GOP Keeps Senate, Loses House, Betting Site Says. – [US political prediction markets] – by Ronald Kessler – 2006-10-24

One theory is that prediction markets are influenced by the results of opinion polls. But if that were true, individual polls would also influence each other. Moreover, long before the Internet and opinion polls came into existence, election betting was accurately predicting election outcomes. From 1884 to 1940, betting was conducted on Wall Street by specialized brokers called betting commissioners. The betting odds for each candidate were published daily in the New York Times and other papers. The so-called New York betting markets correctly predicted 12 of the 13 presidential elections between 1884 and 1940, according to Koleman S. Strumpf, Koch professor of economics, University of Kansas School of Business, who co-authored a paper examining the markets. In the one exception, the betting swung to even odds by the time the polls closed. The Gallup Poll, the first scientific opinion poll, began in 1935. The arrival of opinion polls and stricter anti-gambling laws drove out the New York betting markets. The Internet has led to their revival.

Paper: Historical Prediction Markets: Wagering on Presidential Elections – (PDF) – by Paul W. Rhode and Koleman S. Strumpf – 2003-11-10

My Question: Before 1935 (that’s when George Gallup crafted the first scientific polls), what the hell those political prediction markets were aggregating, for Christ’s sake??? And where is our good doctor Koleman Strumpf when we need him?