Modeling Credit Risk: An Application of the Rough Set Methodology

Abstract

The Basel Accords encourages credit entities to implement their own models for measuring financial risk. In this paper, we focus on the use of internal ratings-based (IRB) models for the assessment of credit risk and, specifically, on one component that models the probability of default (PD). The traditional methods used for modelling credit risk, such as discriminant analysis and logit and probit models, start with several statistical restrictions. The rough set methodology avoids these limitations and as such is an alternative to the classic statistical methods. We apply the rough set methodology to a database of 106 companies that are applicants for credit. We obtain ratios that can best discriminate between financially sound and bankrupt companies, along with a series of decision rules that will help detect operations that are potentially in default. Finally, we compare the results obtained using the rough set methodology to those obtained using classic discriminant analysis and logit models. We conclude that the rough set methodology presents better risk classification results.

Application of the MARS Method to the Evaluation of Grant Applications and Non-Returnable Instruments of Start-Up Business Financing

“…Another advantage of our approach is that it does not require an assessment of the relevance of the criteria (weights) of the credit application evaluations, which could constitute an additional difficulty for the expert. Further research will deal with verification of the empirical usefulness of the model proposed, as well as with identification of other methods using verbal scores, such as MACBETH, ZAPROS, methods based on holistic approach such as UTA, GRIP (Figueira, Greco, Słowiński, 2008, applications of rough sets (Pawlak, 1982;Medina, Cueto, 2013), or fuzzy reasoning (Konopka, 2013) for the evaluation of credit applications.…”

“…Also, the theory of rough sets is used in research on risk involved in start-up business financing (Pawlak, 1982). The decision problem consisting in granting or not granting funding can be represented using a decision system in which the conditional attributes are variables from the model, and the conclusion (the system decision) is a dichotomic variable denoting a "good" client and a "bad" one (Medina, Cueto, 2013). Fuzzy concluding can be a useful tool in the assessment of risk involved in starting an individual business, where those assessing a grant or loan application have limited information on both the applicant and the microeconomic environment of the future businessperson (Konopka, 2013).…”