A SIMPLE RECURSIVE ESTIMATION METHOD FOR LINEAR-REGRESSION MODELS WITH AR(P) DISTURBANCES Academic Article uri icon

abstract

  • An exact transformation that reduces the AR(p) process into white noise is well known in statistics, see Fuller (1976). However, practitioners still use and econometric textbooks still recommend the Cochrane-Orcutt procedure for p>2, see Greene (1990). This paper derives an alternative exact transformation for the AR(p) process which is computationally simple. Based on this transformation, a GLS estimator is proposed, requiring only least squares regressions and recursive computations. This is illustrated for the AR(3) case. 1994 Springer-Verlag.

published proceedings

  • STATISTICAL PAPERS

author list (cited authors)

  • BALTAGI, B. H., & LI, Q.

citation count

  • 3

complete list of authors

  • BALTAGI, BH||LI, Q

publication date

  • December 1994