An Empirical Study On Credit Risk Measurement Of Listed Companies Based On KMV Model

In recent years,with the further development of China's financial market,the credit risk problem has become increasingly prominent,credit risk analysis,measurement and control is relatively lagging behind,the current domestic credit risk analysis method can not solve China's deepening and complex Financial market issues.How to measure the credit risk of our government regulatory authorities,major financial institutions and a majority of investors and important research topics.One of the most frequently studied and used forecasting models is the KMV model,which is derived from the ground-breaking work of Merton(1974)and developed by commercial companies--KMV.The model treats the firm's debt as a call option with its asset value.When the debt is paid at maturity,its asset market value is lower than the liability,the firm will choose to default.The distance between the value of the enterprise asset and the default is called the default distance.From the historical database,the model can map the expected default probability and the actual default probability through the default distance.As the main use of the stock market in real-time trading data,reflecting the many investors on the company's expected judgments,making the model of credit risk assessment has a certain forward-looking,can better measure and predict the company's credit risk.This paper first introduces some foreign modern credit risk measurement models,focusing on the KMV model.Combined with the domestic A-share market,conducted empirical analysis,taking into account the traditional method of EM iterative method and parameter correction.Finally,the KMV model is discussed and studied under the BSDE and non-linear expectation framework.In this paper,we extract the listed companies listed in the A shares for more than one year in 2016 and the relevant data are relatively perfect.The traditional two-party legislation is used to calculate the default distance and the expected probability of default.Statistical analysis of the overall cross-sectional data,with low rating and ST as a reference.Choose two tier-one industry(real estate and construction)and a secondary industry(ferrous metal smelting and rolling processing under manufacturing)listed for 10 years listed companies as a sample,using single-equation EM iteration method,the specific industry Time series analysis,the industry default distance,expected default probability,credit transfer matrix,industry average asset yield and default distance sequence changes.In the empirical process,from the calculation and analysis found:First,the single-equation EM iterative calculation method is feasible and reliable,and to some extent better than the simultaneous equation method,there are direct use of raw data,information loss is small,It does not need to consider the domestic market on the volatility of assets and stock volatility between the structural relationship between the rationality,more reasonable in the quarter of the default number of default as a quarter of default value.Second,through the ST company and credit Low-rated companies to replace the lack of default cases and database default parameters estimation and correction is feasible,but need to be targeted.In the current stock market,as long as investors have affirmative,some of the above sample of its equity value has not rapidly depreciated,and then take into account the small companies of their very low debt data and the market is not active in the transaction,and the sample size is too small Factors easily lead to their relatively large distance,can not take these samples as the industry benchmarks and flags to estimate the parameters and adjust the model,but the need for specific analysis of the sample.