Prediction markets are betting exchanges in which participants buy financial contracts based on the outcome of future events, like political elections. While many offshore exchanges offer these markets, the majority of prediction market publicity in the mainstream media is due to the Iowa Electronic Markets (IEM, http://www.biz.uiowa.edu/iem/). The IEM is the only legal online prediction market in the U.S., specializing in political contracts like the winner of the U.S. Presidential election, although they also offer markets on Federal Reserve Monetary Policy, performance of U.S. equities, and even movie box office performance.
Prediction markets have been acclaimed as accurate predictors of real-world events, and evidence has shown that they are better than polls in predicting election results. Markets have been proposed as a way to best forecast events ranging from business decisions to terrorist attacks (a previous MajorWager column advocated for the legalization of prediction markets: http://www.majorwager.com/frontline-672.html).
But while prediction markets are thought to be the best way to collect the "wisdom of crowds", there are still two controversial questions about the real-world application of these markets. First, are the markets efficient? In other words, do market prices really reflect the chances of an event happening? Reams of research have been published on this topic, and while markets are not completely efficient, they seem to give better predictions than any available alternative.
The second issue is whether prediction markets can be manipulated. Market manipulation is a serious issue for any trader, since it can seriously impact trading strategies and, therefore, profits. A sudden price change in a given market may mean that an "insider" has received breaking news that he is trading on. Unfortunately, it may also just be the result of someone attempting to influence the price for their own reasons.
Unfortunately, manipulation may not always be easy to see. Another complication is the fact that market manipulators may not be motivated just by profit. Some traders might want to manipulate markets to influence public opinion. For instance, manipulating election markets to make one candidate appear weaker than he really is may end up costing the candidate votes. These issues of "strategic" market manipulations may become more common as prediction markets become larger and more influential, and are therefore given more clout from both the press and the public at-large.
A pair of economists have recently turned their attention to the question of market manipulation. Paul Rhode, a professor at the University of Arizona, and Koleman Strumpf, professor at the University of Kansas School of Business, have written a preliminary paper* about their research on manipulation of prediction markets. While this is still a work in progress, the article provides an interesting look at market dynamics in betting exchanges.
The economists looked at political prediction markets using three approaches:
1. An experiment in which actual trades were made on the Iowa Electronic Markets to simulate market manipulation
2. A review of political markets operating in New York from 1880 to 1944 to look for historical evidence of price manipulation
3. An investigation into a series of "speculative attacks" on the 2004 Presidential Election markets at offshore exchange Tradesports, manipulation that was notable enough to receive press in the Wall Street Journal and Time magazine
Interested readers will have to peruse the preliminary draft on their own for the complete details. The bottom line is that in all cases, "speculative attacks" initially moved prices, sometimes quite dramatically, but this move was quickly reversed and prices returned to their previous levels after a short time. The upshot is that, while manipulating political markets is possible, it is, as the authors put it, "difficult and expensive to do for more than a short period".
Besides setting to ease the minds of traders who may worry about the influence of manipulation in the markets, these findings should also help in the push to legalize and expand the reach of betting exchanges. A major criticism towards the use of prediction markets for real-life decision making is that they could be easily manipulated by outside groups. This research suggests that prediction markets can't be easily toyed with, and that market traders in general are sophisticated enough to recognize market manipulation for what it is.
For the more casual reader not interested in details of the economic theory, the anecdotal evidence in the article is still interesting. For the investigation of manipulation at Tradesports, the exchange provided extensive trade data to the researchers, allowing for a rare glimpse inside the market dynamics of one of the most active offshore exchanges. The authors do a good job of tying this market data into real-world news events during the 2004 election. For history buffs, the background on the New York political markets is worth a read.
*Paul W. Rhode and Koleman S. Strumpf. Manipulating Political Stock Markets: A Field Experiment and a Centur of Observational Data. A preliminary draft can be found at http://www.unc.edu/~cigar/papers/ManipIHT_June2008(KS).pdf.
If you would like to make or read comments about this article, you may do so by visiting the Mess Hall forum at MajorWager where a thread has been started. Please click HERE