New draft: Friedman’s instrumentalism in F53. A Weberian reading

In this paper I use Weber’s methodology of ideal types to argue for the instrumentalist interpretation of Friedman’s methodology of positive economics. I characterize Weber’s ideal types as a mix of descriptive inaccuracy and causal adequacy. Based on some recent structuralist results in the philosophy of science I show how intimately causal understanding and the properties of entities are related. The main contrast between Weber and Friedman lies in the emphases they placed on the causal properties of agents. It is argued that Friedman’s instrumentalism results from his neglect of properties for no causal understanding can be placed upon neglected properties. By identifying some possible channels through which methodological Weberianism could spread, the paper raises the possibility of actual Weberian tendencies in Friedman economic methodology.

I opened a session at Any suggestions or comments are warmly welcome.


Some notes on Lucas and the reducibility/supervenience of macro

If we are ready to accept that macroeconomics contains genuine and irreducible macroeconomic entities, and that agents act in an environment that has a plethora of non-individualistic properties, in the micro-founded models there must be something over and above the agents. The two spheres are inseparably mixed and interrelated. Microeconomic universe also contains irreducible macroeconomic concepts. Consequently, as Hoover emphasized, a complete elimination is unfeasible. This is the reason why the critique against Lucasian economics can be turned into a positive argument. Criticizing Lucas for eliminating everything other than individuals and their properties is a misplaced attack. Recall that for Lucas a complete elimination was rather a vision or a mission statement than an accomplishment. What is more, the island models abound in such unreducible elements. So where is the problem? If macroeconomics rejects to be exclusively reduced to micro, then Lucas’ incomplete elimination is not a shortcoming – rather the completion of a plausible strategy. Admittedly, one can justifiably lay the blame on Lucas for envisioning complete eliminativism – but paying lip service to complete elimination on the one hand and completing that on the other hand are completely different stories. The critique against Lucas for his incompleteness results from an erroneously narrow concept of microeconomics where microeconomics is conceived to contain nothing but agents and their properties. By contrast, Lucasian microfoundations project can be assessed as a successful attempt for it contains only incomplete elimination. It is our duty to harness Hoover’s critique regarding the exact nature of the microeconomic universe and to construct a fair critique against the microfoundations project. If we are ready to properly understand the microfoundations project (what can and what cannot be done), we can separate Lucas’ accomplishments from his ambitions. Thus our question below regards whether Lucas so constructed his representative agent as to seize the relevant causal properties of real agents. In Chapter 4 of The Friedman-Lucas transition in macroeconomics I argue that his effort to construct his representative agent out of the properties of its real counterparts was an important step towards re-establishing the causal relevance in macroeconomics and turning away from Friedmanian non-causalist instrumentalism. I characterize this return to a sound causal analysis as the most important facet of the Friedman–Lucas transition in macroeconomics. Even though the Lucasian islander as an abstraction is very far from reality, it is still an abstraction, i.e. built from some isolated elements of reality. Similarly, even though Lucas’ society consists of only the replicas of the one and only representative agent, it is still a society where social relations can be studied in an extremely primitive form. For Lucas, such a simplification seemed unavoidable for modelling heterogenous agents is fraught with both computational and aggregation difficulties.


Mr. Friedman and Mr. Ronnie O’Sullivan

Out of his three examples in F53, Friedman’s billiard or snooker player is a special case. In his philosophy of science, Michael Polany draws attention to the tacit component in our scientific (and everyday) knowledge. Accordingly, our knowledge contains components, in our acts we rely on skills that we know only in an unconscious way. On these tacit or unaware components there is no focal awareness placed. Out of his many examples it is the cyclist that is relevant as far as F53’s snooker player is considered (Polanyi, 1958/2005, p. 51). There is a simple rule every cyclist obeys. When he starts falling to the right he turns the handlebars to the same direction, so that the path of his bike is deflected along a curve towards the right. This manoeuvre triggers off a centrifugal force pushing the cyclist to the left and thus offsets the gravitational force dragging him down to the right. This correction throws the cyclist off balance to the left, which he needs to counteract by turning the handlebars to the left. The cyclist can keep himself in balance by manoeuvring along a series of appropriate curves. The path the cyclist ought to follow can simply described in mathematical terms: for a given angle of unbalance the curvature of each winding is inversely proportional to the square of the speed at which the cyclist is proceeding. Consequently, a cyclist factually applies a physical rule, thus his behaviour can be well described with an appropriate formula. His awareness of the formula is subsidiary. Even if he is familiar with the physical background of his cycling, riding his bike is not supported by a constant calculation process. Moreover, if his focal awareness was directed at the physical rule and its constant application, this would inevitably result in his crashing into a tree. Thus, describing the behaviour of a snooker player by a mathematical formula is neither abstraction nor idealization. The feasibility of such an epistemological strategy of ours is underpinned by the structure of human knowledge we want to model.


Polanyi, M. (1958/2005). Personal knowledge. Towards a post-critical philosophy. London: Routledge.


