Between Keynes and Neoclassical: Narratives that create sense and nonsense in modern macroeconomics

_ Franz R. Hahn, economic researcher, Austrian Institute for Economic Research (WIFO). Translated into English by Yuri Kofner. Published for debate. 20 September 2020.

In the 1970s the foundations of modern macroeconomics could have been clarified. However, this possibility was not taken up or deliberately ignored by the ranks of economists.

The 1970s are seen in the annals of economics and economic policy as a period of fundamental change. The Keynesian school of thought, which had dominated economic policy until then, lost its primacy in this decade and was replaced by an understanding of politics that was close to the principles of neoclassical economics. This neo-classical political view was the result of a reorientation, or rather a return, of academic macroeconomics to its classical, but formally elegantly draped doctrinal views that are in keeping with the times.

What is remarkable about this paradigm shift is that this reorientation of macroeconomics took place just at the time when basic economic research not only questioned classical economically liberal doctrines but refuted them as unfounded. The findings of the formal mathematical analysis of the General Equilibrium Theory (GET) published in the 1970s primarily refute the central economically liberal postulate that the neoclassical trinity of rationality, self-interest, and market efficiency necessarily and unambiguously result in the socially best possible overall economic result (i.e. a clearly definable stable and generally benign equilibrium).

Although the formal foundations of the GET and their analysis are very abstract and have no direct equivalent, in reality, the ‘ideal-typical theory building’ of the GET is not only considered the nucleus of modern economic liberalism but above all as the central reference system of the (i.e. classical or neoclassical) economic theory in the measurement of general economic issues. Fundamental localizations such as whether market economies systemically require political control or only temporarily, e.g. in the exceptional case of a supply shock (e.g. Covid-19 crisis), are not only directly or indirectly related by academic macroeconomists, but often also by economic practitioners to the guiding principles of the GET. Economic liberals usually take positions from classical economics, especially the presumption of the self-organizing (competitive) markets (the Smith’s ‘invisible hand’) for the benefit of all. Market-critical (e.g. Keynesians), on the other hand, emphasize the cognitive dissonances in the central ‘narrative strand’ of economic liberalism and therefore primarily negate the economic policy conclusions that economic liberals draw from the neoclassical trinity. They, therefore, advocate systematic public economic management.

Up until the 1970s, these diametrically opposed fundamental macroeconomic theses were indissoluble even on the basis of abstract-formal basic research.

In the 1970s, in a remarkable series of formal-analytical papers – which even for academic economists often seem almost esoteric – the foundations were laid for bringing the most important fundamental questions of macroeconomics to a final clarification. With the abstract methods of the GET it was proven that none of the central leitmotifs of the ‘economically liberal narrative’ can be substantiated under the ideal laboratory conditions of neoclassical economic theory.

The fatal blow was among others one of the leading figures of the GET, the French-born US mathematician Gerard Debreu. The so-called Sonnenschein-Mantel-Debreu theorem (SMD theorem) is in fact a finding that can hardly be surpassed in terms of ‘destructive power’.

The central message of the SMD theorem is that the neoclassical assumptions of rationality, self-interest, and market efficiency are necessary, but not sufficient for the (theoretically founded) validity of central classical doctrines [1]. The neoclassical trinity does not necessarily lead to ONE stable, macroeconomically efficient equilibrium, but to a large number of general market equilibria, which are generally clearly determined locally, but whose stability properties all remain undetermined.

Meaning-making’ Keynes – ‘Nonsense-making’ Walras?

The SMD theorem seals the untenability of classical doctrines with a devastating overall result. The economically liberal narrative of self-organizing, efficient markets lacks the theoretically concise foundation. It is not possible to theoretically justify ‘classical market forces’ (invisible hand) with sufficient strength that would be able to keep (ideally competitive) market economies in balance or to bring them back there after external shocks.

These properties all turn out to be ‘typical’ for competitive market economies and not, for example, as ‘black swan phenomena’ or theoretically rare marginal phenomena [2].

