The concept of " risk " is the most crucial concept to be considered while making a financial investment decision. Determining policies within the scope of risk management is the concept of "risk". Predicting different risk levels that may be encountered in the future with an appropriate method is of great importance in preparing for these risks and making the right decisions. Making accurate forecasts is only possible by determining the models with the highest statistical performance. In the study, the return-based "ARCH (1) Model" selected among symmetric and asymmetric models and range-based "Conditional Autoregressive Range (CARR) Models" have been estimated based on the weekly data of the BIST100 index for the period 3 January 2011 - 24 July 2020 to obtain volatility estimates of the index return and to determine the model with the highest statistical performance. As a result of the model comparisons, the most appropriate model to be used in estimating the BIST100 return volatility is the "WKODGX (1,1) Model" with leverage effect and Weibull distribution in error terms based on the "interquartile range" measure, which is resistant to extreme values. When all the findings are evaluated together, it is concluded that the range-based models significantly improve the statistical performance in BIST100 index return volatility modelling.
Benzer Makaleler | Yazar | # |
---|
Makale | Yazar | # |
---|