Comparative Study of Modelling and Forecasting Volatility: The Case of Egypt, and Japan

 The  purpose  of  this  paper  is  to  evaluate  the  forecasting  performance  of  linear  and non-linear  (GARCH)  models  in  terms  of  their  in-sample  and  out-of-sample  forecasting accuracy  for  EGX30  and  Nikkei225  indices  as  an  example  of  an  emerging  and  developed markets respectively. We  employ  GARCH,  GARCH-IN-MEAN,  EGARCH,  GJR-GARCH, Multivariate GARCH,  and  Nelson's  EGARCH  for  forecasting  using  daily  price  data  of  the  indices  for the  period  of  2001  to  2019.We  find  that  the  volatility  shocks  on  the  indices  returns  are quite  persistent.  Furthermore,  our  findings  show  that  the  indices  have  leverage  effect,  and the  impact  of  shocks  is  asymmetric,  and  consequently  it  can  be  stated  that  the  impact  of negative shocks on volatility are higher than positive shocks.The results suggest that the Nelson's EGARCH model is the most accurate model in the   GARCH   class   for   forecasting,   as   this   model   outperforms   the   other   models. Additionally,  we  find  that  emerging  stock  markets  have  higher  volatilities  than  those  in developed  markets.  Further,  these  results  imply  that  the  EGARCH  model  might  be  more useful  than  other  models  when  implementing  risk  management  strategies  and  developing stock pricing model. This paper contributes to the literature by comparing two significant global markets; one  of  the  largest  developed  economies  in  the  world,  Japan,  and  one  of  Africa’s  largest developing economies, Egypt.