Modeling the Dynamics of Conditional Dependency between Financial Series
Eric Jondeau and Michael Rockinger*
Forthcoming in a book edited by Emmanuel Jurczenko and Bertrand Maillet, Springer Verlag.
Abstract
We develop a new methodology to measure conditional dependency between daily stock-market returns, which are known to be driven by complicated marginal distributions. For this purpose, we use copula functions which are a convenient tool for joining marginal distributions. The marginal model is a GARCH-type model with time-varying skewness and kurtosis. Then, we model the dynamics of the dependency parameter of the copula as a function of predetermined variables. We provide evidence that our model fits the data quite well. We obtain that the dependency parameter is both large and persistent between European markets. Our methodology has many potential applications, such as VaR measurement and portfolio allocation in non-gaussian environments.
Keywords: Asset allocation, Stock returns, Non-normality, Utility function.
JEL classification: C22, C51, G12.
* HEC Lausanne