On criticism and reflexivity


Recently, I was silent on the blog because I was very busy with completing my new book manuscript ‘Fundamentals of Critical Economics’ (in German: Grundlegung einer kritischen Theorie der Wirtschaft, publisher: Metropolis). Depending on your perspective, this weighty volume of 620 pages is a courageous or lunatic attempt at changing economics from the bottom up. I think that the approach is highly relevant for our INSOSCI project, however.

Traditionally, economics is driven by research on the micro- and the macro-level, as far as the core disciplines are concerned. Both types of research tend to adopt mechanistic perspectives in a naïve way, as they claim to develop linear causal explanations of economic phenomena. Take macroeconomics: The economy is conceived as a kind of mega-machine which processes exogenous causes to produce economic outcomes. In old-style post-war Keynesian economics, that was the view of the aptly christened ‘hydraulic’ approach of structural macro-models. After the rational expectations revolution, a strange beast emerged in the shape of dynamic stochastic equilibrium DSGE models: These work with highly abstract theoretical models of the economy that deliberately eschew the claim of the older structural models to closely describe the mechanisms of the real economy (such as working with constructs such as the ‘representative firm’), but at the same time submit strong claims on causality in econometric analyses of the impact of external shocks on the economy. However, in both strands of macroeconomic thought something very essential is missing: The real-world individual.

If we go back to the birth-time of modern macroeconomics, Keynes himself adopted a very different view. On the one hand, he referred to real-world individuals in grounding macroeconomics on behavioural theory, such as in his theory of savings and consumption as habitual phenomena, and, of course, in his approach to expectations. However, at the same time he argued that there is a distinct, emerging level of macro-phenomena which is not related at all to micro-level phenomena: This is his theory of aggregate demand, famously enshrined in the ‘paradox of savings’. Modern macro in terms of DSGE models simply nullifies this essential movement from micro to macro in doing away with the aggregation problem in a system of complex accounts and dynamics of stocks and balances which drives individual expectations. At the core is the peculiar reference to reflexivity in the so-called ‘Lucas critique’ of Keynesian macro models.

The Lucas critique was one of the greatest errors in the history of economic thought, I think. Only very recently economists try to get rid of that burden: For example, the leading macroeconomist Blanchard (2017) started to argue that one needs two types of models, with structural models receiving equal attention as DSGE models, especially in policy advice. Well, he does not take the embarrassing final step to acknowledge that the way from the Lucas critique to the DSGE models was a cul-de-sac! (more explicitly, see Buiter 2016) If we cannot use models for policy design and realistic analysis of the economy, what is the purpose of pure theory? Should we continue to think that the earth is the centre of the world, while orienting practical action at the view that the Earth cycles around the sun?

The ‘rational expectations revolution’ was a fundamental error in two respects, and this exactly refers to the two aspects that Keynes brought to the fore. First, it prevented us from studying the role of behaviour in shaping macroeconomic developments. That became clear in the Great Financial Crisis: Herding, risk-taking, expectations and uncertainty, all that and much more must be part and parcel of macro. Combining rational expectations with the modern macro-foundations program was simply wrong. Second, it did away with the need to study the interaction between institutional structures and macro-phenomena, with the former channelling behaviour in a way that transcends properties that individuals have. For example, recent research shows the strong impact of financial network structure on crisis dynamics (for example, Acemoglu et al. 2015).

However, the Lucas critique was so powerful because it revealed a fundamental weakness of hydraulic Keynesianism: This is the exclusion of reflexivity. Human agents learn about economic models, and they will anticipate what happens if models are applied, thus changing their behaviour, and eventually undermining the assumptions of the models. The hydraulic models deleted reflexivity, which was well recognized by Keynes, for example, in his famous parable of the ‘beauty contest’ and the formation of expectations.

Thus, I think that in designing new economic theories we need to take reflexivity as a central assumption. In the incipient new approaches to behavioural macro, there is a great danger that behavioural economics results are applied in a non-reflexive way (compare Baddeley 2014). That would almost certainly revive the Lucas critique, and rightly so: What happens with our macro models, if people learn about the predictions of behavioural economists and adapt their behaviour accordingly? An even more fundamental paradox lurks here: Can we imagine of rational actors who respond optimally to behavioural economics research? Do we need a theory of reflexivity that would analyse how real-world individuals use scientific knowledge about real-world individuals to take decisions?



Acemoglu, D., A. Ozdaglar und A. Tahbaz-Salehi (2015): Systemic Risk and Stability in Financial Networks, American Economic Review 105 (2), S. 564–608

Baddeley, M. (2014): Rethinking the Micro-foundations of Macroeconomics: Insights from Behavioural Economics, European Journal of Economics and Economic Policies: Intervention, 11 (1), 2014, S. 99–112

Blanchard, O. (2017): The Need for Different Classes of Macroeconomic Models, Peterson Institute of International Economics,

Buiter, W. (2009): The Unfortunate Uselessness of Most ‘State of the Art’ Academic Monetary Economics’, VOXEU, March,


Funded by: