A Comparison of Parametric, Semi-nonparametric, Adaptive, and Nonparametric Cointegration Tests

H. Peter Boswijk, Andre Lucas, Nick Taylor

Research output: Working paperProfessional

Abstract

This paper provides an extensive Monte-Carlo comparison of severalcontemporary cointegration tests. Apart from the familiar Gaussian basedtests of Johansen, we also consider tests based on non-Gaussianquasi-likelihoods. Moreover, we compare the performance of these parametrictests with tests that estimate the score function from the data using eitherkernel estimation or semi-nonparametric density approximations. Thecomparison is completed with a fully nonparametric cointegration test. Insmall samples, the overall performance of the semi-nonparametric approachappears best in terms of size and power. The main cost of thesemi-nonparametric approach is the increased computation time. In largesamples and for heavily skewed or multimodal distributions, the kernel basedadaptive method dominates. For near-Gaussian distributions, however, thesemi-nonparametric approach is preferable again.
Original languageEnglish
Place of PublicationAmsterdam
PublisherTinbergen Instituut
Publication statusPublished - 1999

Publication series

NameDiscussion paper TI
No.99-012/4

Fingerprint

Gaussian distribution
Costs

Cite this

Boswijk, H. P., Lucas, A., & Taylor, N. (1999). A Comparison of Parametric, Semi-nonparametric, Adaptive, and Nonparametric Cointegration Tests. (Discussion paper TI; No. 99-012/4). Amsterdam: Tinbergen Instituut.
Boswijk, H. Peter ; Lucas, Andre ; Taylor, Nick. / A Comparison of Parametric, Semi-nonparametric, Adaptive, and Nonparametric Cointegration Tests. Amsterdam : Tinbergen Instituut, 1999. (Discussion paper TI; 99-012/4).
@techreport{7c755ea32fc74d5cb8aafaee4c8ef904,
title = "A Comparison of Parametric, Semi-nonparametric, Adaptive, and Nonparametric Cointegration Tests",
abstract = "This paper provides an extensive Monte-Carlo comparison of severalcontemporary cointegration tests. Apart from the familiar Gaussian basedtests of Johansen, we also consider tests based on non-Gaussianquasi-likelihoods. Moreover, we compare the performance of these parametrictests with tests that estimate the score function from the data using eitherkernel estimation or semi-nonparametric density approximations. Thecomparison is completed with a fully nonparametric cointegration test. Insmall samples, the overall performance of the semi-nonparametric approachappears best in terms of size and power. The main cost of thesemi-nonparametric approach is the increased computation time. In largesamples and for heavily skewed or multimodal distributions, the kernel basedadaptive method dominates. For near-Gaussian distributions, however, thesemi-nonparametric approach is preferable again.",
author = "Boswijk, {H. Peter} and Andre Lucas and Nick Taylor",
year = "1999",
language = "English",
series = "Discussion paper TI",
publisher = "Tinbergen Instituut",
number = "99-012/4",
type = "WorkingPaper",
institution = "Tinbergen Instituut",

}

Boswijk, HP, Lucas, A & Taylor, N 1999 'A Comparison of Parametric, Semi-nonparametric, Adaptive, and Nonparametric Cointegration Tests' Discussion paper TI, no. 99-012/4, Tinbergen Instituut, Amsterdam.

A Comparison of Parametric, Semi-nonparametric, Adaptive, and Nonparametric Cointegration Tests. / Boswijk, H. Peter; Lucas, Andre; Taylor, Nick.

Amsterdam : Tinbergen Instituut, 1999. (Discussion paper TI; No. 99-012/4).

Research output: Working paperProfessional

TY - UNPB

T1 - A Comparison of Parametric, Semi-nonparametric, Adaptive, and Nonparametric Cointegration Tests

AU - Boswijk, H. Peter

AU - Lucas, Andre

AU - Taylor, Nick

PY - 1999

Y1 - 1999

N2 - This paper provides an extensive Monte-Carlo comparison of severalcontemporary cointegration tests. Apart from the familiar Gaussian basedtests of Johansen, we also consider tests based on non-Gaussianquasi-likelihoods. Moreover, we compare the performance of these parametrictests with tests that estimate the score function from the data using eitherkernel estimation or semi-nonparametric density approximations. Thecomparison is completed with a fully nonparametric cointegration test. Insmall samples, the overall performance of the semi-nonparametric approachappears best in terms of size and power. The main cost of thesemi-nonparametric approach is the increased computation time. In largesamples and for heavily skewed or multimodal distributions, the kernel basedadaptive method dominates. For near-Gaussian distributions, however, thesemi-nonparametric approach is preferable again.

AB - This paper provides an extensive Monte-Carlo comparison of severalcontemporary cointegration tests. Apart from the familiar Gaussian basedtests of Johansen, we also consider tests based on non-Gaussianquasi-likelihoods. Moreover, we compare the performance of these parametrictests with tests that estimate the score function from the data using eitherkernel estimation or semi-nonparametric density approximations. Thecomparison is completed with a fully nonparametric cointegration test. Insmall samples, the overall performance of the semi-nonparametric approachappears best in terms of size and power. The main cost of thesemi-nonparametric approach is the increased computation time. In largesamples and for heavily skewed or multimodal distributions, the kernel basedadaptive method dominates. For near-Gaussian distributions, however, thesemi-nonparametric approach is preferable again.

M3 - Working paper

T3 - Discussion paper TI

BT - A Comparison of Parametric, Semi-nonparametric, Adaptive, and Nonparametric Cointegration Tests

PB - Tinbergen Instituut

CY - Amsterdam

ER -

Boswijk HP, Lucas A, Taylor N. A Comparison of Parametric, Semi-nonparametric, Adaptive, and Nonparametric Cointegration Tests. Amsterdam: Tinbergen Instituut. 1999. (Discussion paper TI; 99-012/4).