The Science Behind PPX

Consider the "Saddam security," a proposition on intrade.com that paid out only if Saddam Hussein was out of power by July 30, 2003. It functioned, in the days before the war started, as a measure of the likelihood that the U.S. would invade (traders did not put much stock in the possibility that Saddam would retain power if American forces entered Iraq). Wolfers and Eric Zitzewitz, a professor at Stanford University's graduate school of business, tracked the movement of the Saddam security and compared it with changes in a variety of global markets to analyze how traders were responding to the possibility of war. Their methodology was straightforward: "If a 10 percent increase in the probability of war caused the stock market to fall by one and a half percentage points," Wolfers says, then he could extrapolate to 100 percent and conclude that "the difference between war and peace would cause the market to decline by 15 percent." That's bigger than the drop in the days following September 11, 2001. Wolfers and Zitzewitz found that short-term oil prices also shot up when war appeared imminent but long-term prices fell, presumably because of the chance of cheap Iraqi oil eventually hitting the market.

All these inferences were clear in the weeks and months leading up to the final decision to invade, but this was also a special case: There was arguably no more important factor in the price of oil or the health of the U.S. economy in early 2003 than a U.S. invasion of Iraq. The price also moved around enough to let us tease implications from its movement. But what if we wanted to uncover the effects of an event that had a fairly stable probability-how the capture of Osama bin Laden would affect the airline industry, for instance.Or what if we wanted to determine how the market would react to a relatively minor input, like a new trade accord?

The best way to understand these types of "if/then" questions is to create a conditional market, also known as a decision market. These simply offer two complementary propositions: "How much will x be if y occurs?" and "How much will x be if y does not occur?" Substitute anything you want for x and y: "The price of the S&P 500"/ "Troops withdraw from Iraq." "Gas prices"/ "Drilling begins in the Arctic National Wildlife Refuge." Then take a look at how the two complementary propositions are trading. If the market expects the S&P 500 to reach 1,650 if troops withdraw but only 1,500 if troops do not withdraw, then you have a clear view of the financial consequences the stock market attaches to our military policy: a difference of 150 points. "I'd love to see more of these complicated questions being asked, like what will happen in the housing market if Hillary Clinton wins the election," says David Pennock, a senior research scientist at Yahoo and head of its prediction-market research. "We could put forth the clear decisions we have to make as a country and examine their implications."

The trouble with these markets is in getting enough people to participate. It's legally precarious to set up a decision market in the U.S. (trading these propositions looks suspiciously like Internet gambling to some-although by those standards, so does investing in pork-belly futures). Most American ventures use play money, which surprisingly enough, generates predictions as accurate as those of real-money markets. These fake-money markets need to attract traders by being entertaining, not by offering the possibility of riches. It's no wonder the U.S. market with the most trades is the Hollywood Stock Exchange.

Without the lure of real money to be made off other, perhaps less-savvy market participants, it's hard to recruit enough traders into a decision market that creates insight into serious yet perhaps more mundane issues. But it can be done. We have the crystal ball at our fingertips. All that's left is to look into it together.

Michael Moyer is the executive editor of

Armed with your new knowledge, start an account on PPX today.