The (De)merits of using Integral Transforms in Predicting Structural Break Points

Abstract

The structural break points of returns and volatility are generally illustrated by using uni-and-multivariate time series models. Despite the elegance of uni-and-multivariate models, the interchangeability of different structural break points is not well accounted for in those models. This study uses integral transforms (Fourier and Laplace) to illustrate the interchangeability of structural break points of indices. Furthermore, structural break points are validated with commonly used unit root structural break tests [(i) augmented Dickey Fuller (ADF), (ii) ADF-generalized least squares (GLS), Phillips Perron (PP) 1988 and Zivot-Andrews (ZA) 1992 tests]. The results illustrate persistent interchangeability and interconnectedness patterns of structural break points throughout the time series. Moreover, the structural break points tests confirm the findings of the integral transforms.