Categorization and Learning from Financial Information

This paper examines the role of coarse categories in individuals’ learning from financial information. In particular, we (i) test theoretical predictions about categorical over- and underreaction to information by Mullainathan (2002) in an investment context, (ii) explore differences in category-based belief formation and (iii) link category-based beliefs to investment behavior. Our findings document that information aggregation along prominent categories in financial markets, such as industries, can affect people’s beliefs and investment decision-making. Interestingly, we find differences across category types. Subjects form category-based beliefs when the observed stock belongs to “good” stock categories associated with gains. People then overreact to category changes, form overly optimistic beliefs, and invest significantly more in the stock. Yet, we find the opposite pattern if the stock belongs to “bad” stock categories associated with losses. People then seem to be generally sensitive to the stock’s outcomes.
Category changes do not distort their beliefs.