Analysis of Italian Financial Market via Bayesian Dynamic Covariance Models

Abstract

The attempt to provide a quantitative view on the evolution of the temporal and geo-economic relations between the Italian Stock Market Index FTSE MIB and the major financial markets before and during the global financial crisis of 2007–2012 motivates the search for statistical methodologies able to accommodate flexible dynamic structure of dependency among assets and to answer the main issues of multivariate financial time series analysis. This work compares, through an application study, some recent advances in Bayesian covariance regression, with a particular interest in the local adaptive smoothing of the stochastic processes under investigation in order to allow the covariances among returns to vary flexibly over continuous time.