Known Moving-Average Transformations and Autoregressive Processes
The errors in the linear models which are used so widely by economists may be generated by mixed moving-average autoregressive processes. If errors on an aggregative equation are generated by a mixed moving-average autoregressive process and the weights of the moving-average component of this process are known, then the least-squares procedure can yield consistent estimators of both signal and autoregressive parameters if two adjustments are made to the equation. The autoregressive transformation is combined with pre-multiplication by a Moore-Penrose inverse based on the known weights of the moving-average component.
Year of publication: |
1972-02
|
---|---|
Authors: | Rowley, J.C.R. ; Wilton, D.A. |
Institutions: | Economics Department, Queen's University |
Saved in:
freely available
Saved in favorites
Similar items by person
-
Autocorrelation in Empirical Studies of Wage Determination
Rowley, J.C.R., (1972)
-
Empirical Foundations for the Canadian Phillips Curve
Rowley, J.C.R., (1973)
-
Temporal Spillover and Autocorrelation in Some Aggregate Models of Wage-Determination
Rowley, J.C.R., (1972)
- More ...