Evaluating the Likelihood of Default on Delinquent Loans

Abstract

N The literature on credit and default risk management [2, 3, 6, 7, 9, 10, 11, 13] focuses primarily on the credit-granting decision. Yet, in managing loans and receivables, the point at which a loan becomes past due also requires a decision by the creditor. As summarized in [11], credit-granting decision models are often criticized for (i) using subjective methods for selecting customer attributes to be used in the model; (ii) relying on qualitative attributes; and (iii) relying on incomplete information (i.e., they are characterized by information asymmetry). Managing past due accounts presents similar problems; decision makers must choose among alternative responses, often relying on subjective judgment and incomplete information. Little research has focused on identifying delinquent borrower or loan attributes significantly related to the probability of default. Previous analyses of delinquencies used borrower information and contract terms when the loan was granted as explanatory variables [6, 8, 14]. The results provide insufficient information for choosing the most effective response to delinquent accounts.