Testing for Flexible Nonlinear Trends with an Integrated or Stationary Noise Component

Pierre Perron, Mototsugu Shintani, Tomoyoshi Yabu

Research output: Contribution to journalArticle

Abstract

This paper proposes a new test for the presence of a nonlinear deterministic trend approximated by a Fourier expansion in a univariate time series for which there is no prior knowledge as to whether the noise component is stationary or contains an autoregressive unit root. Our approach builds on the work of Perron and Yabu () and is based on a Feasible Generalized Least Squares procedure that uses a super-efficient estimator of the sum of the autoregressive coefficients α when α = 1. The resulting Wald test statistic asymptotically follows a chi-square distribution in both the I(0) and I(1) cases. To improve the finite sample properties of the test, we use a bias-corrected version of the OLS estimator of α proposed by Roy and Fuller (). We show that our procedure is substantially more powerful than currently available alternatives. We illustrate the usefulness of our method via an application to modelling the trend of global and hemispheric temperatures.

Original languageEnglish
JournalOxford Bulletin of Economics and Statistics
DOIs
Publication statusAccepted/In press - 2017

Fingerprint

Deterministic Trend
Wald Test
Chi-square Distribution
Testing
Generalized Least Squares
Efficient Estimator
Fourier Expansion
Unit Root
trend
Prior Knowledge
Test Statistic
Univariate
Time series
Estimator
time series
Alternatives
Coefficient
statistics
Modeling
knowledge

ASJC Scopus subject areas

  • Statistics and Probability
  • Social Sciences (miscellaneous)
  • Economics and Econometrics
  • Statistics, Probability and Uncertainty

Cite this

Testing for Flexible Nonlinear Trends with an Integrated or Stationary Noise Component. / Perron, Pierre; Shintani, Mototsugu; Yabu, Tomoyoshi.

In: Oxford Bulletin of Economics and Statistics, 2017.

Research output: Contribution to journalArticle

@article{516f688d16244db0a1640fa247cbdb7d,
title = "Testing for Flexible Nonlinear Trends with an Integrated or Stationary Noise Component",
abstract = "This paper proposes a new test for the presence of a nonlinear deterministic trend approximated by a Fourier expansion in a univariate time series for which there is no prior knowledge as to whether the noise component is stationary or contains an autoregressive unit root. Our approach builds on the work of Perron and Yabu () and is based on a Feasible Generalized Least Squares procedure that uses a super-efficient estimator of the sum of the autoregressive coefficients α when α = 1. The resulting Wald test statistic asymptotically follows a chi-square distribution in both the I(0) and I(1) cases. To improve the finite sample properties of the test, we use a bias-corrected version of the OLS estimator of α proposed by Roy and Fuller (). We show that our procedure is substantially more powerful than currently available alternatives. We illustrate the usefulness of our method via an application to modelling the trend of global and hemispheric temperatures.",
author = "Pierre Perron and Mototsugu Shintani and Tomoyoshi Yabu",
year = "2017",
doi = "10.1111/obes.12169",
language = "English",
journal = "Oxford Bulletin of Economics and Statistics",
issn = "0305-9049",
publisher = "Wiley-Blackwell",

}

TY - JOUR

T1 - Testing for Flexible Nonlinear Trends with an Integrated or Stationary Noise Component

AU - Perron, Pierre

AU - Shintani, Mototsugu

AU - Yabu, Tomoyoshi

PY - 2017

Y1 - 2017

N2 - This paper proposes a new test for the presence of a nonlinear deterministic trend approximated by a Fourier expansion in a univariate time series for which there is no prior knowledge as to whether the noise component is stationary or contains an autoregressive unit root. Our approach builds on the work of Perron and Yabu () and is based on a Feasible Generalized Least Squares procedure that uses a super-efficient estimator of the sum of the autoregressive coefficients α when α = 1. The resulting Wald test statistic asymptotically follows a chi-square distribution in both the I(0) and I(1) cases. To improve the finite sample properties of the test, we use a bias-corrected version of the OLS estimator of α proposed by Roy and Fuller (). We show that our procedure is substantially more powerful than currently available alternatives. We illustrate the usefulness of our method via an application to modelling the trend of global and hemispheric temperatures.

AB - This paper proposes a new test for the presence of a nonlinear deterministic trend approximated by a Fourier expansion in a univariate time series for which there is no prior knowledge as to whether the noise component is stationary or contains an autoregressive unit root. Our approach builds on the work of Perron and Yabu () and is based on a Feasible Generalized Least Squares procedure that uses a super-efficient estimator of the sum of the autoregressive coefficients α when α = 1. The resulting Wald test statistic asymptotically follows a chi-square distribution in both the I(0) and I(1) cases. To improve the finite sample properties of the test, we use a bias-corrected version of the OLS estimator of α proposed by Roy and Fuller (). We show that our procedure is substantially more powerful than currently available alternatives. We illustrate the usefulness of our method via an application to modelling the trend of global and hemispheric temperatures.

UR - http://www.scopus.com/inward/record.url?scp=85017400848&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85017400848&partnerID=8YFLogxK

U2 - 10.1111/obes.12169

DO - 10.1111/obes.12169

M3 - Article

JO - Oxford Bulletin of Economics and Statistics

JF - Oxford Bulletin of Economics and Statistics

SN - 0305-9049

ER -