A friend of mine who bets on Intrade (he made money correctly betting that Rod Blagojevich would survive into this year) alerted me to the fact that Intrade now has a market for whether the U.S. will go into “depression” in 2009 (warning: that link will resize your browser window). Their definition of “depression” is “a cumulative decline in GDP of more than 10.0% over four consecutive quarters,” but they don’t really mean that. What triggers the payout is if the sum of the quarterly annualized GDP growth rates for four consecutive quarters is less (more negative) than -10.0%. (To see the difference: GDP in Q3 2008 was 0.13% smaller than in Q2 2008, but this was reported as an annualized rate of -0.5%.) This would mean that the total economic contraction over those four quarters would be more than (about) 2.5%. This would make the current recession the worst since at least 1981-82 (which had a total peak-to-trough decline of 2.6%), but not necessarily anything that anyone would call a depression.
On to the interesting bit: the last price for this market was 56.3, meaning that the market assigns a 56% probability to the occurrence of a “depression” as defined by Intrade. The average forecast collected by the Wall Street Journal shows a “cumulative decline” of 7.8% (from Q3 2008 to Q2 2009 the forecasts are for contractions at annual rates of 0.5%, 4.3%, 2.5%, and 0.5%), or a peak-to-trough contraction of about 1.9%. Of the 54 individual forecasts collected by the Journal (you can download the data to a spreadsheet), 22, or 41%, are predicting a depression by Intrade’s definition.
So Intrade is more pessimistic than the experts. There has been a lot of talk about the accuracy of prediction markets like Intrade, but a lot depends on the liquidity of the individual market, and this one doesn’t have much (you can see all the outstanding bids and asks). We’ll just have to wait and see who wins this contest.