can't gain enough support to make it to the general election under Approval, then they don't have enough support to win.

    Third Parties get almost universally ignored in the Plurality general election (minus horse race talk), so I hardly see how it gets much worse with top-two. Under Approval Voting, those candidates get an accurate reflection of their support, and they can't be marginalized by the media in the primaries. The media won't be able to ignore them when they have an accurate measure of support. It's perfectly reasonable then for third parties to be able to build momentum.

    So, I see the two conclusions as:

    1. Plurality distorts support and kills third party chances, making any negative effect of top-two overkill and hardly relevant. Approval Voting fixes the accuracy in measure of support.

    2. Primary nominations are not representing the voter population at large. Blanket open primaries fix this, so long as Approval is used. And there's the bonus of a[n absolute] majority winner within the general election given only two candidates move on.

  8. Sounds like we agree a bit here, Aaron.

    My approach is called Strategic Election Reform. Roughly, I argue for more multi-seated elections in "more local" elections and more "two stage" elections in "less local" elections. This will handicap the rivalry between the two major parties and foster a dynamic centrism that provides more protection for the rights of minorities, especially if there's a proliferation of LTPs or Local Third Parties that specialize in contesting "More Local" elections and vote strategically together in "Less Local" elections.

    As such, I'm okay with the use of better versions of "Top Two primary" in "less local" elections where the economies of scale for winning them already tilts things against third parties, regardless of what sort of options (FPTP, IRV, Approval, Cumulative Voting) are given voters.


  9. You bring up some good points, Aaron; I think I'll have to write a new post to address them!

  10. What do you think of this idea:

  11. (Man, good thing I have an RSS feed for comments; how do people even find these older posts? ;)

    To be blunt, ljp, I don't think it'll go to far.

    Please take a look at this older post:

    And also my latest post:

    We're kind of stuck with having just two major parties (as least for as long as we run our elections the way we do) and while the issues are important, focusing on finding the median-view for each issue in each constituency isn't going to be able to overcome that; there's just too much bound up with parties and issues. You need to free up the two-party stranglehold FIRST in order to free up issues. I don't think it'll work the other way around.

    Also, I think that the increasing level of communications actually HURT third parties: regionalism is one of the few ways 3rd parties have found any success, and easy national communications makes it easier for national parties to absorb any 3rd party movement.