Kolmogorov-Smirnov type statistics for structural breaks - a new adjusted-range based self-normalization approach

Yongmiao Hong,Brendan McCabe,Jiajing Sun,Shouyang Wang
DOI: https://doi.org/10.2139/ssrn.3850894
2021-01-01
Abstract:In this paper, we introduce adjusted-range based Kolmogorov-Smirnov (KS) type statistics to test for structural breaks in the mean of a process and also in a more general setting. We propose a normalization based on the adjusted-range of a partial sum, which is stochastically proportional to the true long-run variance (LRV), and is nuisance parameter and distribution free. We find, through simulation studies, that the new statistic offers better performance than the self-normalized KS statistic (Shao and Zhang, 2010) and the KS statistics utilizing the small-b and fixed-b approaches (Kiefer and Vogelsang, 2005). Finally, empirical studies also illustrate the suitability of KS type tests.
What problem does this paper attempt to address?