9 . Forecasting poli , t ' lcal developments with the help of financial markets

It is a truism to say that forecasting political and economic trends continues tobe a booming business (Sherden 1997). Within academia, however, many still conceive of prediction as a less noble task than explanation. The wide range of attitudes, between attraction and skepticism, leveled at professional forecasters can even be found among those who have transformed forecasting from an obscurantist business associated with astrclogy, crystal gazing, and "propheteering" into a key part of "normal science." The example of Oskar Morgenstern, one of the founders of modern game theory, illustrates this ambivalence nicely. In one of the earliest treatises on economic forecasting he stated apodictically: "Economic prognosis is ... impossible because of objective reasons" (Morgenstern 1928: 108, own translation). Yet, this co-founder of game theory and leading strategic analyst continued to publish on the topic and evaluated for instance the random walk thesis together with Nobellaureate Clive Granger (Granger and Morgenstern 1970). Nevertheless, the acceptance and popularity of prediction as an academic endeavor vary greatly across social scientific disciplines. While the combination of explanation and prediction is the natural way to do research in demography or econometrics, political scientists are generally still hesitant to engage in this seemingly dirty business. One of the traditional reasons for this unwillingness has been the inaccuracy of standard techniques like opinion polling to forecast specific developments such as election outcomes. Forecasting political events, however, no Ionger warrants this approach. The development of two techniques betting markets and expert interviewbased decision models has considerably changed the predictive accuracy of attempts to forecast the outcomes of political qecision-making. In this chapter, I present an approach that marries some of the advantages that these two approaches offer. I test whether the collective information that