Skewness seeking in a dynamic portfolio choice experiment

Abstract

We conduct a controlled laboratory experiment in which subjects dynamically choose to allocate their portfolio between (i) a safe asset, (ii) a risky asset, and (iii) a skewed asset with a negative expected value (a “bet”). Subjects repeat this process over multiple independent investment cycles and can sometimes acquire information about the performance of their peers. In this informationally rich environment, skewness seeking emerges as predominantly context- and state-dependent, rather than driven by risk preference traits. Although more than half of subjects persist in buying the bet, there is substantial learning to avoid it. Purchases are most frequent when subjects have more funds available and at the very end of the investment cycle. Subjects also show interest in their relative standing and buy more bets when their investment performs well but they unexpectedly learn that their peers are outperforming them.