Fat-finger event and risk-taking behavior

Studies of the effect of market shocks on individuals’ risk-taking behavior often cannot determine whether the effect arose due to a change in an individual’s preferences or beliefs. This paper investigates the effect of a specific fat-finger shock – an external algorithmic trading error – that generated no information about market fundamentals and thus should not have induced a change in beliefs. Using both OLS and propensity-score matching DID methods, we find that investors who traded during the shock significantly reduced risk taking, while those who did not trade did not change their behavior. The significant effect was short-lived, mainly driven by the experience of small realized losses, and found only among small individual investors but not among large individual investors. The evidence suggests that the change in risk taking is likely attributable to a change in preferences.