Tous les séminaires
16H15 : Peter BOSWIJK (University of Amsterdam) : "Asset Returns with Self-Exciting Jumps: Option Pricing and Estimation with a Continuum of Moments"
17H30 : Valérie CHAVEZ-DEMOULIN (University of Lausanne - HEC-Lausanne) : "Generalized Additive Models for Conditional Dependence Structures"
16H15 : Peter BOSWIJK (University of Amsterdam)
ASSET RETURNS WITH SELF-EXCITING JUMPS: OPTION PRICING AND ESTIMATION WITH A CONTINUUM OF MOMENTS
We propose an option pricing model with a self-exciting jump component inducing jump clustering. We develop a procedure to imply the latent state variables of the model from a panel of option prices and estimate the model parameters via the generalized method of moments employing a continuum of moments. Monte Carlo simulations show that our estimation procedure has good finite sample properties. Based on a panel of S&P 500 index option prices, we find strong evidence of self-excitation. The model's performance is shown to be superior to that of two alternative models with stochastic volatility and jumps.
This is a joint work with Roger Laeven and Andrei Lalu.
17h30 : Valérie CHAVEZ-DEMOULIN (University of Lausanne - HEC-Lausanne )
GENERALIZED ADDITIVE MODELS FOR CONDITIONAL DEPENDENCE STRUCTURES
We develop a generalized additive modeling framework for taking into account the effect of predictors on the dependence structure between two variables. We consider dependence or concordance measures that are solely functions of the copula, because they contain no marginal information: rank correlation coefficients or tail-dependence coefficients represent natural choices. We propose a maximum penalized log-likelihood estimator and derive its root-n-consistency and asymptotic normality. Finally, we present the results from a simulation study and apply the new methodology to areal dataset.
This is a joint work with Thibault Vatter.
Prochaine séance : Jeudi 26 Mai