糖心原创

Probability Weighting as Evolutionary Second-best

Abstract: According to the concept of the second-best, multiple deviations from a first-best optimum are optimal once the first-best itself can no longer be achieved because of constraints. In an evolutionary framework, we apply this idea to behavior under risk. We argue that non-linear probability weighting is a second-best complement to payoff valuation as postulated in prospect theory. Previous work has shown that an adaptive S-shaped value function may be evolutionarily optimal if decision-making is subject to cognitive or perceptive constraints. We show that overweighting of small and underweighting of large probabilities can be seen as a partially compensating distortion to such a value function. We also discuss the impact of loss aversion on optimal probability weighting and the empirical testability of our model.