Belief Updating and Investment Decisions : The Impact of Good or Bad News Varies With Prior Returns

Investors' belief updating differs for investments in a gain position versus those in a loss position and by the favorability of the news, leading to a anomalies in investment decisions. We propose a context-sensitive reinforcement learning model unifying these empirical findings. In a preregistered experiment we show that the model captures investors' belief dynamics better than competing models. We observe stronger changes in beliefs from negative information in gain compared to loss positions, leading to profit-harming decisions. Providing participants with additional information mitigates these effects. Our findings have implications for theories incorporating belief dynamics and practical applications for improving financial and economic decisions