The econometric estimation and testing of DARP models

Abstract. DARP, acronym for Drift Analysis of Regression Parameters, originated as a heuristic technique for the investigation of parametric drift in any arbitrary `expansion space', geographic or otherwise. DARP was intended as an exploratory tool useful to aid with the formal specification of parametric drift. In this paper, the DARP technique is reformulated in terms of `DARP models', and the estimation and testing of these models by GLS, FGLS, and ML are discussed. The ML estimation of a spatial DARP model is demonstrated using empirical data.