Monday, October 15, 2007

it's seems wrong that something with a 31.8% chance of not happening should feel like a foregone conclusion, but it does

Current market-based estimates of the probability of candidates winning the 2008 Democratic nomination:
Clinton: 68.6%
Gore: 11.5%
Obama: 11.2%
Edwards: 3.5%
So, Obama has fallen behind someone who has given no indication of running for President. It's becoming harder to imagine what that 1 in 3 scenario would be under which Clinton does not win the nomination.

I am no longer as pessimistic about Clinton's chances of winning the general election, although I'm not sure if this is just me being lulled into denial about how nasty the Republican negative campaigning against her is going to be.

Prediction markets, incidentally, have consistently failed to reflect the idea that Clinton faces a particular disadvantage over other Democratic candidates should she get the nomination. So while people such as myself like to opine that idea, it doesn't have traction among anyone willing to put money where their mouth is.


Chip said...

I don't play these prediction funds. Would it be straightforward to place a "win the nomination, lose the general election" bet?

jeremy said...

Chip: You can play all three of the necessary bets:

1. Clinton to win nomination
2. Democrat to win Presidency
3. Clinton to win Presidency

Ang said...

I was told the other day that "people don't dislike Clinton as much as you think they do." That may be, as I pretty much thought that no one wants her for POTUS. At any rate, I used to think a Dem winning the whole shebang was obvious. Now, I have no idea.

Brady said...

Ugh. If this is true, that means we've got a pretty good chance of having had the same two families holding the presidency for over twenty years (almost thirty if she wins twice, at which point Jeb will probably run...argh).

Does this strike anyone else as odd?

Chip said...

sorry, I was a little opaque there. I am wondering what the wisdom of these market says about the likelihood of Hillary her own self winning the nomination and losing the general. (I think that's my worst case scenario - because I am reasonably confident any other Democrat would win the general.)

Ken Houghton said...

Ang's got it. At the margin, no one wants to vote for Hillary, even though many are willing to do so.

If Election Day is predicted for heavy rain in OH, FL, etc., I'll bet against her in size--and I don't have the cash flows to justify that right now, and probably will be worse off in a year.

Ken Houghton said...

chip - the current intrade market has her at 45 last, with the market at 44.7/45.9.

Chip said...

So, maybe I've got this wrong, but it seems to me that conditional on her winning the nomination, the prediction markets are giving her a 2 out of 3 chance of winning the presidency.

2 out of 3 for the nomination sounds about right, but 2 out of 3 against the Republican given she wins the nomination seems way too high.

If I were to put my money where my mouth is, then, I would bet the entire field (Dem and Repub) against Hillary in the general.


jeremy said...


You can, you can bet against Hillary. It's as easy as betting for her.

Your estimate of the probability is basically right. The thing is, Tradesports also offers a contract on the probability of a Democrat getting elected (regardless of who it is). That has been trading around 63. So, if the market thought Hillary's chances were clearly inferior to other Democrat's chances, the probability of her getting elected conditional on her being the nominee would be lower than the probability of a Democrat being elected.

(Note, that her conditional probability is perhaps higher than the overall Democrat probability does not mean that people like her conditional chances better than other candidates because the other major Democrat candidates also have higher conditional probabilities. I think it's an irrationality in the market due to it not having enough volume, and there not being profit for someone to arbitrage it away since the general election payout is still more than a year away.)