Futures Markets and the Election

The second issue of The Economist's Voice is out, and it contains a brief but interesting article Experimental Political Betting Markets and the 2004 Election. It discusses some contingent contracts that were created on Tradesports in order to test what market participants think about correlations between events (such as Osama capture and Bush re-election). So what does this mean?

Suppose we make a contract that pays off only if Bush is re-elected, and Osama is captured before the election. Gamblers call this a parlay. Now basic probability tells us that if the events are independent, the odds of the parlay are just the product of the odds of the event. If Bush is 60% to be re-elected, and Osama is 10% to be captured, the parlay should be 6%. However, what if the events are not independent? Surely Osama being captured increases the chance of Bush being re-elected. If the capture guarantees a Bush victory, then the parlay would be exactly 10%.

Because of this, the price of the parlay, compared to the price of the individual events, tells us what the market thinks about the correlation between the two events. Robin Hanson has suggested applying this technique in some interesting ways. For example, the government could gauge the effect of a policy on the economy by creating a market in contracts of the form "What will the GDP be if this policy is enacted?". The price of this contract will tell them the opinion of traders about how policies will affect the economy. So its a clever way to get people to price hypotheticals.

My futures market betting has gotten rather uninteresting, as my final position going into the election has very little variance. If Bush gets anywhere between 200 and 350 EVs, and the Democrats don't sweep House/Senate/Pres, I win all my bets. Only if some odd event occurs (?alien invasion?) to cause a Dem or Rep blowout will I lose. But I won't complain about the free money.

Share this