The Improvement Of CreditMetrics Based On Mixture Vector Autoregressive Model

Commercial banks bear the credit risk management,so how to prevent and resolve the credit risk has become issue. With the further opening, only if the commercial banks in China establish the credit risk system as soon as possible,will they enhance competitiveness and achieve sustained healthy and stable development.As the bank interior to guard against and dissolve the risk managemen-t tool, the economic capital measurement is the most important. The so-called economic capital is to protect the bank in1year against the economic bankruptcy risk required quantity of capital. At present, the credit economic capital requirement of measurement mainly has two kinds:one kind is capital coefficient method. the single position value is multiplied by a credit capi-tal factor to get the economic capital, then add all positions.Credit capital factor according to the ratings and the duration of the predetermined, this method is simple, but the lack of flexibility.Another method is the portfolio simulation method.One of the most representative of the Morgan JP’s Cred-itMetrices model. CreditMetrices is based on the value distribution of credit risk VaR model. Its core is the credit rating credit migration matrix. Credit migration matrix is used with the historical average credit migration matrix. This method is based on external rating, external rating is generally from the long-term angle across the cycle of enterprize rating, more is to use qualitative method. And the same ratings of different sectors of the actual default rate is not the same. From the angle of capital management, cross cycle rating will be even in the next year the actual default rate, easy to cause the overall credit economic capital underestimate or overestimateBased on the CreditMetrices model based on external rating of the histor-ical average default rate of disadvantage, put forward adopt the internal rating method instead of external rating method. The internal rating method based on point rating method, prediction of the default rate with the economic cycle is changed. the actual reflection of the next years debt issuer credit changes. This paper is based on the mixture vector autoregressive model of internal rat-ing method.Using a mixture vector autoregressive model to the bond. implicit default rate and macroeconomic variables for modeling, dynamic simulation have been predicting one-year default rate. Notes from the implied default rate structure are long-term default rate structure, will be obtained to predict the default rate and1year period of structure analysis, a simulation of the future credit status.The simulation can be obtained after the credit migration matrix. Will get credit migration matrix is substituted in the framework of CreditMetrics can calculate the the distribution of credit losses, and get bond portfolio credit VaR,the next year the economic capital.