It is absurd that the ‘Keynesian narrative’, which was decisive up to the 1970s and whose creation of meaning was fully confirmed by the final results of the GET in the above-mentioned decade, was superseded or replaced as a macroeconomic ladder narrative. The central Keynesian assumption from the 1930s that capitalist laissez-faire market economies are typically indeterminate and structurally unstable ultimately proved to be stringently justified by the findings of the GET. The classic doctrines of the socially benign and efficient competitive markets, which should best be left to work undisturbed for the benefit of the general public, turned out to be mirages, at best as ‘very special cases’.

Although the GET ultimately confirmed Keynes’s hypothesis and rejected the Walras ‘, the latter’s school of thought prevailed, at least in academic macroeconomics.

Smart new classics – dull old Keynesians?

A number of guesses come to mind as to why this could happen. One of the reasons is that in the 1970s and 1980s the overwhelming majority of macroeconomists, not just the Keynesian-oriented school of thought, mistakenly viewed the formal analysis of GET and its results as irrelevant to clarifying fundamental questions of macroeconomics. Ultimately, the Keynesian side also lacked the formal-analytical competence to evaluate the results of the GET from their point of view and to classify them in a meaningful way.

The neo-classical macroeconomists, on the other hand, allowed themselves to be (led) exclusively by the early findings of the GET from the 1950s, above all by the formal proof that a general efficient equilibrium exists under neoclassical assumptions. However, they ignored the, from their point of view, “subversive” results of the GET from the 1970s (keyword: many, qualitatively indeterminable equilibria). They did not develop the GET further but corrupted or shortened the classical theory possibly with the understandable, but under the given circumstances more than questionable justification to at least protect the liberal doctrines from the loss of meaning in academic macroeconomics (keyword: ‘Representative Agent ‘-Model).

The central message of the GET that the premises and market mechanisms of the classical economy are necessary, but not sufficient for a clear and structurally stable general equilibrium solution in a competitive market economy, did not find its way into the defining narratives of modern macroeconomics.

General Equilibrium Theory – An Unfinished Story!

The large number of structurally unstable general equilibria that are compatible with the assumptions of the GET is an unmistakable sign that the GET is telling a ‘potentially meaningful, but obviously unfinished story’. The neoclassical trinity (rationality, self-interest, and market efficiency) is insufficient to close the ‘open-end’ of this theory. A broader axiomatic foundation is required for the last step, the justification of a selection mechanism that clearly determines a structurally stable, generally desired equilibrium from the multitude of qualitatively different equilibria [3].

Again, it was Keynes who ‘instinctively’ guessed how the classic story should be conclusively told to the end. He recognized (probably as the first modern macroeconomist of rank) the fundamental importance of so-called ‘animal spirits’ for the unambiguous determination of a general equilibrium in a market economy. ‘Beliefs’, for example in the sense of rationally unjustifiable moods (optimism, pessimism) or ‘sunspots’, are still viewed by many neoclassical economic theorists as disturbing anomalies, and not as the final part of the solution to the central ‘classical indeterminacy problem’. In fact, beliefs are the “missing axiom” in neoclassical economic theory or the sufficient axiomatic addition to the neoclassical trinity.

Beliefs, and not rationality or rational expectations, if they are shared by market participants, are responsible for the final selection of the determining balance [4]. Fortunately, the importance of ‘beliefs’ for neoclassical economic theory has been increasingly recognized since the financial crisis in 2007 and is now the subject of a very active macroeconomic research program (see e.g. Farmer, REA, The Macroeconomics of Self-Fulfilling Prophecies, The MIT Press, 1993, and the same, Expectations, Employment and Prices, Oxford University Press, 2010).

‘Beliefs’, however, have always been a central analytical component in various forms of economic history or in political-economic studies. The importance of the subtle-constructive interaction not only of formal but above all informal systems of rules (beliefs) and the market economy for prosperity and political stability was highlighted, among other things. impressively confirmed by a series of institutional economic analyses (see e.g. Acemoglu, D., Robinson, JA, Why Nations Fail – The Origins of Power, Prosperity, and Poverty, Crown Publisher, 2012, and the same, The Narrow Corridor: States, Societies, and the Fate of Liberty, Viking, 2019).

