Conditional Probability of Default Methodology

This paper presents the Conditional Probability of Default (CoPoD) methodology for modelling the probabilities of loan defaults (PoDs) by small and medium size enterprises (SMEs) and unlisted firms as functions of identifiable macroeconomic and financial variables. The process of modelling PoDs represents a challenging task, since the time series of PoDs usually contain few observations, thus making ordinary least squares (OLS) estimation imprecise or unfeasible. CoPoD improves the measurement of the impact of macroeconomic variables on PoDs and consequently the measurement of loans’ credit risk through time, thereby making a twofold contribution. First, econometrically, it recovers estimators that show greater robustness than OLS estimators in finite sample settings under the Mean Square Error criterion. Second, economically, on the basis of economic theory and empirical evidence, CoPoD can incorporate a procedure to select a relevant set of macroeconomic explanatory variables that have an impact on the PoDs. We implement CoPoD with information from Norway and Mexico. ∗The author would like to thank Paul Castillo, Jon Danielsson, Jeremy Grant, Charles Goodhart, Joachim Inkmann, Robert Kosowski, Dennis Kristensen, Ryan Love, Carlos Montoro-Llamosas, Bob Nobay, Pablo Padilla, Andrew Patton, Francisco Penaranda, Ashley Taylor and Mercedes Vera Martin for helpful discussions and useful comments. The author would also like to express gratitude to Eivind Bernhardsen, Kjersti-Gro Lindquist, Thorvald Grung Moe and Kjell Bjørn Nordal at the Norges Bank, as well as to Claudio Borio and Philip Lowe at the BIS for access to their databases, their encouragement and enthusiasm. The views expressed in this paper do not necessarily reflect the views of the IMF, the Norges Bank, the CNBV or the BIS.