Trading under ambiguity and the effect of learning ∗

Abstract In this experimental study, I investigate the effect of ambiguity on investors’ willingness to trade under different information conditions. The results confirm the prediction of a wide set of theoretical models, that ambiguity aversion reduces willingness to trade in incomplete markets. Participants choose significantly wider bid-ask spreads when return distributions are ambiguous rather than objectively known. This effect also persists when subjects learn probabilities progressively. However, in the latter case, different information generates more divergent quotes. The more extreme quotes are consistent with a particular updating rule, which is conditional smooth preferences. These findings highlight the role of gradual information release for market underand overreaction in ambiguous markets.