Information Flow Dependence in Financial Markets

In response to empirical evidence we propose a continuous-time model for multivariate asset returns providing a two-layered dependence structure. The price process is subject to multivariate information arrivals driving the... [ view full abstract ]

In response to empirical evidence we propose a continuous-time model for multivariate asset returns providing a two-layered dependence structure. The price process is subject to multivariate information arrivals driving the market activity modeled by non-decreasing pure-jump Lévy processes. The jump dependence is determined by a Lévy copula allowing for flexible dependence beyond the conditional aspects of the return distribution. Assuming that conditional asset returns are jointly normal, their dependence is modeled by a Brownian motion allowing for correlation. We apply novel multivariate models to equity data and obtain estimates having an economic intuition with respect to the two-layered dependence structure.