Credit Risk Assessment and the Opportunity Costs of Loan Misclassification

This paper investigates the opportunity costs of loan misclassification in credit risk assessment, and develops a model that directly incorporates the opportunity costs of loan misclassification into the risk assessment criteria. The performance of this model is assessed by comparing the empirical results with the results of the logit credit scoring model using data provided by the Farm Credit Corporation of Canada. Results indicate that if both the models make a Type I and Type II error with a $1-million loan, the expected costs of error would be $600 and $6, 500 less, respectively, in the cost minimization model compared with the logit model. The expected lenders's profit for a $1-million loan would be $17, 000 more in cost minimization model compared with the logit model. The cost minimization model performed better than the logit model, both in prediction accuracy and also in reducing the costs associated with the errors in classification.