Political Prediction Markets and Manipulation
For many armchair politicos, prediction markets have become ubiquitous resources, more so than even daily tracking polls. Prediction markets work by setting market prices, usually between $0 and $1 (or $0 and $10.00) on binary event contracts. For example, Intrade.com runs multiple current markets for the U.S. Presidential election, but the two most popular contracts by trade volume unsurprisingly are “Barack Obama to be reelected President in 2012,” and “Mitt Romney to be elected President in 2012.” If Obama were to be reelected, the first contract would pay out to all who hold shares at $10.00, and the Mitt Romney contract would pay out $0 to all who own shares. Buyers purchase shares at whatever they feel is the lowest probability that the event will occur. Sellers will sell at a price exceeding his purchase price, and that price reflects the highest probability at which he thinks the event will occur (otherwise he would not sell). In this way, the average current price of a contract on the market is equivalent to the average perceived probability that the event will occur.
These types of markets are fairly clearly a form of a second-price auction. This is so because a buyer need only buy at the price equal to (or marginally greater than) the value offered by the next lowest bidder. In a generalized second-price model, those looking to buy shares of contract, i.e. those who feel an event will occur at a higher probability than it is currently listed at, should always bid at their true value. They will thus never lose value by purchasing contracts and will only stand to gain value if the contract succeeds. Furthermore, sellers should always “ask” to sell at a price equal to the maximum probability that they feel an event can occur, so as to never lose value.
But does it always make sense for one in a prediction market to bid and ask “true” value?
As it turns out, prediction markets are susceptible to many forms of manipulation, as outlined in this entry in the Journal of Business, Entrepreneurship & Law. Some forms of these manipulation, interestingly enough, provide opportunities for astute investors to make “free money” by simply sticking to the proven strategy of betting their fair value; that is to say, even when these markets feature significant forms of manipulation, the best strategy for an investor in such a market is still to invest with his or her true value. For example, in May 2007, a large institutional investor arbitrarily inflated the price on the “Hillary Clinton to win 2008 Presidential Election” contract from approximately 26% (or $2.60 per share) to about 40% (or $4.00). This inflation created an opportunity for investors to make easy money, as those who were looking to sell could now sell for an additional profit of $1.40 per share by selling shares to those regular investors who believed Ms. Clinton actually had a 26% chance to win the Presidency at that time. Thus, they could exploit the difference between their “true” price on Clinton’s chances and the arbitrarily inflated price. Clearly, as discussed in class, it is always best to bid true value.
Source:
http://digitalcommons.pepperdine.edu/cgi/viewcontent.cgi?article=1044&context=jbel&sei-redir=1&referer=http%3A%2F%2Fwww.google.com%2Furl%3Fsa%3Dt%26rct%3Dj%26q%3Dinformation%2520cascades%2520in%2520betting%2520intrade%26source%3Dweb%26cd%3D10%26ved%3D0CHEQFjAJ%26url%3Dhttp%253A%252F%252Fdigitalcommons.pepperdine.edu%252Fcgi%252Fviewcontent.cgi%253Farticle%253D1044%2526context%253Djbel%26ei%3DQb2ZUJ3WFM7RqAGg74CABQ%26usg%3DAFQjCNFQA-tcBIOEB0TZnNL6ikL2HiuBgw#search=%22information%20cascades%20betting%20intrade%22
-jbs296