Subjective Learning of Trading Talent: Theory and Evidence from Individual Investors in the U.S.

Recent studies show evidence that investors learn about their trading abilities. This paper focuses on understanding how investors learn about their talent and proposes a unifying framework that explains many puzzling facts about individual equity investors. In my model, the investor forms subjective beliefs both about the expected return of the current stock-in-holding and about her trading talent represented by the expected return of the next replacement stock, and updates beliefs through learning with fading memory. I calibrate the memory decay parameters to individual trading records, and show that talent learning is about 7 times more sensitive to return signals than stock-in-holding learning. Consequently, the model indicates that stock switching always happens following good performance of the current stock because switching requires a sufficiently large wedge between expected returns of the replacement stock and the current stock to cover the fixed cost, which strongly predicts disposition effect in a learning perspective. This framework also accounts for the performance-contingent trading intensity and attrition, and explains why a negative shock would lead to attrition when an investor has several years of experience, which is inconsistent with the decreasing-gain updating under standard Bayesian learning.