July 27, 2004

Crooked Timber Financial Newswire

Posted by Daniel

You haven’t seen it reported elsewhere, but on the Iowa Electronic Markets”, Kerry overtook Bush a couple of days ago. I don’t know if this is a “convention pop”; needless to say it would be pretty bad news for market efficiency if it were. In related news, the Kerry vote-share contract I bought a while ago is now back into profit, and I am still long. If/when I can be bothered reproducing the files, I will update my system’s equity curve - to be honest it doesn’t look that great, although one might argue that the system did the right thing in keeping me long.

Posted on July 27, 2004 06:23 PM UTC
Comments

How can you rate or de-rate the “efficiency” of a market in something that has no priceable fundamentals? No assets, no liabilities, just guesses, or should I say, just some bets placed.

It’s a pure probability estimation, correct? If the probabilities change, the price ought to change. Can you somehow discount the change in probabilities induced by convention bounce? I’d like to know how that’s done. A well-managed convention surely WILL influence the final result, won’t it? Otherwise, why have conventions at all? They are time-consuming and expensive.

Seriously, I am not trying to be snide, and I am not an economist. I really do wonder what you mean.

Posted by Ralph · July 27, 2004 07:33 PM

If I remember right, Daniel feels that (certain?) polls do a better job of predicting outcomes than markets, and he’s going to try to prove it (or at least provide strong evidence for it) by betting on the IEM.

If his model is more accurate, he can hopefully buy and sell depending on deviations from his model.

If the convention bounce is a short term phenomena, then he’s right to discount it. It may be that the convention did a “better than expected” job, in which case the move would be significant.

Posted by Jason · July 27, 2004 07:42 PM

Isn’t the convention bounce a negative number right now?

Does this mean all the people whose only justification in claiming “Bush will be re-elected” has been the IEM will find some tea leaves?

Posted by Ken Houghton · July 27, 2004 07:48 PM

Looking at the graph more carefully, the flip appears to have occured on 21 July. So it’s not a convention-blip; it’s more likely the result of people reading Chapter Eight of the 9/11 Commission Report. (New information moves markets.)

Posted by Ken Houghton · July 27, 2004 07:52 PM

The convention is a public, scheduled event. It ought to therefore be anticipated and there should not be a contemporaneous movement in the price. Otherwise there would be a simple profit opportunity; buy before the convention and sell into the bounce. In fact, I don’t think that the big movement in the KERR and DEM contracts were convention pops, but if they were, this would imply that the IEM was not semistrong form efficient.

Posted by dsquared · July 27, 2004 07:55 PM

On tradesport, Bush victory is 50.5 bid.

I don’t think either market is efficient.

Posted by Andrew Boucher · July 27, 2004 07:59 PM

Hey Crooked Timber, now are you going to promote Atrios to one of your academic sections now that we know he’s an economics professor?!

Posted by walrus smith · July 27, 2004 09:06 PM

Also note that on tradesports, all of the state-by-state contests have gone into big, big leads for one side or the other save 3: Nevada, Ohio, and Florida. And when you add those numbers in your Electoral College calculator, you find that Nevada doesn’t make a bit of difference and Bush needs to take both of the others to win.

(So I suspect that there’s some internal arbitrage to be found on tradesports…)

Posted by Jeff R. · July 27, 2004 09:51 PM

Jeff, why not go ahead and find it, then profit from it. Tradesports has a fairly hefty margin, so you’d need to be a little cautious, but it would be a good thing for you if you could grab it.

I wonder if the predictions are close to right. I had a quiok look, and there were a few other states in the 25-75 range which I would hardly discount as sure things either way.

An interesting thing to look at is the popular vote numbers vs the electoral college numbers. Is it really that likely that Kerry will win one and lose the other?

Posted by Jason · July 27, 2004 11:13 PM

Looking at current polls and predictions, I’d say there’s a pretty significant chance that Kerry will get more of the total “popular” vote but lose in the electoral college. If this does happen again, I wonder if it will motivate more people to push for changes to the electoral system.

(While a constitutional amendment is unlikely for now, individual states can at least change the way they allocate electors, as in Maine and Nebraska.)

Posted by Matt Brubeck · July 28, 2004 01:36 AM

This will be a good test of the “academics are completely incapable of making money in financial markets” theory, e.g. Long Term Capital.

Posted by Thom · July 28, 2004 06:13 PM

I’ve been travelling. The places I visited have changed since I last saw them. Upon my return I find that the same holds for home.

I like this reality a little bit better than the one I left. More like this, please.

Posted by bad Jim · July 29, 2004 08:46 AM
Followups

This discussion has been closed. Thanks to everyone who contributed.