Semiparametric EGARCH-M Model And Its Application Of China Stock Market

Volatility measures the uncertainty of asset returns and has such an important role in the stock market.If the volatility of financial asset returns analyzed,the risk premium is estimated,Investors can make judgments accordingly,the effective allocation of their own portfolio.However,financial assets tends to have the characteristics of spikes,persistence and aggregation.The classical analysis and processing will assume that the volatility does not change with time and can not describe these characteristics of the yield.So there have been GARCH,GARCH-M,EGARCH and other models,we call it the parameter GARCH model.In the analysis of financial assets,the volatility risk premium is estimated,commonly used is the GARCH-M model.However,the parameter GARCH model needs to give a specific model form and give assumptions about the model error.If these assumptions are not established,then the statistics inferred with a greater error,or even no practical significance.In order to compensate for these shortcomings,non-parametric models are used,but it is difficult to make a reasonable explanation of the estimation results.In this paper,we combine parametric model with the nonparametric and propose a semiparametric EGARCH-M model.Based on the idea of local polynomial and two-stage estimate,with their asymptotic normality and convergence proved.In this model,the parametric part can explain the result,while the nonparametric part could reduce the bias.We use the proposed semi-parametric EGARCH-M model to study the volatility of this set of sequences.We use the semi-parametric EGARCH-M model to study the volatility of this group.Under the criterion of MSE and QLIKE,the semiparametric is superior to the EGARCH-M,moreover,the risk premium curve depicted by the semiparametric model has a reasonable explanation in practice.