Publication: Fractional integration versus trend stationary in time series analysis
Loading...
Identifiers
Publication date
1997-01
Defense date
Authors
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The objective of this paper is to study the effects of spurious detrending of a nonstationary fractionally integrated process (NFI(d), d ~5) on the performance of the traditional least squares estimators and tests. We extend previous work on the subject undertaken by Durlauf and Phillips (1988) which considered only the leading difference stationary (d = 1) case. Moreover, we also consider the possibility of a double misspecification both in the stochastic and in the nonstochastic trends. Standard t-Student tests are shown to diverge in distribution invalidating any inference concerning the presence of time trends. On the other hand, we prove that, even under this double misspecification, the Durbin-Watson statistic remains to be a useful misspecification test,
Description
Keywords
Spurious detrending, Nonstationary fractionally integrated processes, Specification tests, Misspecification