December 1, 2020
An event study generates only a lower bound on the full effect of an event unless researchers know the probability that investors assigned to the event before it occurred. We develop two model-free methods for recovering the market’s priced-in probability of events. These methods require running event studies in financial options to complement the standard event study in stock prices. Validating both approaches, we estimate that the 2016 U.S. election outcome had a 12% chance of occurring.