Past studies typically have focused on whether people perceive more rare risk after experiencing catastrophic disasters. We show that people can also feel less risk with unexpected lucky disaster experience. By exploring a novel identification strategy based on households’ expectations, we find that households perceive less (more) risk when they experience disasters that have lower (higher) fatalities than what was expected. This opposite experience effect of rare disasters is substantial. A one standard deviation increase in the negative (positive) experience shock is associated with a 1.71% decrease (a 1.31% increase) in the life insurance-to-portfolio ratio. We discuss three possible mechanisms to account for our empirical findings: incomplete information learning, salience theory, and change in risk preferences.

We study here say-buy/whisper-sell behavior wherein analysts issue optimistic recommendations to attract retail investors while providing more accurate information to fund managers in private, sometimes resulting in fund managers selling the recommended stocks. We test whether fund managers return the favor using their votes for analysts in a Chinese “star analyst” competition. Managers are more likely to vote for analysts who exhibit greater “say-buy/whisper-sell” behavior toward these managers. This suggests that analysts reduce the accuracy of their public recommendations, thereby maintaining the value of their private advice to funds. Our findings help explain several empirical puzzles about analyst public recommendations.