A Study Of Electricity's Volatility Modeling Based On Realized Volatility

Volatility is one of the most vital studies about financial assets' modeling.In recent years,many countries start pushing on marketization of electricity.Derivatives of electricity become more and more common,which leads to the situation that the study of electricity's volatility becomes more and.more important But electricity's volatility is radically different from other assets because of its non-storage,seasonality and periodicity.So the studies of modeling electricity's volatility catch much attention.There as many methods of estimating volatility of assets,but with the development of computer technology and Availability of high-frequency data,realized volatility,which is computed based on high-frequency data,becomes popular in academe.So models based on realized volatility start increasing sharply.Most of the existing studies of electricity's volatility modeling studies just transpose the models of other assets into electricity market,which seldom points on the characters of electricity such as the "inverse leverage effects".Specific to this question,improvement is done to HAR model,which is one of the common models of realized volatility.LR test,DM test,SPA test and MCS test are used to test the in-sample fitting performance and out-of-sample forecasting performance of the new models.The Empirical results show:1)The volatility of Australia electricity price do have the characters like the"inverse leverage effects",Leptokurtosis and volatility of volatility,which is consistent with other studies.The result of LR test shows the modification of model do improve the data-fitting performance of the model.2)The result of DM test,SPA test,MCS test shows that the modification of model does improve the forecasting performance of the model,which is consistent with the in-sample results.Besides,GARCH model and FIGARCH is more suitable as the model of volatility of volatility.