Friday, October 7, 2022

Lars P. Syll —On the validity of econometric inferences


This is essentially a recap of Keynes's criticism of Tinbergen's approach to econometrics. The basis of the issue is the assumption that the social sciences are sufficiently like the natural sciences to enable use of similar methodology. 

Keynes argued that this assumption is wrong. The subject matter of economics is historical and institutional, that is, variable, and in social systems, the data lacks the homogeneity of the natural sciences. See the blog post by Lars, Econometrics: The Keynes-Tinbergen controversy (8 Nov 2018).

We now know much more about systems than was known in the 1930 and 40's when this controversy began that was sparked by Keynes. The problem that the social sciences, including economics, face is modeling complex adaptive systems, which societies are. This means that not only are there known unknown owing to complexity but also unknown and likely unknowns that are not knowable in advance owing to emergence as a result of adaptation through learning from feedback.

This is a reason that heterodox approaches to economics reject the conventional approach based on formalism. For example, MMT is based on operational analysis of institutional arrangements, specifically the existing monetary system as a general case and the various national systems as specific cases.

Lars P. Syll’s Blog
On the validity of econometric inferences
Lars P. Syll | Professor, Malmo University

No comments: