Modeling Credit Risk through the Austrian Business Cycle: An Update of the OeNB Model

In quantitative financial stability analysis, the link between the macroeconomic environment and credit risk is of particular importance when assessing the risk hidden in loan portfolios. Macroeconomic stress testing, in particular, which aims at measuring the impact of an economic crisis on individual banks or on the entire financial system, depends on means to quantitatively assess this link. Hence, the objective of this paper is to provide a methodological update of the OeNB’s previous credit risk model that improves the capture of the relation between macroeconomic variables and probabilities of default for the main Austrian corporate sectors. In addition to the standard model based on individual macroeconomic variables, the paper explores solutions to two important challenges: first, the challenge related to the exploitation of potential information inherent in a larger macroeconomic data set and second, the problem that accounts for potential nonlinearity in the relation between credit and business cycles. The first issue is addressed via a regression model based on a principal components analysis that takes in a wider range of macroeconomic variables than commonly practiced. The second issue is addressed via a threshold approach. This paper presents the estimation results for the three different models and discusses them on the basis of an illustrative example.