As everyone who has ever played the ponies knows, sometimes a 50-1 long shot wins. After all, when the betting markets predict, as they were as of this writing, that Trump has a roughly one-in-three chance of being elected president, that’s not a trivial possibility. Most pollsters, meanwhile, accurately predicted both events.īut that’s no reason to ignore these markets. Most striking were the failure of online-exchange gamblers and London bookies to get the Brexit vote right last summer and the failure of the betting markets to anticipate Trump’s nomination to head the Republican ticket. This year, in particular, they’ve hit potholes. The betting markets are far from infallible. A 2008 study by three University of Iowa researchers found that the Iowa market was more accurate than 74% of the 964 presidential polls conducted during the five elections between 19.

That market has generally been more accurate than polls in a wide swath of executive, legislative, national and local elections. The Iowa market’s results have been striking. Formed by business professors at the University of Iowa in 1988 for students and teachers to wager on that year’s presidential election, the Commodities Futures Trading Commission licensed IEM to take bets from the public on subsequent elections because of its academic purpose-but the commission put a $500 limit on an individual’s “investments.” In the U.S., the Iowa Electronic Markets is the granddaddy of these exchanges.

Presidential election bet online