Lars Syll — The nodal point of the macroeconomics debate

Microfoundations today means more than anything else that you try to build macroeconomic models assuming “rational expectations” and hyperrational “representative actors” optimizing over time. Both are highly questionable assumptions.

Read it at Lars P. Syll's Blog
The nodal point of the macroeconomics debate
by Lars P. Syll | Professor, Malmo University

To deliver macroeconomic models building on rational expectations microfoundations the economists have to constrain expectations on the individual and the aggregate level to be the same. If revisions of expectations take place they typically have to take in a known and prespecified precise way. This squares badly with what we know to be true in real world, where fully specified trajectories of future expectations revisions are no-existent 

Further, most macroeconomic models building on rational expectations microfoundations are time-invariant and so give no room for any changes in expectations and their revisions. The only imperfection of knowledge they admit of is included in the error terms, error terms that are assumed to be additive and to have a give and known frequency distribution, so that the models can still fully pre-specify the future even when incorporating these stochastic variables into the models.

This is also the basis of George Soros's general theory of reflexivity. IN The Alchemy of Finance, for example, Soros views economics and finance as a complex information system in which feedback is a principal determinant. 

Feedback is a reflexive information flow alters dependent information flows in ways that cannot be anticipated any more than actual events can be known advance. There are not only known unknowns but also unknown unknowns affecting complex information systems, and social sciences have to take this into account in their theories and modeling. To the degree they don't they will produce insufficient explanations and this will be revealed by failed predictions. 

This is basically an epistemological problem with implications for probability and statistics, as Lars points out. The present approach to microfoundations glosses over the epistemological issues since it is essentially static with respect to information in presuming both a representational agent and rationality in the way it does. Understandable methodological choice in light of model tractability, but inadequate to model complex information systems.

For an investigation of the epistemology see How George Soros Knows What He Knows: Towards a General Theory of Reflexivity by Flavia Cymbalista, Ph.D.

Reflexivity In Social Systems: The Theories Of George Soros by Stuart Umpleby, George Washington University

Mathematical analysis of Soros’s theory of reflexivity by C.P. Kwong