Joint paper with with Bordalo, Burro, Coffman and Shleifer
How do we form beliefs about new risks, about which we have little or no direct experience? We address this question using a representative survey of beliefs about covid's lethality. The evidence is consistent with a memory based mechanism in which similar enough experiences are used to simulate covid death, acting as sources of pessimism while less similar experiences interfere, acting as sources of optimism. The model reconciles average overestimation of covid risks with large belief dispersion, with some people systematically underestimating covid risks.