washington, dc

The Democratic Strategist

Political Strategy for a Permanent Democratic Majority

Are Gamblers More Reliable Than Polls?

Far be it from us to (gasp) encourage gambling. But Slate has just launched a fun feature of interest to political junkies, which provides:

…a comprehensive guide to all the big political prediction markets. From now until Election Day 2008, we’ll publish regular updates of the key data from Iowa Electronic Markets, Intrade.com, Newsfutures.com, and Casualobserver.net. (Casualobserver has not yet launched its 2008 political prediction market, but we will add it as soon as it goes up.) In these early days of the campaign, we are tracking four markets: 1) Democratic nominee for president, 2) Republican nominee for president, 3) presidential victor, and 4) party control of the presidency. We’ll add Senate and House races as they heat up next year.

This is not just a greed game for bread-heads. As Slate notes:

The thrill of prediction markets for political junkies is that they harness “the wisdom of crowds.” A single person’s bet on an election outcome isn’t very good, but thousands of bets, with real stakes, are more likely to predict the correct result than even the best pundit. The Iowa Electronic Markets, the big daddy of the political prediction markets, is consistently better at forecasting winners than pre-election polls.

University of Iowa Biz School scholars offer some verification for the claim here. There may be a few pollsters out there who beg to differ. Might be fun for pollsters and gamblers to make a little group wager.

One comment on “Are Gamblers More Reliable Than Polls?

  1. Sam on

    Any analysis that concludes political markets are a better prediction of election results than are polls does not take into account discounting over time (which economic principles demand be applied in this situation–buyers and sellers in the markets inherently include discounting of political information in their calculations, while polls represent a snapshot in time). With the proper discounting coefficient, models using time-series polling data outperform political markets as a predictor of vote results. Check out this paper by Robert Erikson, one of the foremost political science authorities on American elections:
    http://www.temple.edu/ipa/workingPapers/Documents/Erikson%20and%20WlezienWAPOR%202005.pdf

    Reply

Leave a Reply

Your email address will not be published.

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.