52 Pages Posted: 15 Nov 2017 Last revised: 26 Mar 2019
Date Written: November 30, 2018
We estimate and analyze the ex ante higher order moments of stock market returns. We document that even and odd higher-order moments are strongly negatively correlated, creating periods where the return distribution is riskier because it is more left-skewed and fat tailed. Such higher-moment risk is negatively correlated with variance and past returns, meaning that it peaks during calm periods. The variation in higher-moment risk is large and causes the probability of a two-sigma loss on the market portfolio to vary from 3.3% to 11% percent over the sample, peaking in calm periods such as just before the onset of the financial crisis. In addition, we argue that an increase in higher- moment risk works as an "uncertainty shock" that deters firms from investing. Consistent with this argument, more higher-moment risk predicts lower future industrial production.
Keywords: asset pricing, financial economics, higher order moments, tail risk
JEL Classification: G00, G1, G12, G13, G17
Suggested Citation: Suggested Citation