At the moment of this writing, traders on Intrade—probably the most prominent and active information market in the world—have Barack Obama at about 53.5% to win the 2008 election. Meanwhile, over at the Iowa Electronic Markets—maybe the second-most prominent information market in the United States—traders have Senator Obama at approximately 63% to win the election. This type of broad schism is not supposed to occur in information markets (or any other type of market, for that matter). It raises a number of interesting questions about why the divergence has developed, why it has been allowed to persist, and which of the two figures is to be believed.
The second of these is easier to answer. In a typical market, arbitrageurs would exploit this type of spread to the tune of nearly ten cents per transaction, causing Obama’s Intrade price to rise and his Iowa price to fall until the spread disappeared. Here, however, the Iowa markets are too thin to permit any serious arbitraging; there simply aren’t enough buyers and sellers at any given moment. Intrade is somewhat deeper, but of course an arbitrageur needs adequate volume on both markets to be successful. Intrade also charges transaction costs, which will eat into the available margin. It’s not hard to imagine that arbitrageurs are not excited about expending the time and effort to open accounts on both markets and place multiple trades in order to trade just a few hundred shares at a profit of only ten cents per share.
The more interesting—and perhaps more difficult—puzzle is why this discrepancy exists in the first place. The answer may lie in the different populations of people who trade on each of the two markets. Intrade is available to anyone with a web browser and a bank account, but it is a for-profit British company, and it may not be entirely legal for Americans to trade on it (though many undoubtedly do). The Iowa Electronic Markets, on the other hand, are run by the University of Iowa’s business school as a non-profit research tool. This fact has two principal ramifications: first, that Americans can trade on the Iowa Electronic Markets without fear of legal repercussion; and second, that the conventional wisdom holds that a disproportionate share of traders on the Iowa markets are university faculty and students. (This is a widely shared assumption, but if data exist I have not seen them.) Faculty and students are different from the general population of internet users in a variety of respects, but I’ll pick just two salient ones: they are typically better educated, and they are more likely be Democrats. It seems reasonable to assume, then, that a greater percentage of traders on the Iowa markets are American citizens, and that these traders are better educated and more likely to support Barack Obama than those on Intrade. These factors may account for the Iowa-Intrade divergence; Iowa traders may have a stronger understanding of the political climate, or they may have something of an optimism bias in favor of the Democratic candidate.
The more general point is that thin markets, high transaction costs, and difficulties in undertaking certain types of transactions (for instance, it is rather cumbersome and expensive to short a stock on the Iowa markets) can lead to mispricing. On Intrade and the Iowa markets this doesn’t particularly matter; very little is at stake. But mispricing of markets that are actually used to allocate valuable goods can engender real problems.
Update: a blogger at FiveThirtyEight.com notes that Betfair.com, another British information market with volume similar to Intrade's, also has Barack Obama stock trading approximately ten points higher than Intrade does. The suspicion is that someone has been manipulating the Intrade market. This raises an entirely new set of questions, but for the moment it is perhaps most important to note that if in fact manipulation is occuring, it will become extremely expensive to maintain before too long. It will be interesting to watch how these markets move in the next few days.