The ontological status of the models of the business cycles is widely debated. The fundamental question regards whether these models can give approximately true descriptions of the causal structures underlying the relevant sets of socio-economic phenomena—or they are useful instruments only capable of generating predictions with considerable empirical performance. This is the conflict that underlies the debate between economic realism and instrumentalism. Although Uskali Mäki provided efficient arguments for the realism of economics, according to the currently standard interpretation Friedman put economics on an instrumentalist track. Moreover, both Lucas’ and the succeeding RBC-theory are conceived as further steps along this line. This is a rather worrying interpretation that can deprive economics of the possibility of looking for the causes of macro-social phenomena (i.e. the possibility of understanding in a Weberian sense). This interpretation may delimit even the scope of economic policy applications.
Kevin D. Hoover made efforts to argue for Friedman’s realism along a more modest but tenable line of reasoning. Hoover highlighted Friedman’s causal realist achievements. For Hoover, Friedman set up instrumentalist agent-level assumptions in order that causal realism could be achieved at a macro-level. Due to his interpretation, the earlier debate between full-blown realism and instrumentalism could be set aside. According to the most careful interpretation currently available causal realism can be achieved by instrumentalist agent-level assumptions. However, our knowledge of a causal realism built on agent-level instrumentalism in economics is rather limited.
My ultimate purpose is to consider the plausibility of the mix mentioned above, to extend that to be a sensitive analytical framework in which a methodologically-oriented history of modern business-cycle theory can be outlined. The suggested framework is Epistemic Structural Realism (ESR) worked out by John Worrall and others. ESR describes the relations connecting entities (here: economic agents) so that the first-order properties of agents can remain hidden. Here almost everything is structure and structure is interpreted at the level of the structure itself. As causal mechanisms can work only along the relations that connect entities, ESR is the general philosophical framework in which it can be analysed whether structural/causal realism underpinned by entity-level instrumentalism is a viable option both as an interpretation for the history of economics and as a research strategy.
According to my hypothesis, the answer is negative. Chakravartty and others called attention to the fact that knowledge of a structure involves knowledge of the entities in the object-based ontology that characterizes the micro-foundations project of modern macroeconomics. It means that providing a realistic (verisimilar) description of a structure is not possible on the basis of assumptions that define the related entities in an instrumentalist manner. A given real causal structure is not compatible with an entity-ontology chosen arbitrarily. In this approach the “microfoundations project” will obtain a new meaning, since the relationship between the microfoundations and the macro-level knowledge can be clarified. Microfoundations are the instruments of connecting with reality. On the basis of the critically reconsidered ESR-based reconstruction we can judge which chapters of the theoretical evolution are anchored to socio-economic reality. Moreover, the methodological breaks that make the linear evolution a dynamic relationship with reality can also be revealed. Thanks to this analysis, our knowledge of the way some assumptions are set up will considerably improve. In the methodological corpus these assumptions are uniformly labelled as “unrealistic”. Whilst such assumptions have only limited empirical content, they can be created by different cognitive techniques. One of them is the “cooked-up” (instrumentalist) technique advocated by Friedman. This procedure is justified only by the empirical performance of the models, whilst the act of understanding is not supported here. In the other case, however, by performing careful abstraction we can preserve the relevant features (i.e. the structural properties) of the agents in order to analyse their macroeconomic/macro-level consequences. By performing this methodological analysis, I can describe the evolution that has modified the relevance of theories with respect to understanding socio-economic reality. Methodological considerations turn the story into a dynamic evolution interspersed with methodological breaks.
Robert E. Lucas Jr.
Finn E. Kydland
Edward C. Prescott