A locally optimal seasonal unit-root test
Journal of Business and Economic Statistics
Taylor & Francis Inc.
349 - 356
Item Usage Stats
MetadataShow full item record
This article proposes a locally best invariant test of the null hypothesis of seasonal stationarity against the alternative of seasonal unit roots at all or individual seasonal frequencies. An asymptotic distribution theory is derived and the finite-sample properties of the test are examined in a Monte Carlo simulation. My test is also compared with the Canova and Hansen test. The proposed test is superior to the Canova and Hansen test in terms of both size and power.
KeywordsLocally best invariant test
Maximum likelihood estimation
Seasonal differencing filter