IKTISAT ISLETME VE FINANS, vol.28, no.330, pp.81-104, 2013 (SSCI)
This study uses a value-at-risk method to investigate the GARCH(p,q), GJR-GARCH (p,g), and EGARCH(p,g) models for the ISE 100, Financial, Industrials, and Services indices to determine the most accurate GARCH-type model. The results show that, at the 99% confidence level, the most accurate are the GARCH(1,1) model for the ISE 100 and Financial indices and the GJR-GARCH(1,1,1) and GJR-GARCH(2,1,1) models for the Industrials and Services indices, respectively However, at the 95% confidence level, the most accurate models are the GJR-GARCH(1,1,1) model for the ISE 100 and Financial indices and the EGARCH(1,1,2) model for the Services indices. Furthermore, it is found that the GARCH and GJR-GARCH models perform better than the EGARCH model in most cases.