The Study Of Credit Risk Of Chinese Commercial Banks Based On CPV Model

The business character of commercial banks determines their character that take risks and gain profit depend on operating risks. During the last semi century, the international banking have experienced large changes of risk management methods. The promulgating of Basel agreement in 1988 marked the all-around risk management principle system of international banking had been mainly formed. As to the state of Chinese commercial banks, credit risk is still the largest and uppermost risk they faced. The study and management of credit risk is quite important for the development of commercial banks.While the risk management method developing, the risk measure methods are developing fast. Accompany the innovation of mathematics and the developing of financial community, the measure methods of credit risks have experienced the process from simplicity to complexity. When the international risk management technology keeps developing, the credit risk management technology in China still falls behind. Most commercial banks are still using financial factors even subjective judgment to analyze risks. The result is that NPAR in China keeps a high level. The condition is not better until the four assets management companies established in 2000. But peeling off NPAR is not a fundamental way to solve the risk management problem. The commercial banks should improve their management level to resist risks and make assets better.Especially nowadays, the financial storm is sweeping all over the world. Economic crisis is influencing every aspects of our life. Trace it to its cause, it was that the money corporations haven't control the risks at first. And then the financial derivatives make the risks larger and lager and make serious result. Today, many countries are experiencing economic contraction, the unemployment rate is higher and higher, and even the public order is worse and worse. So, controlling the risks is important to politics, economy and society.Nowadays, people use the standpoint of credit portfolio to analyze credit risk. The paper contact three of them, and choose the Credit Portfolio View(CPV)model to make further analysis. CPV model is put forward by McKinsey in 1997. The model considers that the transition probabilities of credit degree at different periods are not fixed, but affected by macroeconomic factors such as countries, economic cycle, unemployment rate, increment speed of GDP, long-term interest rate, rate of exchange, government expenditure, aggregate savings, industry factor, and so on. Suppose if macroeconomic variables are AR(2). This is a Logistic probability model.Macroeconomic index in model is given by multifactor model. Macroeconomic variables are specific to each country. When we have enough data, the model can be calibrated in different countries (or industries).The paper makes a discussion about the data of PD (probability of default). The PD means probability of default. The Basel agreement makes relevant definition of PD. However, the industry of Chinese commercial banks do not have united definition of default, the government also do not have relevant regulation. The Commercial Bank of China still uses five-category assets classification for the main method for risk management. To compare with the definition of five-category assets classification, the PD is similar definition. The paper use CPV model to research the relation between PD and macroeconomics factors, however, not the exact measure of the PD, so the paper choose the probability of the non-performing loan instead of PD.The paper makes empirical analysis about the PD of whole country instead of the analysis of the whole industry. The combination of the macroeconomics index Y for the macroeconomics indexes is chosen mainly according to the requirement of the CPV model and the difficulty of procure of the data,. According the requirement of the cycle analysis, paper also chooses some common indexes of cycle analysis.Initially, the paper choose 12 indexes including GDP, increment speed of GDP, government expenditure, increment speed of industrial added value, investment in the fixed assets, annual interest rate, foreign investment in actual use, gross income per capita in one quarter of city, total retail sales of consumer goods, M2, rate of exchange, CPI. Finally, by using the test of multi-co linearity,7 indexes including increment speed of industrial added value, investment in the fixed assets, foreign investment in actual use, gross income per capita in one quarter of city, total retail sales of consumer goods, M2, CPI are chosen. By the test of serial correlation and adjustment of regression formula, degree of fitting is very high. That's mean these seven macroeconomics index and relevant coefficients could be use to explain the macroeconomics index Y well, and then explain the PD. At last, we used the model to forecast the probability of default in the 3rd quarter of 2008. The difference between estimated value and true value is quite little. It has also proved the validity of the model.According the analysis, after CPI and seasonally adjustment, the paper consider eight indexes, including increment speed of industrial added value, investment in the fixed assets, foreign investment in actual use, gross income per capita in one quarter of city, total retail sales of consumer goods, M2, CPI, could explain the macroeconomics index Y well, goodness of fittest has in excess of 95 percent.. And then the macroeconomics index Y has direct mathematics relation to translate to PD, so these seven indexes could also have high level of goodness of fittest to the PD. The forecast has also proved the conclusion.That mean the macroeconomics have direct influence to the PD. To sum up, the CPV model is applicable to macroeconomics circumstances of China, especially the part of regression.The paper has two original ideas. Firstly, the paper do not research the PD with different industries, however, the paper consider general status of PD of the commercial bank of China. Secondly, after initially choosing some relevant index , the author use the multiple regression to sift the index. The advantage of the process make sure the indexes have high correlation with the PD, that's different with other method in the past.The paper also has its own deficiency. Firstly, the paper do not have exact definition of data of PD which need industry of Commercial Bank and government provide clear definition. Secondly, the paper do not testy the all steps of CPV model, and the paper also do not discuss about the transfer matrix.