An Empirical Analysis of Credit Risk Measurement of Commercial Banks —— Credit Risk Measurement and Risk Management of Commercial Banks

I. Literature Summary The credit risk of commercial banks mainly refers to the possibility that the borrower is unable or unwilling to repay due to various reasons, which leads to the loss caused by the creditor's default. With the continuous innovation and development of financial industry and financial market, the competition in financial market is becoming increasingly fierce, which makes the measurement of credit risk in financial markets at home and abroad become the focus of many scholars' research.

Credit risk measurement model includes traditional measurement and modern measurement. Common measurement models are: ZETA model, MDA model, Z-scoring model, Logit model and neural network model belong to traditional measurement models; At present, credit measurement model, credit risk+model, KMV model, credit portfolio model and KPM model are the most popular and studied modern measurement models. In the west, the measurement model of credit risk was simulated and summarized in 2000. Gordy and Crouhy(2000) evaluated the static portfolio at the same time, and thought that the data analysis results of different models were close. At the same time, Nickell(2004) made a comparative analysis of the risk measurement model, focusing on the combination data of actual assets, but the results verified that the model was not accurate enough in risk assessment. Similarly, Zhang (2009) and others made a comparative analysis of the credit risk of more than 65,438+0,000 loan enterprises in Chongqing with the data from four state-owned commercial banks in Chongqing. Through the analysis of the data, Zhang got a research result close to nickel(2004), and the research conclusion also showed that the effect of risk prediction based on model deduction was weak. Another point of view is that KMV model proposed by KMV Company of the United States in 1993, that is, using the option idea and excluding the credit monitoring model, shows that KMV model is the most widely used in China practice at present, mainly due to the constant standardization of capital market management in China and the approaching of market value and intrinsic value of listed companies. The premise of applying KMV model is that the data of capital market is more effective and the symmetry of information is higher.

KMV model is widely accepted and used by academic and practical circles. After the emergence of new features in China's financial development, it is of great theoretical and practical significance to study KMV mode in depth. As we all know, the main theoretical basis of KMV model is Merton option pricing theory, and the input variables are the market value and volatility of enterprise equity. At the same time, according to the debt status of the target enterprise, the default execution point is calculated and the default distance is calculated. Then according to the relationship between the default distance and the expected default rate, the default rate of the enterprise is obtained, and the default rate is regarded as the credit risk of the enterprise. Recently, many domestic scholars have also begun to study KMV model in theory and practice for China market. Liu Yingchun (2004) and others agree that the application of KMV model by foreign scholars is based on modern option theory, and on this basis, they predict the relevant disclosure information in the capital market, but the relevant information does not include the historical book reports in the market. The forecast results can truly reflect the credit risk status of listed companies in the current capital market. However, Liu Yingchun's research also has its shortcomings, that is, the application effect of non-listed companies is very poor, mainly for listed companies. Wu (2005) and others think that the EDF index in the model is mainly aimed at displaying the real-time data of stock prices, and it is continuous in measuring the credit risk of companies. At the same time, KMV model provides a reliable credit analysis and measurement model for relevant regulatory authorities and investors. The limitation of this model lies in the lack of historical data of corporate defaults and related bankruptcies in China capital market, and the inability to convert the default distance and expected default rate, thus affecting the wide application of KMV model in China.

To sum up, there are many theoretical research literatures about KMV model in China, but few empirical research literatures about KMV model. Therefore, this paper uses KMV model to predict and estimate the equity value and its volatility of all listed banks in China, and finally obtains the probability of EDF index loss.

Second, the setting of credit risk measurement model and variable selection

Firstly, the credit risk measurement model is introduced. In the 1990s, some financial institutions at home and abroad began to conduct technical research on enterprise credit risk, mainly studying some new credit risk measurement models, including KMV model, credit measurement model and Credit risk+ model. Modern risk measurement model is a theoretical risk measurement based on financial market. At the same time, the method of mathematical statistics and the research method of system engineering are added to the model, which mainly analyzes, identifies, measures and monitors the risks faced by commercial banks. KMV model was put forward by KMV Company in 1989, which mainly provides credit risk management and other services. This model measures credit risk according to the stock price in the stock market. KMV model, also known as credit monitoring model, is a credit monitoring model for estimating rights and interests according to option pricing theory. By analyzing the fluctuation of equity value of listed companies, that is, the intrinsic value of enterprises, the possibility of default of listed companies is predicted. Its application object is listed companies, and the data of listed companies are easy to obtain.

According to KMV model, the relevant indicators are defined: the total value of a company is the equity value (owner's equity) plus the debt value of the enterprise. According to the balance sheet, when the total value of the company is greater than the corporate debt, the creditors of this part of the debt will be fully repaid, and the shareholders will get the residual value; When the total value of the company is lower than the debt value, that is, when the assets are insolvent, the company will be unable to repay the debt, that is, default will occur. At this point, the shareholder value is zero or even negative. Therefore, if the company's total assets are below a certain value, it will cause default risk to shareholders and creditors. At this time, the corresponding total company value is set as the company default point, that is, the point where the company's asset value and liability value are equal.

The model assumes that the distribution characteristics of the company's total assets value in a certain period are described by the expected value (E) and standard deviation (volatility) of the company's total assets. The difference between the average future assets of an enterprise and the book value of its liabilities is set as the default distance. Determine the corresponding linear relationship between the expected default rate and the default distance, that is, use the default distance to represent the default rate, establish a functional relationship and estimate the corresponding default rate. The expected default rate indicates the probability that an enterprise may default in a certain period of time under the normal market environment. The KMV model assumes that when the average value of future assets of an enterprise is lower than the book value of liabilities to be paid by the enterprise, default will occur. Because it is impossible to accurately judge whether the loan enterprise will default, it can only estimate the default risk, that is, the possibility of default.

Second, the parameter estimation of KMV model. Details are as follows:

The first is the estimation of the company's stock value. This paper takes listed banks as the research object, assuming that the asset prices of these 16 listed banks will remain unchanged for a certain period of time. In Merton's model, it is assumed that holding the company's equity is equivalent to holding the company's call option, so the value of the company's debt is also the final exercise price of the option.