The (Time-Varying) Importance of Disaster Risk

How much of the historical 7% per year equity risk premium could have been risk compensation for disasters that just happened not to have occurred? The answer can be found in below-the-money put prices, which would have protected against such disasters. Using the cost of rolling over one-month index put options, I show that the maximum possible premium for crash risk could not have accounted for more than about 2% per year, thus leaving about 5% per year for reasons other than sudden disasters. I also provide a novel "conservative diffuse prior" approach for dealing with black swan risk.