Semiparametric adaptive estimation of the GARCH model using Matlab
We report Matlab code for semiparametric (adaptive) estimation of the GARCH model; moreover, we report a Monte Carlo simulation which shows that the semiparametric estimator is more efficient than the Quasi Maximum Likelihood estimator. The theory for semiparametric (adaptive) estimation of the GARCH model is reported in Drost and Klaassen (1997), Journal of Econometrics, Volume 81, pp. 193-221. For the sake of brevity, theoretical details on semiparametric estimation are not reported here, and we refer the reader to Drost and Klaassen (1997) for theoretical details. We use the t5-student innovation for the GARCH process.
OTHER DETAILS: All Matlab code files must be included in the same folder, and the folder must be added to Matlab path. The main Matlab file which includes the Monte Carlo simulation is entitled "MainFile.m". All other Matlab files included in this repository -i.e., "MLE_normal_new.m", "MLE_t5_NEW.m" and "mycon.m"- are ancillary files that are used to estimate the parameters of the GARCH model. Finally, we report all Matlab files also in the zip folder "Zip_Folder_Matlab_codes.zip".