ABSTRACT
Since the 1990s, the value at risk (VaR) has become a new way to measure risk. In the process of risk measurement and portfolio strategy selection, the most important thing is to determine the distribution of assets, and we generally use the normal distribution to characterize it. However, when the kind of assets is increasing and the same kind of assets has different marginal distributions, traditional assumptions will generally fail. By introducing the Copula technology into the financial markets, you can have a deeper understanding of dependencies between financial assets and make better predictions of risk. If you want to predict the risk, it is the most appropriate to introduce the GARCH volatility model which is the most commonly used model to reflect characteristics of the market changes, and it can effectively capture the clustering and heteroscedasticity phenomenon of the volatility of return on assets. After the combination of them, we can not only analyze the VaR of portfolio through diversified distribution, but also effectively
capture the nonlinear relationship between financial markets.