The meaning of ‘beliefs’ is also reflected in the results of economic-historical analyses of the interaction between modern capitalism and different political regimes. In particular, recent economic history clearly shows that the market economy or capitalist economic system can achieve and sustainably secure a high level of efficiency under different political (belief) systems. The two currently largest economic powers, the USA and China, are outstanding current examples of this (see e.g. Milanovic, B., Capitalism, Alone: The Future of the System That Rules the World, Harvard University Press, 2019, and Plumpe, W., Das cold hearts: Capitalism: the story of an ongoing revolution, Rowohlt Berlin, 2019).

Findings and research activities like these, at least those that seem suitable to bring the open narrative of neoclassical economic theory to a concise conclusion, are ultimately inevitable if academic macroeconomics is not to become meaningless.


In addition to the monographs mentioned in the text, the author recommends reading the relevant chapters in the following publications to supplement and deepen:

Arrow, K., Hahn, F., General Competitive Analysis, North-Holland, 1971

De Vroey, M., A History of Macroeconomics from Keynes to Lucas and Beyond, Cambridge University Press, 2016

Ellickson, B., Competitive Equilibrium – Theory and Applications, Cambridge University Press, 1993

Hildenbrand, W., Kirman, A. P., Equilibrium Analysis: Variations on Themes by Edgeworth and Walras, North-Holland, 1988

Mas-Colell, A., Whinston, M., D., Green, J., R., Microeconomic Theory, Oxford University Press, 1995

Rizvi, S. A. T., The Sonnenschein-Mantel-Debreu Results after Thirty Years,

History of Political Economy, 38 (annual suppl.), Duke University Press, 2006

Varian, H. R., Microeconomic Analysis, 2nd Edition, Norton, 1984

Wickens, M., Macroeconomic Theory – A Dynamic General Equilibrium Approach, 2nd Edition, Princeton University Press, 2011


1 The most important subjunction of the SMD theorem for macroeconomics is that it is not the individual but the aggregated demand that is of decisive importance for the existence, uniqueness, and stability of the competitive general equilibrium. However, the formal conditions for the aggregated excess demand are so general within the GET that in fact every continuous function that is homogeneous of degree zero in relation to goods prices and that fulfils Walras‘ law is compatible with general competitive market equilibria. The theoretical conclusions from this are very far-reaching and profound. Among the most serious are the fact that rational optimizing behaviour on the individual level is irrelevant for the qualitative properties of the aggregated excess demand and thus also for the existence of the general equilibrium. This eliminates the need or the formal requirement for a microeconomic foundation of macroeconomics (at least from the perspective of the GET). Due to the SMD theorem, the GET ultimately also loses its claim to being scientific. It turns out to be fundamentally non-falsifiable since any arrangement of strictly positive prices can be assigned to a suitable general equilibrium of a competitive market economy.

2 This also applies to one of the global characteristics of competitive market economies, structural instability. Shocks lead, for example, to qualitative changes in the dynamic properties of equilibria. Local equilibria typically not only change their local stability properties, even after only minor distortions, but can also assume indeterminate behaviour in the sense of chaos theory. Comparative-static and comparative-dynamic analyses are therefore typically not effective.

3 It is not uncommon in the history of science that a theory only turns out to be incomplete after decades or even centuries. The best-known example in this context is the classical mechanics founded by Isaac Newton, which was ‘told to the end’ at the beginning of the last century by the theory of relativity and quantum mechanics. Unfinished narratives often turn out to be an agent of scientific progress in science.

4 It is well known that rational expectations correspond to optimal forecasts if they are formed on the basis of neoclassical theory. ‘Beliefs’, on the other hand, are self-fulfilling, especially if they are shared by market participants without reference to neoclassical theory.


Leave a Reply

Your email address will not be published. Required fields are marked *