Journal article
Testing the autoregressive parameter with the t statistic
Journal of econometrics, Vol.27(2), pp.143-161
02/01/1985
DOI: 10.1016/0304-4076(85)90084-3
Abstract
This paper considers a Gaussian first-order autoregressive process with unknown intercept where the initial value of the variable is a known constant. Monte Carlo simulations are used to investigate the sampling distribution of the
t statistic for the autoregressive parameter when its value is in the neighborhood of unity. A small sigma asymptotic result is exploited in the construction of exact non-similar tests. The powers of non-similar tests of the random walk and other hypotheses are estimated for sample sizes typical in economic applications.
Details
- Title: Subtitle
- Testing the autoregressive parameter with the t statistic
- Creators
- J.C. Nankervis - City of London Polytechnic, London EC2M 6SQ, UKN.E. Savin - Trinity College London
- Resource Type
- Journal article
- Publication Details
- Journal of econometrics, Vol.27(2), pp.143-161
- DOI
- 10.1016/0304-4076(85)90084-3
- ISSN
- 0304-4076
- eISSN
- 1872-6895
- Publisher
- Elsevier B.V
- Number of pages
- 19
- Language
- English
- Date published
- 02/01/1985
- Academic Unit
- Economics
- Record Identifier
- 9984963102502771
Metrics
1 Record Views