Testing for ARMA(1,1) Disturbances in the Linear Regression Model

Serious alternatives to the AR(1) disturbance model in econometric applications of linear regression include MA(1) disturbances and the sum of independent white noise and AR(1) disturbance components. All three are special cases of ARMA(1,1) processes. This paper reports an empirical power comparison of tests for AR(1), MA(1) and ARMA(1,1) disturbances assuming ARMA(1,1) disturbances. Tests compared include the Durbin-Watson test, the locally best invariant test and various point optimal invariant (POI) tests. The results suggest that the power of the POI test is largely invariant to the choice of AR(1) parameter value at which power is maximized. This conclusion is strengthened by a theoretical result.


Issue Date:
Jan 01 1992
Publication Type:
Working or Discussion Paper
Language:
English
Total Pages:
25
Series Statement:
Working Paper No. 1/92




 Record created 2018-01-30, last modified 2018-01-31

Fulltext:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)