Tail Risk, Volatility, and Return Predictability

• Activity Number: 49
• Type: Invited
• Date/Time: Sunday, August 9, 2015 : 4:00 PM to 5:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Viktor Todorov*
• Companies: Northwestern University

Abstract

We develop new methods for the estimation of time-varying risk-neutral jump tails in asset returns. In contrast to existing procedures based on tightly parameterized models, our approach imposes much fewer structural assumptions, relying on extreme-value theory approximations together with short-maturity options. The new estimation approach explicitly allows the parameters characterizing the shape of the right and the left tails to differ, and importantly for the tail shape parameters to change over time. On implementing the procedures with a panel of S\&P 500 options, our estimates clearly suggest the existence of highly statistically significant temporal variation in both of the tails. We further show that the time variation in the shape of the tails and the compensation demanded by investors for bearing jump tail risk help predict future market returns in addition to standard predictors used thus far.