;Logistic ; Credit Risk Beaver [3] ( ; ; ; ); [1] [2]

Transcription

1 1,2 3,4 1 (1., ; 2., ; 3., ; 4., ),, ; ;Logistic ; [1] Credit Risk [2] Beaver [3] Altman [4] 5 Z-score 1977 Altman [5] ( ; ; ; ); ;,, ; -, ;, 78 Vol.24 No

2 Z ZETA 1995 Altman [6] Z Z 2002 Shi [7] Martin [8] logit Ohlson [9] Platt [10] logit 1992 Lawrence [11] logit Coats [12] Zhang [13] [14] [15] [16] [17] [18] 1333 on-line analytical mining OLAM [19] AHP-ANN MANAGEMENT REVIEW Vol.24 No

3 Resident 1 Age Education 1 Company 1 Income Average Income Monthly repayment to household income Loan to household income Card House 1 1 Unemployment Insurance 0 Medicare 0 Pension 0 Housing Fund Cumulative number of loans overdue The number of consecutive late Credit Record Providing security for others 2 0 Payment Ratio Vol.24 No

4 / / 4 / / / / 4 + / / /[ + )/2] 3 /[( + )/2] / ( - / - / 4 - / - / 7 )/2] / + )/2 / + )/2 / + )/2 7 / / / 2 / : =( - )/( - ),, 2 : =( - )/ MANAGEMENT REVIEW Vol.24 No

5 1 Logistic Logistic Logistic y i 0-1 y i =0 y i =1 Py i =1 x i 1 P logit P/1-P x i 2 Logistic x i Py i =1 x i = x i Py i =1 x i = Σb i x i =b 0 +b 1 x 1 +b 2 x 2 + +b i x i 2 P= 1 1+exp[- Σ b i x i ] Ln P 1-P =Σb ix i [20] MCLP MCLP {G 1 G 2 } Tr={G 1 G 2 } n Ai=1 i n r b G 1 G 2 X=x 1 x 2 x n R r 3 A i G 1 A i X<b 3 A i G 2 A i X b α i b A i G 1 G 2 A i b α i A i X-b β i b A i G 1 G 1 A i b β i A i X-b fα α i gβ β i 4 A i b X A i X P MCLP Minimize fαand Maximize gβ 4 A i X-α i +β i -b=0a i G 1 A i X+α i -β i -b=0a i G 2 α i β i 0i=1 n 82 Vol.24 No

6 Logistic SAS :1 2:1 1:1 Logistic :1 2:1 4: % 83.82% 82.52% 83.50% 77.67% 72.82% 87.38% 86.89% 84.95% AIC SC LOGL Deviance Pearson 1 1 <.0001 Likelihood <.0001 <.0001 < Logistic Pr>ChiSq Intercept Resident Age Education Company Monthly repayment to monthly household income <.0001 Loan to household income <.0001 House <.0001 Unemployment Insurance <.0001 Housing Fund Cumulative number of loans overdue Number of consecutive Late Credit Record Down payment ratio <.0001 Providing security for others R Square (12 ) (3 ) MANAGEMENT REVIEW Vol.24 No

7 Ln P 1-P = x x x x x x x x x x x x x x 14 P= 1 1+exp[- Σ b i x i ] 5 / / / / K-S ROC 85.44% 83.5% K-S 0.62 ROC K-S ROC % 83.50% 87.38% K-S 4 ROC 84 Vol.24 No

8 ST ST ST ST ST ST 20 H 20 ST 20 7 MCLP 90% 95% 92.5% 4 K-S ROC 5 6 K-S 0.7 ROC MCLP ST H ST H % % % 19 95% % % % 18 90% % % % 18 90% % % % 17 85% % % % % % % % % % % % 19 95% % % % 17 85% % % % % % % % 19 95% 98.00% % 90.00% 94% 8 MCLP Logistic MCLP Logisitc 8 KS ROC MCLP 90.00% 92.5% Logistic 78.75% 82.5% MCLP K-S 6 MCLP ROC MANAGEMENT REVIEW Vol.24 No

