Experimental Economics

Research project P 13919 Experimental Economics Ulrike LEOPOLD-WILDBURGER 11.10.1999 In the study of economic behavior we are frequently confronted with the question of rationality. Difficulties arise when comparing theoretical models of rational behavior with observable reality. Modem experimental economics is a useful tool for the analysis of behavioral patterns of both individuals and groups. It is based on laboratory experimentation, in which the economic environment is completely under the control of the experimenter. Experimental economics is closely related to economic psychology, game theory and the theory of rational behavior. In a certain sense economists and psychologists have common intellectual roots: The field of economic psychology as a connecting link. between economics and psychology goes back to a paper by the French author Gabriel Tarde (1902). In its present form experimental economics was founded by H.Sauermann & R.Selten (1959) and by Siegel & Fouraker (1960). Controlled experiments allow researchers to analyze and forecast economic processes systematically. Recent improvements in the technique of forecasting on the basis of a priori hypotheses involve mathematical statistics and as an optimization approach Operations Research. In the present experimental study we will restrict ourselves to forecasts of time series, especially in relation to financial markets. We want to evaluate graphical extrapolation of a time series as a simple and popular forecasting technique. Recognizing regularities of a given curve by visual inference enables the filtering out of outliers and helps to make out the patterns of the stochastically superimposed formation rules according to which a certain series of numbers is formed. We are interested in an a priori model for explaining individual and average forecasts. On the basis of a pilot study we have created a validated model for which simulation studies and extensions of various dimensions are intended.