Credit Risk Modeling: Spline based logistic regression Survival approach

Subject Areas : Journal of Investment Knowledge

Mohammad Ali Rastegar ¹ (Financial Engineering Group, Industrial & Systems Engineering, Tarbiat Modares University)
Mahdi Eidi Goosh ² (Financial Engineering Group, Industrial & Systems Engineering, Tarbiat Modares University)

Received: 2018-01-15 Accepted : 2018-07-25 Published : 2020-08-22

Keywords: spline-based logistic regression survival, credit risk, Cox regression, ROC,

Abstract :

Nowadays, banks in the country are faced with serious problems in terms of their assets. One of the factors that led to this situation is the poor quality of banks' assets, which can be attributed to the lack of a rating system and an improper assessment of credit risk. This study predicts the probability of default during a specific time using the Cox regression model as well as the survival model of spline-based logistic regression. For modeling of credit risk, using these two methods, 10 variables related to 2861 customers of an Iranian bank were used. We compared two models using ROC method, the Cox regression model with AUC = 0.799 was more efficient than the spline-based logistic regression model with AUC = 0.746.

References:

[1] حامد کرانی و مولود آقایی‌پور. (1390). کاربرد نظریه تحلیل بقا در مدیریت ریسک اعتباری دریافت کنندگان تسهیلات؛ مطالعه موردی بانک مسکن. فصلنامه روند، شماره 65، 175-200.

[2] مرضیه ابراهیمی و عبداله دریابر. (1391). مدیریت ریسک اعتباری در نظام بانکی- رویکرد تحلیل پوششی داده ها و رگرسیون لجستیک و شبکه عصبی، فصلنامه دانش سرمایه‌گذاری سال اول/ شماره دوم/ تابستان 1391.

[3]        Narain, B. (1992). Survival analysis and the credit granting decision. In: Thomas L., Crook, J. N. and Edelman, D. B. (eds.). Credit Scoring and Credit Control. OUP: Oxford, 109-121.

[4]        Cao, R., Vilar, J.M., Devia, A., Veraverbeke, N., Boucher, J.P. and Beran, J., 2009. Modelling consumer credit risk via survival analysis.

[5]        Miller, S., 2010. Risk Factors for Consumer Loan Default: A Censored Quantile Regression Analysis.

[6]        Stepanova, M. and Thomas, L., 2002. Survival analysis methods for personal loan data. Operations Research, 50(2), pp.277-289.

[7]        Luo, S., Kong, X. and Nie, T., 2016. Spline based survival model for credit risk modeling. European Journal of Operational Research, 253(3),.869-879.

[8]        Dietz, K., Gail, M., Krickeberg, K., Samet, J. and Tsiatis, A., 2002. Statistics for Biology and   Health. Survival Analysis, Edition Springer.

[9]        Malik , M . and Thomas L . (2006). Modeling Credit Risk of Portfolio of Consumer Loans . University of Southampton , School of Management

 [10]     Stein , R . (2005). The Relationship Between Default Prediction and Lending Profits : Integrating the ROC Analysis and Loan Pricing. Journal of Banking and Finance.Vol. 29, pp 1213-1236.

[11]      Blöchlinger , A . and Leippold , M . (2006). Economic Benefit of Powerful Credit Scoring. Journal of Banking and Finance.Vol. 30, pp 851-873