We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
UNBIASEDNESS OF PREDICTIONS FROM ESTIMATED AUTOREGRESSIONS WHEN THE TRUE ORDER IS UNKNOWN.
- Authors
Dufour, Jean-Marie
- Abstract
The article considers the case of estimating the parameters of an autoregressive model by ordinary least squares. It describes the so-called jointly symmetric processes, a wide class of stochastic processes enjoying a simple symmetry property. Then, it demonstrates that any autoregressive model estimated by ordinary least squares yields unbiased forecasts when applied to a jointly symmetric process. The article suggests that for a symmetric autoregressive process, one will get unbiased forecasts even if the order of the estimated model is lower than the actual one.
- Subjects
AUTOREGRESSION (Statistics); FORECASTING; LEAST squares; ESTIMATION theory; MATHEMATICAL statistics; PROBABILITY theory; STATISTICS; ECONOMETRICS; MATHEMATICAL economics; MATHEMATICAL models; ECONOMETRIC models; ECONOMIC statistics
- Publication
Econometrica, 1984, Vol 52, Issue 1, p209
- ISSN
0012-9682
- Publication type
Article
- DOI
10.2307/1911469