The controversy over the selection of ‘growth regressions’ was precipitated by some remarkably numerous ‘estimation’ strategies, including two million regressions by Sala‐i‐Martin [ American Economic Review (1997b) Vol. 87, pp. 178–183]. Only one regression is really needed, namely the general unrestricted model, appropriately reduced to a parsimonious encompassing, congruent representation. We corroborate the findings of Hoover and Perez [ Oxford Bulletin of Economics and Statistics (2004) Vol. 66], who also adopt an automatic general‐to‐simple approach, despite the complications of data imputation. Such an outcome was also achieved in just one run of PcGets , within a few minutes of receiving the data set in Fernández, Ley and Steel [ Journal of Applied Econometrics (2001) Vol. 16, pp. 563–576] from Professor Ley.
No takes yet. Share an insight, caveat, or question.
Hendry et al. (2004) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: