GARCH Model And Empirical Analysis Of China's Stock Market Of The Agricultural Section

When we use mathematical tools, especially the statistical tools to study thefinancial problems, we should first solve the problem of volatility modeling. Discretemodel to deal with the problem of volatility is the GARCH model. GARCH modelresiduals distribution is often assumed for the standard Gaussian distribution can beGaussian distribution does not fully reflect the financial data of the thick tails, we firstused the Student t-distribution and generalized error distribution instead of thestandard Gaussian distribution to the establishment of the GARCH model and theShanghai Composite Index analysis, we found that the EGARCH model fits best thedistribution of generalized error distribution. GED distribution of the tail is difficult tofully express the fat-tail data, the Generalized Pareto distribution of the thickness ofthe end of interval analysis, specifically consider the two-parameter generalizedPareto distribution as a mean correction of the sequence, the establishment of theARCH-the GPD model, and compared with the normal and student's t distributionARCH model. In the empirical analysis conducted by Standard&Poor's500Index,the results show that the ARCH-the GPD model than the t distribution ARCH modelto better fit the data.Test the Chinese stock market of the agricultural sector, especially the fat tailcharacteristics of the leading agricultural section, the application of the GED-GARCHmodel fitting, the results show that the GED-GARCH model fitting accuracy is higherthan the normal distribution of the GARCH model. Finally TARCH model andGED-EGARCH (1,1)-M-model analysis of data, found that agriculture leading platehas a significant leverage effect, while between income and the fluctuation rate alsoexists significant relationship, the last dummy variable found Chinese stock market agriculture leading plate has significantly with the week effect, the positive Tuesdayeffect and negative Friday effect...