Volatility Forecasting in Borsa Istanbul: A Comparative Analysis of Participation and Conventional Indices

Volatility is a key input for investment decisions and portfolio management. Volatility modeling and forecasting are essential for risk analysis and risk management in financial markets. In this study, we investigate volatility characteristics of the four indices in Borsa Istanbul (Istanbul Stock Exchange) using GARCH family models. The primary motivation of this research is to model and forecast the volatility of participation and conventional indices. First, we estimate the model parameters using 809 days of data and explain the volatility features of the index return series. Second, we generate 57 days of forecast series and calculate realized volatility in the out-of-sample period. Finally, we analyze the predictive power of the forecast series of the models in relation to realized volatility, calculated using intraday returns. We use measures such as RMSE (Root Mean Squared Error), QLIKE (Quasi-Likelihood Loss), and R-squared (R²) to evaluate the predictability of GARCH models. The results reveal that participation indices exhibit distinct volatility characteristics and that asymmetric models—especially EGARCH and TGARCH—produce satisfactory predictions.