Testing for autoregressive against moving average errors in the linear regression model

Research output: Contribution to journalArticleResearchpeer-review

17 Citations (Scopus)


This paper presents a new approach to hypotheses testing problems which are non-nested in the classical sense and which concern the covariance matrix of the disturbance vector of the linear regression model. In particular, the application of the approach to testing for AR(1) disturbances against MA(1) disturbances is explored in some detail. Practical difficulties are discussed and selected upper bounds for the test's five percent significance points are tabulated. The small sample power of four versions of the new test are compared empirically and a clear conclusion is made in regard to the best overall test.

Original languageEnglish
Pages (from-to)35-51
Number of pages17
JournalJournal of Econometrics
Issue number1
Publication statusPublished - 1983

Cite this