Volatility modeling in Borsa Istanbul stock market: An application on BIST banking index

In this study, the volatility model of BIST Banking index is estimated based on the daily closing data of the index between January 4, 2010 and December 31, 2019. In this context, firstly, the stationary of the price series was investigated with the help of Augmented Dickey-Fuller unit root test and it was determined that it was a first-degreed stationary series. Then, it was determined that the best mean equation model is ARMA (2,2) among the autoregressive models. In addition, it was observed that there was an ARCH effect in the error terms of the mean equation, and it was tested that with which conditional heteroscedasticity model or models, the series of BIST Banking index can be explained. According to the test results obtained, it was determined that the model that gives the best results in estimating the volatility modeling of the BIST Banking series is TGARCH (0,1,1) compared to the information criteria and EGARCH (1,1,1) compared to the forecasting performance. Keywords: Volatility, BIST Banking Index, Heteroscedasticity, ARCH, GARCH