9 MCLP Logistic 1 Logistic / / MCLP Logistic [1]. [M]., 2004 [2]. [3] Beaver, W. Financial Ratios as Predictors of Failure[J]. Journal of Accounting Research, 1966,(4): [4] Altman, E. I. Financial Ratios Discriminant Analysis and the Prediction of Corporate Bankruptcy[J]. Journal of Finance, 1968,23(4): [5] Altman, E. I., Haldeman, R., Narayanan, P. Zeta Analysis a New Model to Identify Bankruptcy Risk of Corporations[J]. Journal of Banking & Finance, 1977,1(1):29-54 [6] Altman, E. I., Saunders, A. Credit Risk Measurement: Developments over the Last 20 Years[J]. Journal of Banking & Finance, 1997, 21(11): [7] Shi, Y., Peng, Y., Xu, W., Tang, X. Data Mining via Multiple Criteria Linear Programming: Applications in Credit Card Portfolio Management[J]. International Journal of Information Technology and Decision Making, 2002,1(1): [8] Martin, D. Early Warning of Bank Failure: A Logit Regression Approach[J]. Journal of Banking & Finance, 1977,1(3): [9] Ohlson, J. A. Financial Ratios and the Probabilistic Prediction of Bankruptcy[J]. Journal of Accounting Research, 1980,18(1): [10] Platt, H. D., Platt, M. B. A Note on the Use of Industry-relative Ratios in Bankruptcy Prediction[J]. Journal of Banking & Finance, 1991,15(6): [11] Lawrence, E. L., Smith, S., Rhoades, M. An Analysis of Default Risk in Mobile Home Credit[J]. Journal of Banking & Finance, 1992,16(2): [12] Coats, P. K., Fant, L. F. Recognizing Financial Distress Patterns Using a Neural Network Tool[J]. Financial Management, 1993,22 (3): Vol.24 No

10 [13] Zhang, G. P., Hu, M. Y., Patuwo, B. E., Indro, D. C. Artificial Neural Networks in Bankruptcy Prediction: General Framework and Cross-validation Analysis[J]. European Journal of Operational Research, 1999,116(1):16-32 [14],. [J]., 1999,13(1):5-8 [15],. [J]., 1999,(9):25-32 [16],,. [J]., 2005,7(2):81-85 [17],,. [J]., 2000,(3):50-55 [18],. OLAM [J]., 2004,(10): [19],. AHP-ANN [J]., 2004,9(3):94-98 [20] Shi, Y., Wise, M., Luo, M., Lin, Y. Data Mining in Credit Card Portfolio Management: A Multiple Criteria Decision Making Approach[C]//Koksalan M., Zionts S(Eds.). Multiple Criteria Decision Making in the New Millennium, Springer, Berlin, 2001: An Research on Customers Default Rate of Commercial Banks in China Based on Credit Scoring Models Wang Ying 1,2, Nie Guangli 3,4 and Shi Yong 1 (1.Research Center on Fictitious Economy and Data Science of CAS, Beijing ; 2.Treasury Department, the Export-import Bank of China, Beijing ; 3.Guanghua School of Management, Peking University, Beijing ; 4.Postdoctoral Program of Agricultural Bank of China, Beijing ) Abstract: Nowadays commercial banks are facing more and more complex risk factors. Credit risk is the most important and complex one. Basel II presents two methods on credit risk management including Standard Approach and Internal Rating-Based Approach (IRB). It also points out that banks which are suitable to do so should apply the IRB approach to estimate the probability of customers default by constructing models on historical data. Taking the IRB approach and the actual situation in China into consideration, this paper discusses problems of estimating the probability of customers default. Key words: personal default rate, corporate default rate, logistic, MCLP MANAGEMENT REVIEW Vol.24 No