Abstract
The essential roles of methodology in the development of sound econometric models and of computers in making econometrics operational are noted. The former has led to the theory of reduction as a basis for general‐to‐specific (Gets) modelling procedures to build congruent empirical models. Software for data management, graphics, estimation, testing, and simulation have underpinned many of the great strides in the discipline. A new generation of computer programs (such as PcGets) bids fair both to resolve many of the methodological issues about model selection that have been the subject of debate—by demonstrating that general‐to‐specific works. PcGets controls the significance level of the selection process, yet attains a power close to the maximum feasible. Automatic modelling both facilitates simulation studies to evaluate its properties and greatly reduces the burdens of empirical modelling.