Detecting Sudden Changes in the Extreme Value Volatility Estimator

Abstract

In this paper, we examine the performance of Sanso, Arago and Carrion's (2004) Iterated Cumulative Sum of Squares (AIT ICSS) algorithm to detect sudden changes in Rogers and Satchell (1991) estimator (RS estimator) and compare it with the performance of the demeaned squared returns by Monte Carlo simulation experiments. We assess the size and power properties of the AIT ICSS algorithm for both proxies of volatility for various data generating processes. We find that the AIT ICSS algorithm exhibits outstanding power properties when applied with the RS estimator. We apply the AIT ICSS algorithm on the major indices from the Advanced Emerging Economies suggested by FTSE and find that most of the structural breaks detected by the RS estimator can be related to major macroeconomic events while very few of the structural breaks detected by demeaned squared returns can be related to macroeconomic events and hence are probably spurious.