Re-examination of volatility dynamics in Istanbul Stock Exchange

It is well known that ignoring the regime changes in standard GARCH models results in overestimation of volatility persistence. In this study, by applying iterated cumulative sums of squares (ICSS) algorithm on weekly data of ISE 30 and ISE 100 indices, sudden change points in variance are endogenously detected. For ISE 30 index, two sudden changes points associated with three different volatility regimes and for ISE 100 index, six break points and seven distinct volatility regimes are identified. This information is then integrated to a GARCH(1,1) model and it is found that the volatility persistence is not as high as it has been previously shown in the literature. Under the circumstances created by the recent global financial crisis, the results have important implications for financial investors who are interested in understanding the volatility pattern in stock markets. The findings also point to the review of common perception that the volatility in financial markets is highly persistent