Testing Alternative Theories of Financial Decision Making: A Survey Study with Lottery Bonds

Abstract

In this paper, we present the results of a simple, easily replicable, survey study based on lottery bonds. It is aimed at testing whether agents make investment decisions according to expected utility, cumulative prospect theory (Tversky-Kahneman, 1992) or optimal expectations theory (Brunnermeier and Parker, 2005, Brunnermeier et al., 2007) when they face skewed distributions of returns. We show that more than 56% of the 245 participants obey optimal expectations theory. They choose a distribution of payoffs which is dominated for second-order stochastic dominance and which would not be chosen according to cumulative prospect theory, for a large range of parameter values. Our results first cast doubt on the relevance of variance as a measure of risk; they show the importance of skewness in decision making and, more precisely, they emphasize the attractiveness of the best outcome, an essential feature of optimal expectations theory. The ranking of outcomes, used in cumulative prospect theory, seems insufficient to characterize the way people distort beliefs. As by-products of this study, we illustrate that agents use heuristics when they choose numbers at random and have, in general, a poor opinion about the rationality of others.