Abstract
I propose a likelihood ratio test for fixed unit root against time switching unit root models. Our test is different from the existing jump detection literature centered on the application of various forms of Augmented Dickey-Fuller tests. Our methodology involves the inclusion of random coefficients, which effectively capture both expansion and contraction behaviors. We show that the contiguous alternatives converge to the null hypothesis at the order of $T^{-3/4}$, where $T$ is the sample size. Our test is asymptotically optimal in the sense that it maximizes a weighted power function. We derive the asymptotic distribution of our test under the null and local alternatives.
Committee Chair
Werner Ploberger
Committee Members
Gaetano Antinolfi, Ismael Mourifie, Fei Tan, Marting Xiumin, Guofu Zhou
Degree
Doctor of Philosophy (PhD)
Author's Department
Economics
Document Type
Dissertation
Date of Award
4-3-2024
Language
English (en)
DOI
https://doi.org/10.7936/5br7-7s24
Author's ORCID
0000-0002-5860-5032
Recommended Citation
Zhang, Li, "Likelihood Ratio Test for Markov Switching Parameters" (2024). Arts & Sciences Theses and Dissertations. 3079.
The definitive version is available at https://doi.org/10.7936/5br7-7s24