Monday, February 27, 2012

Even If It's Deadlocked, It Probably Wouldn't Be Deadlocked

I wrote about this a while ago, but a tweet just now reminded me: InTrade's current market for "brokered convention" is absolutely insane. It's now at 20%, which is approximately 19.999% too high. OK, I'll take seriously an argument that it's really only 19.9% too high, but it's somewhere in there.

Actually, this is sort of an excuse to make a point I've been meaning to add to this discussion about contested or deadlocked conventions. Here's the deal: InTrade defines "brokered convention" as a multiballot convention:
If the Presidential nominee is not decided after the first round of delegate voting at the party convention the convention will be considered "brokered" (i.e. it takes multiple rounds of voting by party delegates to decide the nominee).
But the thing is, even if we have the highly unlikely event that the normal delegate accumulation process doesn't yield a winner, it's still very likely that everything will be settled long before the convention. Not certain, but very likely.

Once again, there are two possible scenarios. One would be a contested situation, in which one candidate comes close close enough that he only needs to win over some of the unpledged or unbound delegates in order to get to 50% + 1. In that case, unless the candidate was clearly unacceptable to party actors (Ron Paul this year), it's almost certain that he would pick up enough of those delegates in plenty of time to have a normal, ordinary convention. Remember, the last event in the primary season is in Utah on June 26* -- but the convention isn't until August 27-30. That's two whole months, more than enough time to resolve it.

On the other hand, in the even less likely scenario of a true deadlocked convention, in which pledged or bound delegates to three or more candidates make it impossible for anyone to get to 1144 delegates...we'll still probably get there in time. Everyone in the party who is not specifically tied to a candidate is going to have a massive incentive to settle things as soon as possible, and the odds are that even in this situation, one of the candidates would concede defeat and drop out without it actually going to the convention. After all, most conventions during the last generation of the old mixed system (through 1968) were resolved on the first ballot, even though they were all "contested" by modern standards. So it would probably get worked out in advance.

Probably. But not for sure. There is a real possibility that everyone digs in and refuses to budge, and that there's real multiballot chaos in Florida. I think that's extremely unlikely this year; it's possible, but unlikely, for it to ever happen under current rules and procedures.

But the main point here is that whatever the odds are that things will not be resolved by late June, the odds are smaller -- and perhaps considerably smaller -- that they won't be resolved by August 27.

*The last event is June 26, but I don't really know when the very last delegates are selected; those states with caucuses or other multilevel procedures for selecting the actual people to go to Tampa will all have begun their process by then, but I don't actually know whether all of them will have finished. Not to mention that Texas could wind up going really late, as the date for their primary is still unresolved.


  1. Nebraska will select its delegates at its July state convention. That's the last one.

  2. One of the unfortunate changes InTrade has made recently (since the last election) is to institute a $5 monthly fee in place of transaction fees. This has basically gotten rid of everybody but the crazed gamblers and day-traders.

    The effect of the monthly fee is felt most strongly in contracts like this one, where the payoff comes several months in the future. If you wanted to bet against a brokered convention, you'd have to risk $200 just to make the $30 monthly fee back. Maybe there are some rational actors who think that's a good use of their money, but I think the results speak for themselves.

    This is something people don't factor in when talking about Intrade this year as opposed to previous cycles.

    1. Interesting point about inTrade. Still though, isn't the point of inTrade that if there are any inefficiencies, then the ruthless marketplace should be able to exploit them, and things converge quickly to the CW? And this makes me wonder, is there a history of political scientists investing in inTrade and making money?? So far, following this blog, I would have bet against Cain, Bachmann, et al, and would have profited handsomely.

    2. There is not, to my knowledge, such a history. Most of the folks I know tend to ignore inTrade. I was not aware of the new $5 fee, but I was going to go on there and short sell as many of these as I could (OK, $500 worth).

      The other thing I like to note about the whole "efficient market" hypothesis is that it requires a market. The gap between buyers and sellers on this question is decently big (at least, to my non-economist eyes): the cheapest anyone will sell (one) share is 24.8%, with 100 for sale at 24.9%. The most anyone will pay is 17.3% (48 shares), next is 10 shares at 17.2%. That's a gap of 7.5%.

