Abstract:
Variable addition tests for parameter stability are based on the use of test variables constructed as the product of dummy variables and the variables whose coefficient is believed to have changed. The object of this note is to analyze some power properties of this class of tests. The main conclusion is that the power of such tests against the null of no change depends on the length of the test period. It is suggested that, in absence of a priori information on the timing of breaks, a suitable strategy can be based on recursive testing. Copyright 1994 by Blackwell Publishing Ltd
Date: 1994
There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.
Related works: This item may be available elsewhere in EconPapers: Search for items with the same title.
Oxford Bulletin of Economics and Statistics is edited by Christopher Adam, Anindya Banerjee, Christopher Bowdler, Gavin Cameron, David Hendry, Adriaan Kalwij, John Knight and Jonathan Temple