Asymptotically UMP panel unit root tests: the effect of heterogeneity in the alternatives

Research output: Contribution to journalArticleScientificpeer-review

Abstract

In a Gaussian, heterogeneous, cross-sectionally independent panel with inciden- tal intercepts, Moon, Perron, and Phillips (2007) presents an asymptotic power envelope yielding an upper bound to the local asymptotic power of unit root tests. In case of homogeneous alternatives this envelope is known to be sharp, but this paper shows that it is not attainable for heterogeneous alternatives. Using limit experiment theory we derive a sharp power envelope. We also demonstrate that, among others, one of the likelihood ratio based tests in Moon, Perron, and Phillips (2007), a pooled GLS based-test using the Breitung and Meyer (1994) device, and a new test based on the asymptotic structure of the model, are all asymptotically UMP. Thus, perhaps somewhat surprisingly, pooled regression-based tests may yield optimal tests in case of heterogeneous alternatives. Although finite-sample powers are comparable, the new test is easy to implement and has superior size properties.
Original languageEnglish
Pages (from-to)539-559
Number of pages21
JournalEconometric Theory
Volume31
Issue number3
Early online date23 Sep 2014
DOIs
Publication statusPublished - 2015

Fingerprint

Panel unit root tests
regression
experiment
Power envelope
Optimal test
Finite sample
Upper bound
Unit root tests
Experiment
Likelihood ratio

Cite this

@article{c6e665ec03e14dd4937697790c926dc4,
title = "Asymptotically UMP panel unit root tests: the effect of heterogeneity in the alternatives",
abstract = "In a Gaussian, heterogeneous, cross-sectionally independent panel with inciden- tal intercepts, Moon, Perron, and Phillips (2007) presents an asymptotic power envelope yielding an upper bound to the local asymptotic power of unit root tests. In case of homogeneous alternatives this envelope is known to be sharp, but this paper shows that it is not attainable for heterogeneous alternatives. Using limit experiment theory we derive a sharp power envelope. We also demonstrate that, among others, one of the likelihood ratio based tests in Moon, Perron, and Phillips (2007), a pooled GLS based-test using the Breitung and Meyer (1994) device, and a new test based on the asymptotic structure of the model, are all asymptotically UMP. Thus, perhaps somewhat surprisingly, pooled regression-based tests may yield optimal tests in case of heterogeneous alternatives. Although finite-sample powers are comparable, the new test is easy to implement and has superior size properties.",
author = "I.G. Becheri and F.C. Drost and {van den Akker}, R.",
year = "2015",
doi = "10.1017/S0266466614000656",
language = "English",
volume = "31",
pages = "539--559",
journal = "Econometric Theory",
issn = "0266-4666",
publisher = "CAMBRIDGE UNIV PRESS",
number = "3",

}

Asymptotically UMP panel unit root tests : the effect of heterogeneity in the alternatives. / Becheri, I.G.; Drost, F.C.; van den Akker, R.

In: Econometric Theory, Vol. 31, No. 3, 2015, p. 539-559.

Research output: Contribution to journalArticleScientificpeer-review

TY - JOUR

T1 - Asymptotically UMP panel unit root tests

T2 - the effect of heterogeneity in the alternatives

AU - Becheri, I.G.

AU - Drost, F.C.

AU - van den Akker, R.

PY - 2015

Y1 - 2015

N2 - In a Gaussian, heterogeneous, cross-sectionally independent panel with inciden- tal intercepts, Moon, Perron, and Phillips (2007) presents an asymptotic power envelope yielding an upper bound to the local asymptotic power of unit root tests. In case of homogeneous alternatives this envelope is known to be sharp, but this paper shows that it is not attainable for heterogeneous alternatives. Using limit experiment theory we derive a sharp power envelope. We also demonstrate that, among others, one of the likelihood ratio based tests in Moon, Perron, and Phillips (2007), a pooled GLS based-test using the Breitung and Meyer (1994) device, and a new test based on the asymptotic structure of the model, are all asymptotically UMP. Thus, perhaps somewhat surprisingly, pooled regression-based tests may yield optimal tests in case of heterogeneous alternatives. Although finite-sample powers are comparable, the new test is easy to implement and has superior size properties.

AB - In a Gaussian, heterogeneous, cross-sectionally independent panel with inciden- tal intercepts, Moon, Perron, and Phillips (2007) presents an asymptotic power envelope yielding an upper bound to the local asymptotic power of unit root tests. In case of homogeneous alternatives this envelope is known to be sharp, but this paper shows that it is not attainable for heterogeneous alternatives. Using limit experiment theory we derive a sharp power envelope. We also demonstrate that, among others, one of the likelihood ratio based tests in Moon, Perron, and Phillips (2007), a pooled GLS based-test using the Breitung and Meyer (1994) device, and a new test based on the asymptotic structure of the model, are all asymptotically UMP. Thus, perhaps somewhat surprisingly, pooled regression-based tests may yield optimal tests in case of heterogeneous alternatives. Although finite-sample powers are comparable, the new test is easy to implement and has superior size properties.

U2 - 10.1017/S0266466614000656

DO - 10.1017/S0266466614000656

M3 - Article

VL - 31

SP - 539

EP - 559

JO - Econometric Theory

JF - Econometric Theory

SN - 0266-4666

IS - 3

ER -