This study examines the performance of the GARCH and stochastic volatility (SV) models incorporating jumps in terms of returns. Therefore, in addition to standard GARCH and SV models, it also takes into account the GARCH with jumps and SV with jumps models for the Turkish stock market including the ISE100, ISE Financial and ISE Industrial stock indices. All the models are estimated using Bayesian techniques and, as a Bayes factor, log ML is applied to compare the performances of the models. The results show that (i) the jump probability is 0.06, equating to nearly three jumps per year in the Turkish stock market; (ii ) incorporating jumps into the models improves their performance; and (iii) the best model is the SV model with jumps, while the standard SV model and SV with jumps model outperform the standard GARCH and GARCH with jumps models.
Keywords: Volatility, Jumps, GARCH model, Stochastic volatility model, Turkish stock market
|