Abstracts

We study the joint properties of option pricing together with the market prices of volatility risks, with the help of a parsimonious three-factor model featuring interdependent volatilities and an unspanned skewness component. Using closed-form expressions for option prices and the market price of variance risk, we perform a consistent estimation for a large panel of S&P 500 options and infer on the dynamics and term structure of variance risk premia. We find that our model produces an excellent pricing performance, improving on that of more parameterized three-factor affine models. It also yields variance risk premium dynamics consistent with economic intuition and with the features of unconstrained volatility predictive regressions. Our findings also highlight (i) the inaccuracy of lower-dimensional (e.g., two-factor) models for capturing the S&P 500 volatility dynamics; (ii) a systematic link between volatility risk premia and model-implied unspanned skewness, (iii) a rich dynamics of the term structure of volatility risk premia, which is systematically related to market-wide conditions and sentiment, (iv) the usefulness of model specifications introducing an unspanned skewness dimension, and (v) the necessity of specifying jumps in the underlying price process, at least.