Friday, October 28, 2005

Betting on The Next Supreme Court Justice (redux)

It's time to look once again at prediction markets for the next Supreme Court Nominee.

As I've posted repeatedly, I'm a BIG fan these markets. For the unitiated, a prediction market (typically, but not always) trades "futures" contracts. These contracts pay $1 if a given event occurs (and $0 otherwise). Some simple math shows that the fair price for this contract is the buyer's (or seller's) subjective probability as to the event occurring.

Prediction markets are interesting tools for aggregating dispersed information. Let's see how they might be helpful (at least in theory) in helping to predict the next Supreme Court Nominee. In this case, I'll use the contracts currently trading at Tradesports as an example. To see the current contracts trading on this market, click here. Note: Information changes, and so will the prices for these contracts by the time you read this. The illustration I'm using here is as of 2:40 p.m.

As of the time of this post, the market has put the highest valuation on the contract for Samuel Alito (last traded at 27). This contract opened the day at 11.9 (reflecting a subjective probability of Alito's nomination of 11.9%). After the market opened, the marginal trader thought that the contract was undervalued at that price. In other words, this trader thought that the chance of Alito's being nominated was greater than 11.9%. So, he would put in a buy (called a "bid") order at above 11.9. As long as the next "marginal" trader" believes that the probability of Alito being nominated was greater than the current price, he would put in another bid order, which would push the price higher. This would continue until the marginal trader thinks that the price exactly equals his assessment of the probability of the nomination.

So why does this help aggregate information? It all rests on the idea that at least some individuals have superior information and are willing to trade on it. If the informed trader (say, someone who works in the Bush Administration, or has their ear to the Washington grapevine) has a good idea that Alito would be nominated, they would start buying if the price was too low (i.e. below their assessment of the probability of nomination). If they thought that the price was too high (i.e. above the price that reflects their assessment of the probability), they'd sell the contract.

This trading would take place until the marginal trader thought that the price was "fair" (i.e. reflected their individual assessment).

For those wanting to learn more, there's an extremely very comprehensive collection of materials on prediction markets at Chris Masse's website here.