Characterizing dependence in financial series
Faculty Advisor: Victor H. de la Pena
In this thesis I address the question, how do financial series move together? In order to do this, I develop a new method of modelling different dependence structures, utilizing a mixed copula approach. This method may be applied in unconditional and conditional settings, and allows natural nesting of symmetric and asymmetric dependence. Moreover, the mixed copula framework is directly linked to issues of downside risk, and characterization of financial market turbulence. The first chapter develops my insights on issues of dependence that are common to various financial settings, and derives a number of useful technical and conceptual contributions. The second chapter builds on these contributions to estimate the unconditional structure of dependence in international financial markets. The third chapter introduces and develops a dynamic extension of the unconditional copula to address the importance of turbulence and quiescence in financial markets.