Information Transmission Among Equity Markets: A Comparison Between ARDL and GARCH Model

This study compares the performance of autoregressive conditional heteroscedastic (ARCH) model and autoregressive distributed lag (ARDL) model in term of relationship detection. The daily, weekly, and monthly data are used from 2005 to 2019 to explore the dynamic linkages among KSE 100, S&P 500, Nasdaq 100, Dowjones 30, and DFMG indices. The results indicate that the ARDL and ARCH model have same power to detect the relationship among financial series. The results show that due high volatility in daily and weekly data the ARDL model is failed to capture ARCH effect. In case of monthly data, the performance of ARDL model is as good as GARCH model. It concluded that on monthly basis or less frequency data the ARDL model can be used as an alternative method to GARCH model for financial time series modeling.