CMStatistics 2019: Start Registration
View Submission - CFE
A0158
Title: The leverage effect puzzle revisited: Identification in discrete time Authors:  Eric Renault - University of Warwick (United Kingdom) [presenting]
Abstract: In spite of evidence that leverage effect (negative correlation between volatility and return) should be present, the estimation of this effect is notoriously difficult. We propose a new discrete time stochastic volatility model with leverage effect that is a natural discrete time analog of popular continuous time affine option pricing model. With an exponentially affine stochastic discount factor, the historical and the risk neutral models belong to the same family of joint probability distributions for return and volatility processes. The discrete time approach allows making more transparent the role of various parameters: leverage versus volatility feedback effect, connection with daily realized volatility, impact of leverage on the volatility smile, etc. Even more importantly it sheds some new light on the identification of leverage effect and of the various risk premium parameters through link functions in closed form. The price of volatility risk is identified from underlying asset return data, even without option price data, if and only if leverage effect is present. However, if leverage effect is close to zero, identification of the volatility risk price may be weak, leading to a new procedure of identification robust inference based on link functions.