The Study Of Conditional Value At Risk Based On Realized Volatility

With the rapid development of financial markets, people will face more complex risks, how to accurately measure the risk is the problem people need to solve. Among them, the measure of financial market risk is particularly important.The quantitative study of Value at risk (VaR) can measure the greatest possible losses of financial assets during a given period, but the current researchs on the VaR measure need to be further deepened. The traditional methods only focused on the Value at risk of assets or portfolio, without considering a number of other factors that might impact Value at risk. The various financial risk factors influence and spread each other, so their measures need to be considered under given conditions. The Realized volatility model based on high frequency data is a volatility measurement model, including more market information, can better reflect the market characteristics. Therefore, the study of Conditional Value at Risk based on the Realized Volatility has a great significance.This paper studied the several common VaR estimations and their advantages and disadvantages, as well as the current conditional VaR, then discussed the theoretical background of Realized Volatility, proved that the realized volatility has been the unbiased estimator of actual volatility, furthermore, analyzed the selection ideas of the Realized Volatility's optimal time interval, found the Realized Volatility's optimal time interval of China's Shanghai stock markets is 15 minutes based on Rusell and Bandi's research by using 1 minute transaction data.Then, this paper analysised the distribution characteristics of the return ratio series and realized volatility by use of Copula methods, analyzed the dependency structure between them, and then estimated its parameters to calculate the conditional VaR based on the realized volatility.