On the relationship of persistence and number of breaks in volatility: new evidence for three CEE countries

In this article, we contribute to the discussion of volatility persistence in the presence of sudden changes. We follow previous research, particularly Wang and Moore (2009), who analysed stock market returns in five Central and Eastern European countries using the Iterated Cumulative Sum of Squares (ICSS) algorithm for detecting multiple breaks and the test (IT) proposed by Inclán and Tiao (1994). We complement this analysis by using the κ1 and κ2 statistic introduced by Sansó et al. (2004), which lead us to the hypothesis that the estimated persistence in volatility depends inversely on the number of breakpoints in volatility. We explored this claim through a simulation study, where by randomizing an increasing number of breakpoints over the sample, we estimated kernel density of the persistence measure. The results confirmed the relationship between persistence and the number of breakpoints. It also showed that the use of break detection algorithms leads to lower persistence estimates, even within the class of models with an equal number of breaks. Therefore, the overall decrease in persistence can be attributed both to the number of breaks and their position, as suggested by the chosen break detection tests.