Lessons in Econometric Methodology: The Axiom of Correct Specification

Abstract

According to Econometric theory, one should start with the right model, and then estimate it. The process of model selection causes many types of problems in estimation and inference. In practice, however, everyone estimates different models eventually selecting one as the “right” one. All of the millions of regressions possible to explain economic growth (or any other suitable variable) cannot all be right; yet, researchers estimate and interpret their own favorite equations as the only right one. A huge amount of work has been carried out on the appropriate methodology for selection of regressors. Although it is known to experts, it is not readily available in textbooks, and hence not commonly used in applied econometrics. The goal of this article is to provide a review of the relevant literature, to make it accessible to practicing econometricians and students.

Keywords: Econometric Methodology, Model Selection, Spurious Regression

JEL Classification: C10, C52