MODELING METHODOLOGY FEBRUARY 2016FEBRUARY 2016CAPITAL MARKETS RESEARCHValidating the Stressed EDFTM Model for Public FirmsHighlights¯ Validation of the Stressed Expected Default Frequency (EDF) model shows that the methodological approach and model parameters continue to adequately capture the relationship between macroeconomic factors and the credit risk of public firms.¯ In order to test the out-of-sample validity of the Stressed EDF model methodology, we calculate perfect foresight Stressed EDF measures for the period July 2014 through September 2015, conditioned on realized macroeconomic data during this time.¯ Perfect foresight Stressed EDF measures conditioned on realized macroeconomic data track realized (unconditional) EDF measures well during the out-of-sample period at the aggregate and firm levels and within sector.¯ This validation exercise covers all four Stressed EDF models: North America, Western Europe, Japan, and Australia/New Zealand.Author Glenn LevineContact Us Americas +1.212.553.1658 clientservices@moodys.comEurope +44.20.7772.5454 clientservices.emea@moodys.comAsia (Excluding Japan) +85 2 2916 1121 clientservices.asia@moodys.comJapan +81 3 5408 4100 clientservices.japan@moodys.com2 JANUARY 2016CAPITAL MARKETS RESEARCHTable of Contents1 Introduction 32 Data and Methodology 33 Validation Results 34 Conclusion 5Appendix A Validation Results by Region by Sector 6North America 6Western Europe 9Japan 12Australia/New Zealand 15References 183 JANUARY 2016CAPITAL MARKETS RESEARCH1 Introduction Stressed EDF credit measures are one-year default probabilities (PDs) conditioned on a hypothetical economic scenario. They depend on measurements of the sensitivity of credit risk to a set of macroeconomic risk drivers, with these sensitivities derived from regression analyses on fixed samples of historical data. Although the relationship between the macroeconomic factors and credit risk can be regarded as structural, these relationships may sometimes change over time, especially when data for a complete credit cycle become available. Moodys Analytics therefore annually reviews the suitability of the model framework and estimated parameters for the current environment. This paper describes the annual validation of the Stressed EDF model.Since Stressed EDF measures are expectations of default risk conditional on the realization of a hypothetical macroeconomic scenario, they are not forecasts per se, and there is little sense in comparing past values with realized EDF measures.1 The performance of Stressed EDFs relative to realized default probabilities is dependent on both the model methodology and the accuracy of the economic forecast. In order to validate the Stressed EDF model, it is necessary to eliminate forecast error due toinaccuracies in the economic forecasts. We do this by calculating Stressed EDFs for a p...
