Study On The Volatility Forecasting Of Crude Oil Market Considering Structural Break

Crude oil,as a type of energy and chemical material,plays an indispensable role in the development of the global economy,and the oil price often arises large fluctuations due to various emergencies in the past decade.Therefore,it attracts much attention worldwide to explore how to forecast the oil price volatility accurately,and the structural break has become an important factor affecting oil price volatility forecasting.However,there are still many doubts and problems in the research on the volatility forecasting of crude oil market considering structural break.Most methods usually regard structural breaks as being sharp and pure jumping,but some scholars propose that it is also reasonable to model the structural breaks as being smooth rather than sharp.Therefore,it is necessary to systematically consider the effect of different ways to solve the structural break to the forecasting performance of the volatility models.By these ways,this paper can systematically analyze and compare the international oil price volatility forecasting models considering structural break,and then look for the best model.Therefore,this paper chooses two methods to depict the structural break,which regard the structural break as a smooth or sharp process respectively.Based on this the paper employes the HAR-type models which using high-frequency data and GARCH-type models which using low-frequency data respectively in the third chapter and fourth chapter,to systematically judge the effect of different ways to solve the structural break to the forecasting performance of the volatility models.The main empirical results indicate that,firstly,the crude oil markets exhibit clear structural breaks behavior,and the incorporation of the structural breaks can significantly improve the forecasting performance of the GARCH-type models and the HAR-type models.It should be noted that a small number of the low-frequency trigonometric terms can accurately capture the structural changes in crude oil price volatility.Secondly,the Flexible Fourier Form(FFF)GARCH-type models and Flexible Fourier Form(FFF)HAR-type models considering smooth shift can accurately model structural breaks and yield superior fitting and forecasting performance to traditional models.Meanwhile,the forecasting results of these models usually outperform the forecasting results of the models which regard structural breaks as being sharp,and it indicates that the incorporation of smooth shift can better capture various degrees and forms of structural changes such as smooth structural changes and structural breaks incrude oil price volatility,thereby improving the forecasting accuracy of volatility models.Finally,the high-frequency data usually contains more intraday trading information and is more conducive to predicting the volatility of the crude oil market.When consider the factor of structural break,the forecasting performance of GARCH-type models are normally worse than the HAR-type models,even if the lagged realized volatility is included as an additional explanatory variable.Besides,the FFF-HAR-type models considering smooth shift often obtain the best prediction results.In this paper,we discuss the volatility forecasting of crude oil market based on structural break,and the research significance mainly includes the following two aspects.On the one hand,it helps many market participants and investors to deeply understand and grasp the law of crude oil market price volatility.On the other hand,the conclusions of this paper provide some support and reference for the decision-makers and market participants.