The Theory Of VaR With Applications To China's Financial Market

This thesis analyzes and summarizes systematically the methodological systemof VaR, which includes the natal background and developmental progress, basicmeaning , calculational principle, portfolio VaR (Marginal VaR, Component VaR ,Incremental VaR), the advantage and disadvantage of VaR, and introduces in detailthree primary calculational methods of VaR ( history/historical simulation method,parameter method (the Variance-Covariance method), Monte Carlo simulation method)and analyzes comparatively their respective strongpoints or shortcomings andapplicability, also introduces some new developmental state of VaR(ES, CVaR,CDaR). CVaR is a new concept of portfolio optimization and has some advantagesover the traditional VaR in linearity and sub-additivity.In addition, the thesis makes demonstrational analysis of the application of VaRto China's financial market.Firstly, the dissertation chooses four exchange rates of (CAD,EUR,GBP,JPY)relative to USD as analytical objects and calculates the daily VaR of these exchangerates through the six methods, which are History Method, Monte Carlo Simulation,Normal Method, Simple Moving Average, RiskMetrics, GARCH Parameter method.The results suggest that the values of VaR from the first three methods and the rest arequite same, with the former being little less than the latter. Furthermore, the GARCHMethod (GARCH(1,1)-M & TGARCH(1,1)) achieves a satisfying simulation to thefluctuation of the four exchange rates.Secondly, in the analysis of application of VaR to portfolio, the thesis adoptsRAROC approach to evaluating the performance of funds, which considerssynthetically the two factors: risk and profit, then makes demonstrational analyses ofthe VaR-based Markowitz model and CVaR-based portfolio optimization model. Theresults show that under the same profitable condition, the portfolio risk decreases afterappending the VaR-restriction;that VaR is also optimized when CVaR is optimized;that the Standard Variance and VaR from CVaR-based model are less than those fromthe other models;and that CVaR model can reduce further the portfolio risk.The demonstrational analyses indicate that VaR is a very useful tool of riskmanagement. It's very important and practical to introduce VaR into China's financialrisk management.