Market- and Book-Based Models of Probability of Default for Developing Macroprudential Policy Tools

The recent financial crisis raised awareness of the need for a framework for conducting macroprudential policy. Identifying as early as possible and addressing the buildup of endogenous imbalances, exogenous shocks, and contagion from financial markets, market infrastructures, and financial institutions are key elements of a sound macroprudential framework. This paper contributes to this literature by estimating several models of default probability, two of which relax two key assumptions of the Merton model: the assumption of constant asset volatility and the assumption of a single debt maturity. The study uses market and banks? balance sheet data. It finds that systemic risk in Luxembourg banks, while mildly correlated with that of European banking groups, did not increase as dramatically as it did for the European banking groups during the heights of the financial crisis. In addition, it finds that systemic risk has declined during the second half of 2010, both for the banking groups as well as for the Luxembourg banks. Finally, this study illustrates how models of default probability can be used for event-study purposes, for simulation exercises, and for ranking default probabilities during a period of distress according to banks? business lines. As such, this study is a stepping stone toward developing an operational framework to produce quantitative judgments on systemic risk and financial stability in Luxembourg.