An Empirical Study Of VaR Techniques Applying To Risk Analysis In Stock Market

In recent years, as a results of many factors like the liberalization of the finance, globalization of economy, competence and unregulation, innovation and technology advancement and so on, fundamental changes have been made in financial market, such as the enlarged scale of financial market, the improved efficiency and the innovation of financial instruments, and as well as the rapidly increased volatility and market risk. Based on that situation, scholars home and abroad have realized that how quantify and control risk is indispensable.This paper provides an overview of a series of typical crisis events home and abroad and comparison of techniques in risk management, in order to indicate that VaR technique has been widely used in financial organization and supervision. So far, both academic scholars and practitioner in financial organization have not made a uniform cognition on VaR technique and inaccuracy has been existed in every technique. This paper involves a brief introduction of principal of VaR, the procedure of measuring VaR and the testing of effectiveness on VaR. Two kinds of widely used VaR techniques are introduced in the paper, namely normal distribution and ARMA-GARCH model. We firstly apply the simplest normal distribution to estimating VaR, then because of the observed volatility clustering and leptokurtosis and fat-tail of the distribution of the returns, the relatively accurate ARMA-GARCH model is been adopted to quantify the value. And with the back testing of Kupiec, it has been proved that ARMA-GARCH model can better reflect the distribution of the daily returns and describe the volatility of financial markets.In order to improve the risk measurement techniques, we must try our efforts to overcome substantial difficulties.The framework of the paper is as the follows:The purposes, the meanings of the paper and the research accomplishments home and abroad have been introduced in the first chapter. The second chapter involves the definition and the sources of risks, the classification of the risks, and a summary of the current situation of our stock market, showing that effective the risk measurement techniques are indispensable.The evolution of the risk measurement techniques have been introduced in the third chapter. Firstly, a number of basic techniques are summarized, such as mean-variance technique, sensitivity analysis and volatility analysis. Then the paper emphasize two kinds of VaR techniques, based on the normal distribution and the ARMA-GARCH model, evaluating the corresponding the VaR.The fourth chapter is to make a contrast of the two VaR techniques with the back-testing of Kupiec, indicating that the ARMA-GARCH model is better suited to describe the returns of stock market in china.The drawbacks of my research are listed in the end, and meanwhile guide to improve the VaR techniques in the future.