Information and efficiency in political stock markets: using computerized markets to predict election results

This study concentrates on four computerized political markets in Germany between 1990 and 1998. While this new method for the prediction of election events worked quite well in the USA it did not perform as well in Germany. Searching for the causes of this distinction it is shown that, in contrast to the findings of Forsythe et al. (1992 Forsythe, R, Nelson, F, Neumann, GR and Wright, J. 1992. Anatomy of an experimental political stock market. American Economic Review, 82: 1142–61.  ), (i) extraordinary profits were achieved less by people who took advantage of other people's anomalies than by those people who acquired an advantage from the existence of asymmetric information, (ii) the marginal trader hypothesis does not hold when applied to German markets, and that (iii) traders relied on public opinion polls. It is argued that these distinctions are caused by the differences in the German and the US voting systems. Additionally, it is shown that to a certain extent (iv) election markets were able to predict the contemporary mood of the electorate without the help of public opinion polls, (v) first of all the informed traders used public opinion polls as a source of information, and (vi) prices themselves became a source of information on which expectations were based.