Credit Risk Measurement And Management Based On MDA And EDF Methodology

Credit risk is the likelihood of the loss occurred to banks, investors or trade counterpart due to the default by debtor borrower(s) or trade counterpart because of unwiliness or inability to fulfill the contract, which includes the likelihood of the loss occurred due to the rating migration and the fluctuation of market value of the asset. This paper, under the framework of commercial bank credit risk management required by 2001 BASEL Agreement, will use the MDA and EDF methodology to analyze the three key issues in credit risk measurement and management, i.e. firm credit rating, credit transit matrix and allocation of economic capital based on the loss given default (LGD). The paper contains eight chapters.Chapter 1 is the introduction of the academic background, the development and the application of the credit risk study and presents the research framework along with the technical route for this study.Chapter 2 outlines the basic concept of credit risk, management theory, the main contents of BASEL Agreement and designs the internal credit risk management system from the point view of commercial bank. In addition, it introduces a variety of the credit risk assessment methodology such as expert method (5Cs, 5W), Zeta model, LPM model, Logit model, neural network, term structure model, mortality rate model, CreditMetrics?, KMV model, Credit Risk~+, Credit Portfolio View and REE.Chapter 3 addresses the credit rating method, its main content, function of rating in credit risk management and the main problems of rating in China.Chapter 4 constructs a 4-variable and 7-variable discriminant analysis model by use of MDA methodology based on financial date of Chinese listing firms and apply them to the corporate rating. The research finds that both of the models constructed have the ability of forecast by 4 years in advance. The creditability of the listing firms are quite good as a whole, but unstable with deterioration in credit risk. Through the observation of 5-year rating transitions of the listing firms, the author constructs the credit transit matrix for Chinese listing firms and finds that the probability of retaining on the initial rating is the largest. The higher ratings (AAA, AA, A) have the least transit probability to default rate (D). The transit probability becomes smaller as the rating classification becomes bigger. However, compared with that of Standand & Pool, the credit transit matrix constructed in this paper shows the features of fluctuation andinstability. In particular, the proportion of the firms at two extremes retaining on the initial rating is quite small.Chapter 5 first made adjustments to the calculation of equity value, the determination of default point for KMV model in accordance with Chinese situation and then applies the adjusted model to the corporate rating. The result shows the model has the ability of forecasting by 4 years in advance for corporate credit risk as a whole and two-year ability of forecasting individual firms in credit risk. Under the three different DPT scenario, the mean-variance of ST firms and non ST firms are becoming large as the approaching of ST year and reaches the largest one year before ST.Chapter 6 makes a comparison study between the MDA and KMV model in the application of credit risk identification. It finds out: (1) the ratings by the two models are rather consistent, but quite different in ratings for the firms at two extremes.(2)Three is a significant negative correlation between theoretic expected default frequency (TEDF) and Z-score, which means firms with larger Z-score are healthier in finance, therefore, TEDF is smaller, credit risk is lower, otherwise, credit risk is higher.(3) Three is a significant positive correlation between Z-score ratings and that of Xinhua Far East China Ratings. The coefficient reaches 0.66. (4) Though DD ratings is positive with Xinhua Far East China Ratings, but not as good as Z-score ratings.Chapter 7 applies the research results to the measurement of loss given default(LGD) and capital allocation for commercial bank loans. In this paper, a systematic evidence study has been carried for a sample of 104 bank loans from the obligators rating, debt rating adjustment, the measurement of loss given default (LGD) and capital allocation based on VaR.The originality of this paper is reflected in the following aspects: (1) Constructs the Z-score model by using the MDA methodology and the financial data of Chinese listings firms, made adjustments to the KMV model in accordance with Chinese situation, then, apply them to the corporate ratings in China. (2) Constructs the credit transit matrix for Chinese listing firms through the observation of 5-year rating transitions of the listing firms. (3 ) The adjustments to the calculation of equity value, the determination of default point for KMV model in accordance with Chinese situation and the application to the corporate rating.(4) A systematic evidence study has been carried out for a sample of 104 bank loans from the obligators rating, debt rating adjustment, the measurement of loss given default (LGD) and capital allocation based on VaR.The research provides an important concept framework and methodology for financial institutions, investors, securities issuer, capital market and regulatory departments in credit risk assessment and management. The application includes: (l)The financial institutions can use the models constructed for corporate ratings, bank loan pricing, credit risk monitoring and regulatory purpose.(2)In accordance with the VaR, banks can re-evaluate those obligators with large value changes and set up a risk-responsibility mechanism for lending and investment. (3) helpful to the determination of the credit line and the efficient allocation of risk capital to meet the regulatory requirements. (4)Firms or investors, through observing the changes of Z-score or DD, can warn the financial distress and take measures in advance to avoid the potential crisis. (5)The government regulatory departments can also, through observing the changes of Z-score or DD, coordinate and optimize the resource allocation so as to reduce the cost. (6)The Z-score model constructed in this paper can help the auditors to lock at quickly those financial firms and facilitate their survey.