Tests for seasonal differencing with an unknown break-point

T.M. NG, Wai Keung LI

Research output: Contribution to journalArticle

Abstract

Some Lagrange multiplier tests for seasonal differencing are proposed; their main objective is to avoid over-differencing due to structural change. The null hypothesis is either the presence of both regular and seasonal unit roots or the presence of a seasonal unit root. Alternative hypotheses allow for stationarity around a possible structural change where the break-point is unknown. The location of the structural change is estimated using the proposed procedures, the asymptotic distribution of the test statistics under the null hypothesis is derived and some useful percentiles are tabulated. An illustrative example based on the Canadian Consumer Price Index is presented. Copyright © 1996 the Statistical Society of Australia.
Original languageEnglish
Pages (from-to)131-153
JournalAustralian Journal of Statistics
Volume38
Issue number2
DOIs
Publication statusPublished - Jun 1996

    Fingerprint

Citation

Ng, T. M., & Li, W. K. (1996). Tests for seasonal differencing with an unknown break-point. Australian Journal of Statistics, 38(2), 131-153. doi: 10.1111/j.1467-842X.1996.tb00670.x

Keywords

  • Break-point
  • Broken trend
  • Lagrange multiplier test
  • Regular and seasonal unit roots
  • Structural change
  • Wiener process