Anyone following the economics blogosphere will have noticed the serious uptick in debates over “microfoundations,” and specifically, the turn towards microfoundations for macroeconomic models in the 1970s. All the macro bloggers – from Paul Krugman to Noahpinion – have chimed in on the question of whether or not economics made a serious error in striving for microfoundations (or, in how it went about striving for microfoundations). David Glasner’s post from earlier in the week was particularly fascinating, as he gave a sort of counterfactual history of what microfoundations in economics could have been (warning, jargon-heavy):

But my own view, in a nutshell, is that the rational-expectations revolution — especially the dogmatic view of how economics ought to be practiced espoused by Robert Lucas and his New Classical, Real Business Cycle and New Keynesian acolytes — has subverted the original aims of the microfoundations project. Rather than relax the informational assumptions underlying conventional equilibrium analysis to allow for a richer and more relevant analysis than is possible when using the tools of standard general-equilibrium theory, Lucas et al. developed sophisticated tools that enabled them to nominally relax the informational assumptions of equilibrium theory while using the tyrannical methodology of rational expectations combined with full market clearing to preserve the essential results of the general-equilibrium model. The combined effect of the faux axiomatic formalism and the narrow conception of microfoundations imposed by the editorial hierarchy of the premier economics journals has been to recreate the gap between the Keynesian theory of involuntary unemployment and rigorous microeconomic reasoning that Alchian, some forty years ago, thought he had found a way to bridge.

Basically, Glasner is saying that instead of trying to build a model which aggregated the limited, imperfect nature of individual human information gathering and processing (and thus explained observed phenomena like unemployment and recessions and all that), the entirety of mainstream macro (both the “freshwater” and “saltwater” schools, aka Real Business Cycle and New Keynesian) choose for its microfoundations a highly technical set of assumptions that privileged formally specified models of individual behavior over accurate, empirical ones. In other words, Glasner argues for a macroeconomics based on how humans actually decide (with limited information, search costs, uncertainty, etc.). Andrew Gelman makes a similar argument here, with added insights into the politics of model choice (e.g., why is rational choice conservative?).

At the same time, I think sociologists have been having similar debates for approximately forever (+/- 1 day) about agency and structure, mechanisms, and all that. Omar’s recent posts on the problem of agency are particularly excellent (both for his summary and new suggestions). At the same time, Dan Little and John Levi Martin have been going back and forth about JLM’s book Explanation of Social Action and their disagreements over the role of “mechanisms” vs. phenomenological accounts in explaining social action. More broadly, the whole “mechanisms turn” fits into longrunning debates about the proper “level” for sociological analysis: individual action or big institutions, agency or structure, etc.

I wonder to what extent a comparison of these debates in economics and sociology would prove fruitful. In some ways, the debate in economics is much “cleaner” – microfounded macroeconomic models look very different from old Keynesian IS-LM models. The ideological lines are also clearer (though as Gelman notes, it’s not 100% clear why the lines were drawn the way they were). In sociology, the debates are a bit more open-ended, and the methodological variation much more extreme. Debates about mechanisms or structure and agency in cultural sociology, economic sociology and stratification (to pick three at semi-random) will draw from radically different kinds of analysis, from hermeneutic close-readings to regression models of Fortune 500 business decisions. The relative consensus, or at least structure dissensus, in economics casts the foundations in much sharper relief, while the plurality of approaches in sociology allows for all sorts of interesting “mismatches” (where, for example, does Harrison White fit? My colleague Lotus Seeley argues that White’s work has as much in common with Foucault and Judith Butler as it does more conventional approaches in sociology, and yet White is famous for dense, formal, mathematical models of market behavior).

What do you think? Is this a useful comparison? Is there a single parallel debate in Sociology, or a spectrum of ill-defined debates?