(FORTUNE Magazine) – Question: What is the best way to predict the winner of a sporting event? Answer, possibly getting familiar in these precincts: Look at the betting line. Academic research firmly states that free markets, in which greedy people put their own money on the line, are the best available predictors of probabilistic events. Next question: What is the best way to predict the winner of a presidential election? Improbable answer: the Iowa Presidential Stock Market (IPSM). This fascinating institution, founded by a determined gang of of rational expectationists in the University of Iowa economics department, opened in June and recently was the only place in the country where you could legally bet on the presidential election. (Nevada law forbids election bets.) Actually, you might have trouble getting your money down even if you read this before November 8: The university has limited the action to students, faculty, and staff. Some of these investors have made relatively small, unserious bets; however, one freshman put up $1,000, and most others have got in deep enough to suggest that their minds are being concentrated. Average daily volume in late October: around $1,300. Total number of market participants: around 230. The ''betting'' in the IPSM is not exactly on who will win the election. Instead, investors have been trying to estimate how Bush, Dukakis, and the minor-party candidates would divide the popular vote. The object: to identify a candidate who would rise in popularity over time. If you invested in Bush at a time when the market said he would get 51% of the popular vote, and if investors later decided he was going to get 56%, you could sell at a profit; alternatively, you could hold your shares until after the election, when stock-exchange management -- the rational expectationists -- pays off everybody who still has a profit. As you might expect, Bush prices were strong, and Dukakis prices weak, during most of October. The market action was especially interesting in the early part of the month, when most national polls were showing the race close and Bentsen was widely judged to have won the October 5 debate with Quayle. Rejecting this perspective, the market sold Dukakis quite steadily between October 3 and 8, signaling a weakness that did not show up in the polls until later. Late in October, the market was showing a Bush lead over Dukakis of 7.6%. If you believed in free markets and lesser evils, you would doubtless have rated that good news.