Measuring Contagion Within a Financial Network: A New Conditional Distance to Default Approach

In the aftermath of the global financial crisis, regulators and financial authorities have recognized the importance of developing methodologies for assessing interconnections among companies. Understanding the propagation of shocks within interconnected networks has emerged as a critical aspect of macro-prudential monitoring. However, existing methodologies have faced significant challenges. Firstly, many suffer from the curse of dimensionality, limiting their ability to effectively capture interconnectedness within large networks. Additionally, while various methodologies offer interconnectedness measures (or rankings), they often fall short in elucidating the practical implications of such interconnectedness in the event of a shock. In this paper, we propose a novel framework of systemic risk measures aimed at evaluating how shocks to individual companies can amplify the risk of default among other entities within the same network. Leveraging the established concept of distance to default, our objective is to equip macro-prudential regulators with practical tools for monitoring financial system stability during turbulent periods. Furthermore, by employing a large time-varying parameter vector autoregression approach, we address both the need for dynamic metrics and the curse of dimensionality. Unlike previous studies, our methodology jointly estimates all bivariate interdependencies simultaneously, offering a more comprehensive assessment of connectedness. While multivariate distance to default approaches have been explored previously, our study represents, to the best of our knowledge, the first application of a large time-varying parameters vector autoregression model to dynamically assess the impact of each entity on the distance to default of others. Finally, we demonstrate the practical implementation of these measures by applying them to the components of the Dow Jones Industrial Average.