Var Measuring Models Based On Extreme Value Theory And Empirical Studies

Extreme value theory is increasingly used to study financial risk management in the research area of financial time series. In this paper General Extreme Value (GEV) distribution is mainly employed to analyze a share market, and maximum likelihood is used to estimate the unknown parameters. On the other hand, the assumption of independence in extreme value theory is critical in practice. Taking it into account, a modified VaR measuring model is also constructed and applied to estimate the risk of HS300Stock Index. The modified model is shown to more precisely describe the extreme fluctuations in the actuarial market, which compensates for the underestimate of risk in traditional extreme value theory. It provides a more useful instrument for institutional investors to make profit at least margin transaction guarantee sum.