Analysis of Realized Volatility in Tehran Stock Exchange using Heterogeneous Autoregressive Models Approach

The present study aims at investigating the behavior of realized volatility for high-frequency data of Tehran Stock Index from April 28th, 2012 to August 8th, 2018. Three different types of HAR models including of HAR-RV-CJ, HAR-RV and HAR-RVJ were used to analyze the Realized Volatility. The obtained results of three diverse models revealed that the estimated Realized Volatility in market was described appropriately by the traders who work daily and in the framework of HAR-RVJ model. Moreover, based on the Heterogeneous Market Hypothesis, we found out that in comparative performance for all of time horizons in this study, the results of four evaluative criteria (including of MSE, RMSE and etc.) in HAR-RVJ model is lower than HAR-RV-CJ
and
HAR-RV. The in-sample forecasting performance of HAR-RVJ, in relation to Future Volatility of Tehran Stock Exchange Index, was better than the results we obtained from the alternative models in the study (HAR-RV
and
HAR-RV-CJ) and the best scores were observed among all the criteria. In addition, for the out-of-sample analysis, the simple HAR-RV model had superiority over the other two models only in the Monthly time horizon.