The Applicability Of KMV Model In China

The definition of credit risk is divided into two categories:one was proposed by the traditional risk management theory which had many shortcomings, such as delay and passive effect. As a result, a new type of definition aroused among the modern risk management which said:Credit risk is the risk of an economic loss from the failure of a counterparty to fulfill its contractual obligations or the credit rating transfers, so that the value of a financial product changes and causes a loss for those product holders. The new definition is no doubt more accord with modern financial market, for it views credit risks as dynamic processes which makes it possible for us to measure them with market risk measure approaches.This paper describes the history of credit risk, the models used to measure it and the empirical study. The main purpose of this paper is to prove the applicability of KMV model in China. I used the figures from automotive industry in A-share market, solved the default distances of all companies (viewed as the measurement of credit risk) and compared them with domestic credit risk data. I concluded that KMV model can, to some extent, forecast the credit risk but no special superiority is observed.In the macro part, the paper first introduces the categories of credit risk, and then puts forward three important factors:Default Probability, Loss given Default and Credit Exposure. Default Probability is a focus, and I use both Actuarial method and market price method to calculate it. Finally, some appraisals and suggestions about applying credit models in Chinese market are listed. These contents are covered in chapter two and six.In the microscopic section, I elaborate credit risk models——first traditional models and then modern models. Traditional models depend on the experts' judgments, including expert analysis, credit ratings and credit score methods. The modern credit risk models, however, starting in 1990th, focus on objective math models such as Credit Metrics, Credit Risk+, Credit Portfolio View and KMV model. These contents are covered in chapter three.In the theoretical part, the paper explains Merton model, which leads to the generate of KMV, in detail. The basic idea of B-S model, mathematical statistics and Ito lemma are employed when deriving DD. I also use GARCH (1,1) model to evaluate volatility of stock returns. These contents are covered in chapter four.In the empirical part, I simplify some parameters so that they meet better with domestic circumstances. As for the estimation of variables, I search references and select those optimum methods:volatility of stock returns is estimated with GARCH (1,1) model; I use the regular one-year (2008)weighted average bank deposit rate as the risk-free interest rate; tradable and non-tradable shares are valued in different prices and several exciting forms of default point are considered and compared. In the process, I use matlab to realize the Iterative algorithm. These contents are covered in chapter five.There are several innovations in this paper:First, a different angle is chose to study KMV model. The samples are limited in the automotive area, where I serve the debt risk as the replacement of credit risk. These figures can then be used to compare with DD. The advantages of this method are obvious:The forecast accuracy of KMV model is tested without the interference of industrial differences and artificial selection factors. Second, I detect different setting methods of default point. Third, the whole paper is organized from macro-level to the micro-level.