INSTITUTE OF MATHEMATICAL ECONOMICS

Wulf Albers

Working-Paper No. 284

January 1998

Abstract

The purpose of this paper is, to show that the theory of prominence is an adequate tool to make predictions for subjects' money equivalents of lotteries of type [x(p),y(q)]. The evaluation principle is similar to that of KAHNEMAN and TVERSKY (1989 and 1992), however the evaluation functions for money and probabilities are not obtained by econometric curve fitting, but by behavioral principles described by the theory of prominence. Moreover, money and probability space are clearly distinguished from the perception space, specificly the result of the evaluation in the perception space is remapped to the money space to obtain the money equivalent.
The surprising result is that a parameterfree boundedly rational model using the insights of the theory of prominence concerning the evaluation of numerical stimuli (in the decimal system) adequately describes median behavior of subjects, even better than the model of KAHNEMAN and TVERSKY. The model is very simple, although it has no parameters that have to be adjusted. It seems that this instrument can be used as a predictor of median subjects' behavior, i.e. the behavrior of subjects, who have - compared with other subjects - no special motives to take or to avoid risk. From this point of view the model defines a benchmark of 'risk neutral' behavior, where 'preferences' are only induced by the principles of perception as defined by the theory of prominence.
I thank Bodo Vogt for many helpful discussions and him, Ralf Sievert, Andreas Uphaus, and Axel Woeller for their assistance to run the experiments.