摘要：The purpose of the present study is to empirically examine the performance of different kinds of volatility modeling and their forecasting performance for the general index of an emerging stock market, namely Dhaka Stock Exchange from the period December 06, 2010 to March 12, 2013. We mainly used Box-Jenkins modeling strategy thereafter the volatility model. The descriptive statistics, correlogram, unit root test, ARMA, ARCH, GARCH, TARCH, EGARCH and several model selection criteria are used in this study. The Butterworth filter is used for removing the noise of the return series of general index. All the parameters in this study are estimated through Maximum Likelihood method. The descriptive statistics show general index decrease slightly overtime with positively skewed and leptokurtic. The return series follows ARMA(1,1) model with volatility provide evidence of the superiority of GARCH(1,1) and GARCH(2,1) over the all order of other GARCH models. Finally, we found that the fitted model on filtered general index of Dhaka Stock Exchange are ARMA(1,1) with GARCH(1, 1) and GARCH(2,1) model. This model can be used for future policy implication through its accurate forecast.
关键词：Butterworth Filter; Return Series; Unit Root; ARMA; GARCH; TARCH; EGARCH; Model Selection Criteria
