orgtheory.net

whoops

with 10 comments

10 Responses

Subscribe to comments with RSS.

1. You’re pointing out an interesting contradiction in the prediction market (an inefficiency if you will). If you looked at all of the individual Senate races on the morning of election day, you would have come to the conclusion that the Senate would go to the Dems. They had Webb beating Allen and Tester beating Burns before the polls started closing. But yet, the contract for the GOP Senate victory was trading at a higher price than the contract for a Dem Senate victory. (And the GOP Senate/Dem House contract was well above .50 for most of the day.) So the market was obviously out of synch. An arbitrageur who noticed the discrepancy should have been able to make a pretty profit.

   brayden

   November 10, 2006 at 6:50 am
2. will it be about….performativity, by any chance?

   Peter

   November 10, 2006 at 1:55 pm
3. Brayden –
   how would you have done this? I’m just noodling around with the combinations, and I’m wondering if there is more than just an ‘arbitrageurs would have corrected this’ assumption. Given, say, Webb trading at 55, Tester trading at 55, and GOP control trading at 55, how would you arbitrage this, given that the end result will be either 100 or 0?

   Possible it’s just early AM and my brain ain’t on yet…

   Peter

   November 10, 2006 at 2:02 pm
4. Well, I doubt you could actually do this in Tradesports because the infrastructure is fairly simple. But in a stock market you’d sell short the GOP Senate contract and go long on Tester and Webb (and every other Senate race that you thought would go Democratic). I think that’s how you’d do it. If Brad DeLong is hanging around, he might correct me on that.

   brayden

   November 10, 2006 at 2:16 pm
5. But was the market really out of sync? For the Dems to take control of the Senate, they needed to sweep all four ‘uncertain’ races. If I’m correct, all four were trading around 55 to 70 in favor of the Dems. The joint probability of all four going to the Dems (and hence, the Dems controlling the Senate) would be around 15-20 percent, which is pretty much where the market was.

   Jerry

   November 10, 2006 at 4:47 pm
6. Peter – the fact that Tradesports doesn’t have the infrastructure to shortsell contracts is probably evidence supporting a performativity argument. Markets don’t naturally arrive at efficient outcomes (you might argue) without specific designs put in place by financial engineers. Not sure if that’s effective or Barnesian performativity.

   Jerry – I hadn’t thought of it in that way. I suppose though that if you were an outsider, with little information about any of the individual races, you could still recognize the opportunity for arbitrage if you perceived that the prices of the Democrat candidates over .50 were good indicators of election outcomes.

   In my thinking, I was dichotomizing the outcome (>.50 = victory &

   brayden

   November 10, 2006 at 7:22 pm
7. “It seems to every administrator that it is only by his efforts that the whole population under his rule is kept going, and in this consciousness of being indispensable every administrator finds the chief reward of his labour and efforts. Whle the sea of history remains calm the ruler-administrator in his frail bark, holding on with a boat-hook to the ship of the people and himself moving, naturally imagines that his efforts move the ship he is holding on to. But as soon as a storm arises and the sea begins to heave and the ship to move, such a delusion is no longer possible. The ship moves independently with its own enormous motion, the boat-hook no longer reaches the moving vessel, and suddenly the adminstrator, instead of appearing a ruler and a source of power, becomes an insignificant, feeble man,” – Leo Tolstoy, “War and Peace.”

   Rumsfeld

   November 11, 2006 at 12:42 am
8. Somehow my earlier comment was cut off. What I meant to say was that if you dichotomize the outcomes, it makes sense to think of a GOP victory being highly unlikely. But if you take the products of the prices (interpreting them as probabilities) of all of the close GOP Senate races, you’d still get a lower overall GOP probability compared to a Dem probability of taking the Senate. I’m too lazy to go do the math now, but it only makes mathematical sense that the GOP price was trading much too high.

   brayden

   November 11, 2006 at 6:02 am
9. “Senate races, you’d still get a lower overall GOP probability compared to a Dem probability of taking the Senate. I’m too lazy to go do the math now, but it only makes mathematical sense that the GOP price was trading much too high.”

   This is not the case, the pricing was perfectly internally consistent (and selling short GOP.SEN is definatly posible in trade sports). The way it worked is that the dems needed to pick up all 6 chairs, so asuming they are indepedent events and where to win all 6 with prob 0.8 youd still only have a 0.26 (0.8^6) change of taking the Senate. ofcourse the independence asumtion here is wrong but i think it makes it easier to ilustrate the point.

   econgeek

   November 11, 2006 at 10:51 pm
10. You’re right. I was forgetting that the GOP only needed to win one of those tight races to maintain control. My bad.

    brayden

    November 12, 2006 at 7:16 am