Integrating macroeconomic risk factors in credit portfolio models

ABSTRACT The recent financial crisis has shown the relevance of macroeconomic factors for forecasting and stress testing credit portfolio models. Despite this, most banks still work with a through-the-cycle approach. We show how to integrate macroeconomic variables into the risk management system of a bank using a multifactor credit risk model with observable macroeconomic and latent variables. In an empirical study, we compare the point-in-time results of this model with those of a through-the-cycle model and explain the deficiencies of the latter. We also provide a solution for the important case in which the bank's credit risk model includes no macroeconomic information so that macro-level stress tests and scenario analyses may be executed in a straightforward and consistent way.