Journal article

Testing for unit roots in seasonally adjusted data

N Olekalns

Economics Letters | Published : 1994

Abstract

The power of the Dickey-Fuller test is known to be low if seasonal data are adjusted via a moving average filter. I extend this result by considering alternatives to moving average filtering; a preliminary regression on dummy variables, and a frequency domain, band-pass filter. © 1994.

University of Melbourne Researchers

Citation metrics