The Construction And Application Of Credit Rating Transition Prediction Model

Since the 1990s, as the key part of financial risk, the credit risk continues to affect the global economy. How to effectively manage credit risk become the most popular topics in every area. So far, the Basel Committee, the world's major banks and rating agencies are working hard to develop a variety of credit risk measurement models. The credit rating is mature method with a long history, which results are generally accepted and the credit rating changes foreshadowed the changes in credit risk to some extent. Therefore, through by building a credit rating transition matrix, we can identify the future trends of credit rating, and thus predict and control the dynamics of credit risk.Based on the concepts related to credit risk, credit rating and credit rating transition, this article learns the research results related to the credit risk management, credit rating transition matrix from their predecessors, finally proposed credit rating transition matrices to predict the dynamic changes. The Markov model and semi-Markov model was constructed as credit rating prediction model, in which Markov model is of more mature, but exists some problems in applicant process, and semi-Markov model is a improvement to Markov model. To compare the effect of these two models and simulations in the Chinese market, the two models were applied to the average one-year transition matrix of Moody's credit rating and the average one-year credit rating transition matrix of the Grand Duke. For the results of Markov model, this paper will gives the trend analysis for the credit rating transfer and the default time prediction. In addition to the corresponding analysis, for the results of semi-Markov model, this paper adds a transition probability analysis of specific residence time based on the statistical properties of semi-Markov. It finds that semi-Markov models predict more stable than the Markov model and is more suitable for China's stable economic development. Because of the Chinese credit system with late construction, inadequate credit database, the two models have better simulation results in foreign credit data. In the End of the article, the author raises an outlook on the research.