    3. But, continuing on, now we can't pretend there's an efficient market. It's a collection of people WILLING TO PAY $5/month to gamble on these things. That's FAR from a bunch of people with good information, as Anon #1's little demonstration shows.

    4. >So far, following this blog, I would have bet against Cain, Bachmann, et al, and would have profited handsomely.

      You would have needed this blog to bet against Cain and Bachmann? What commentators were betting for them? (And I don't mean the occasional namby-pamby "Don't count Bachmann out" sort of post, I mean people who really argued that these candidates were likely to win.)

  3. If Santorum wins Michigan by a surprising (say 5+ point) margin, and also outperforms expectations in Arizona, would that change your assessment of the likelihood of convention chaos?

  4. I think the Republicans are more afraid of convention chaos than they are of Santorum. Unless a lot changes, Romney and Santorum are both well ahead of everyone else, so it would be one of the two.

    Ron Paul will probably try to bargain with the winner (most likely Romney) instead of going 3rd party because his son's career is on the line. And Gingrich will, having fallen behind, likely support anyone who can beat Obama.

    So there probably won't be a second ballot. We could still get a mess, though, if the Republican public doesn't support the winner by delegate math and bargaining.

  5. Not that anyone cares, but as a vocal former skeptic about the low probability of convention chaos, I've seen the light; the good professor is correct, there is realistically almost zero chance of a second ballot in Tampa.

    Suppose at the end of the counting, Romney's at 1100 delegates. Suppose you controlled about 50 Gingrich or Santorum delegates. The phone rings, and an ominous-sounding Romney operative offers you the chance to cough up your 50 delegates. You say heck no! and recite the litany of Romney shortcomings. There's silence, then the ominous voice says, "Suit yourself, I'll call the next person on behalf of soon-to-be President Romney".

    That's not going to happen. If Romney ends anywhere close to the 1144 needed, faithless Gingrich or Santorum delegate supporters will be tripping over themselves to curry favor by providing the delegates needed to send Romney over the top.

    Even in a deadlocked situation, where Romney's not close to 1144, the above-hypothetical phone call would still probably be effective, unless Gingrich and Santorum's delegates were entrenched the way Paul's are. If Romney ends with 700, while Santorum/Gingrich have 650 each, controllers of those Santorum/Gingrich delegates still risk making President Romney's blacklist without (as it now stands) the confidence that the controllers of the rest of their candidates' slate wouldn't fold like a cheap suit, given the fact that no one seems to legitimately like Gingrich or Santorum, anywhere near as much as they don't like Romney.

    So you'd need 1) a true deadlocked situation, and 2) much more public conviction for a Santorum or Gingrich than what we've seen thus far, which would be necessary for Gingrich/Santorum supporters to feel it was worth it to risk falling on their swords.

    Even then, you still wouldn't have a second ballot. That's just the unlikely circumstances needed to get the competition to the negotiation phase! Even if the two highly unlikely criteria above were somehow met, the party would still have two months to iron things out.

    1. But what if Gingrich endorses Santorum?

    2. This is certainly beyond my pay grade, but I'd say it shouldn't much change the low probability of a multiple vote convention.

      If the primary would otherwise be deadlocked, there's a pretty good chance Santorum+Gingrich would represent more than 50% of the delegates; if one sends his delegates to the other, the recipient would have a majority. Even if the combined is just less than 50%, the recipient is then playing the heavy on the Romney delegates, which should settle out very quickly.

      If its just contested, I suppose combining forces makes it a bit easier to fight Romney, but you still have the problem of each individual actor not wanting to tell Team Romney to pound sand, knowing that the next guy down the line will probably cave, and you'll have taken one for the team for nothing but personal pain.

      I guess if Gingrich and Santorum pool resources via endorsement the probability of a multiple ballot convention goes up a little, but it seems from here not very much.


Note: Only a member of this blog may post a comment.

Who links to my website?