A credit risk model for agricultural loan portfolios under the new Basel Capital Accord

摘要

The New Basel Capital Accord (Basel II) provides added emphasis to thedevelopment of portfolio credit risk models. An important regulatory change in Basel IIis the differentiated treatment in measuring capital requirements for the corporateexposures and retail exposures. Basel II allows agricultural loans to be categorized andtreated as the retail exposures. However, portfolio credit risk model for agricultural loansis still in their infancy. Most portfolio credit risk models being used have been developedfor corporate exposures, and are not generally applicable to agricultural loan portfolio.The objective of this study is to develop a credit risk model for agricultural loanportfolios. The model developed in this study reflects characteristics of the agriculturalsector, loans and borrowers and designed to be consistent with Basel II, includingconsideration given to forecasting accuracy and model applicability. This studyconceptualizes a theory of loan default for farm borrowers. A theoretical model isdeveloped based on the default theory with several assumptions to simplify the model.An annual default model is specified using FDIC state level data over the 1985 to2003. Five state models covering Iowa, Illinois, Indiana, Kansas, and Nebraska areestimated as a logistic function. Explanatory variables for the model are a three-yearmoving average of net cash income per acre from crops, net cash income per cwt fromlivestock, government payments per acre, the unemployment rate, and a trend. Net cashincome generated by state reflects the five major commodities: corn, soybeans, wheat,fed cattle, and hogs. A simulation model is developed to generate the stochastic defaultrates by state over the 2004 to 2007 period, providing the probability of default and theloan loss distribution in a pro forma context that facilitates proactive decision making.The model also generates expected loan loss, VaR, and capital requirements.This study suggests two key conclusions helpful to future credit risk modelingefforts for agricultural loan portfolios: (1) net cash income is a significant leadingindicator to default, and (2) the credit risk model should be segmented by commodityand geographical location.