Title: Intersectorial default contagion: A multivariate Poisson auto-regression analysis
Authors: Ana Escribano - Universidad de Castilla-La Mancha (Spain) [presenting]
Mario Maggi - University of Pavia (Italy)
Abstract: Credit rating default dependences are analyzed in a multi-sectorial framework. Using the Mergent's FISD database, we study the default series in the US over the last two decades, disaggregating defaults by industry-sector groups. During this period, two main waves of default occurred: the implosion of the ``dot com'' bubble, and the global financial crisis. We estimate a Multivariate Autoregressive Conditional Poisson (MACP) model to the weekly number of defaults occurred in the different sectors of the economy. We discuss the contagion effect between sectors in two ways: the degree of transmission of the probability of default from one sector to another, i.e., the `infectivity' of the sector, and also the degree of contagion of one sector from another, i.e., the `vulnerability' of the sector. Our results show some differences between the sectors' relations during the first and the second part of our sample. We add to the analysis some exogenous variables and evaluate their contribution to the goodness of fit.