Limits of Econometrics for Testing Economic Hypotheses (A Rejoinder on 'Limits of Econometrics' by David Freedman and Arnold Zellner’s Comments)

David A. Freedman's (DF hereafter) fascinating methodological survey article explains very concisely the principal challenges faced by economists and other social scientists when they aim to test hypotheses using statistical (or econometric) models. He also discusses the principal responses to these challenges. Let me state from the beginning that I agree with most of DF's central methodological assessments, in particular his well-reasoned “considerable scepticism about disentangling causal processes by statistical modeling.” Arnold Zellner's (AZ hereafter) important observations on DF's survey follow directly from the principles of his well-known philosophy of econometrics. I will provide additional commentary on David Freedman's survey and views, by discussing the various proposals how (not) to use econometric methods to test (hypothesised) causal structures in economics. In this context, I will also consider several of AZ's very interesting and helpful methodological recommendations