Predictive Failure and Econometric Modelling in Macroeconomics: The Transactions Demand for Money

Abstract

Abstract This tests the DHSY approach in another subject‐matter area (money demand) to investigate its ability to produce improved models and to encompass previous findings. ‘Simple‐to‐general’ modelling methods are criticized and contrasted with the outcome achieved by general‐to‐simple, sequentially reducing to parsimonious, constant, and encompassing equations. Contemporaneous conditioning is based on weak exogeneity and constancy is stressed as a model‐selection criterion. A successful test of the selected error‐correction transactions’ demand for money model on new data is reproduced from The Econometric Analysis of Economic Time Series.