Some notes on the commonsensibles in economics

Hoover (2001, pp. 229-230) calls attention that Mäki’s commonsensibles interpretation is not cogent for some concepts of economics than for others. For instance, everyday people of folk economics have problems understanding such concepts as the’real GDP’ or ’the general price level’. Hoover distinguishes natural aggregates and synthetic aggregates. For him, only natural aggregates can only be concieved as commonsensibles for these simple sums of averages as measures have the same dimensionality as their individual components. Total unemployment or the average rate of interest can easly be interpreted based on our everyday experience. However, synthetic aggregates consist of components in a way that modifies the dimensionality of the parts, thus no easy or analogous interpretation is available. Interesting, but Arrow (1974, pp. 253-254) talks about economic equilibrium in this vein. In the everyday, normal experience laymen can easily interpret a situation in which the quantity supplied and the quantity demanded are equal. However, this understanding is not so thorough as to cover the understanding how the interplay of underlying mechanisms (i.e. the reallocation of production inputs or interactions between shifts in technology and the allocation of the labour force) can lead to general economic equilibrium. As another example, Hands (2012) suggests contemporary revealed preference theory (CRPT) as a radical departure from this standard of commonsensibles. Thus for him we are not justified to say economics as such operates exclusively with commonsensibles. Some qualifications are needed. However, this has nothing to do with Mäki’s commonsense realism for this framework still works as a normative standard. In spite of the apparent problems, we do have commonsensibles in economics. Citing Lionel Robbins (1935/1984, pp. 75-79), Hands (2008) argues for the significance of introspection in economic theorizing. Over and above obvious empirical facts such as scarcity (which can also be conceived as a commonsensible), some economic concepts of fundamental imprtance (e.g. preferences) come not from experiments or observation but from self-knowledge. Any individual knows by self-observation or inner everyday experience that different things have different importance to him or her and in terms of importance things can be arranged in a certain order. Thus we all have clear mental pictures (which we can make even clearer by apt theorizing) of how economic agents think, decide or act. For the present argumentation it is enough to underline that there is no confusion in our ideas regarding the fundamental properties of economic agents. We can easily judge whether a theoretical description of a ’typical’ agent corresponds to our views on ourselves. Considering the success of abstraction by such an intuitive comparison is an undemanding easy job in most cases.


Arrow, K. J. (1974). General economic equilibrium. Purpose, analytic techniques, collective choice. Nobel lecture. The American Economic Review, 64(3), 253-272.

Hands, D. W. (2008). Introspection, revealed preference and neoclassical economics. A critical response to Don Ross on the Robbins-Samuelson argument pattern. Journal of the History of Economic Thought, 30(4), 453-478.

Hands, D. W. (2012). Realism, commonsensibles, and economics. In A. Lehtinen, J. Kuorikoski, & P. Ylikoski (Eds.), Economics for real. Uskali Mäki and the place of truth in economics (pp. 156-178). Oxford: Routledge.

Hoover, K. D. (2001). Is macroeconomics for real? In U. Mäki (Ed.), The economic world view. Studies in the ontology of economics (pp. 225-245). Cambridge: Cambridge University Press.

Robbins, L. (1935/1984). The nature and significance of economic science. London: Macmillan.


The Friedman-Lucas transition in macroeconomics. A structuralist approach

These months I am working on my new book project. There is a lot to do since I would like to complete all the chapters that I can finish here at home to have only a minor part of the project to be completed in Chicago, at Duke and at Arizona State University. In the meatime the title of the forthcoming book seems to have got a final form: it is

The Friedman-Lucas transition in macroeconomics. A structuralist approach.

In this post I would like to provide a bird-eye view of the contents as planned in the current stage.

Chapter 1. The introduction of the book is devoted to demonstrating the idea that methodology plays a crucial role in forming our ideas in economics. The discussion of this basic tenet in this chapter is twofold. First, I will review some methodological aspects of the neoclassical–institutional controversy. Their special theoretical problems directed both institutionalism and mainstream economics to unique methodological ways. Consequently, it has become impossible for the camps to mutually cultivate their approaches. Due to the special interests and the accompanying methodologies, these schools have been unable to pay attention to certain mechanisms of socio-economic reality.

Second, as the realism of theories or models can be doubted on multiple grounds, in this chapter I will characterize realism as a purpose, desire or ambition. This is the only thing we have for sure: a realist wants to be realist. The most we have about realism is our ambition to be realists. This is the purpose along which we can distinguish realists and instrumentalists. Accordingly, I will underlie that any methodological analysis of theories in the realism-instrumentalism context cannot be complete without scrutinizing the purpose of the theorists. Here understanding (i.e. providing causal knowledge as a purpose) stands in the focus as a feature of realists—in contrast to instrumentalists’ emphasis on mere predictive purposes (or to have knowledge only of the observable phenomena while denying the possibility of knowing the unobservable: e.g. causes).

Chapter 2. This chapter is devoted to a historical overview, as the reviewers wanted to read more about the THEORIES and the changes in econometric practice and the empirical performance of these models. Here I can talk about RBC and new Keynesian models as alternatives to Lucas’ line. I think this is going to be a very interesting chapter since a lot of details will be included about Chicago economics, the transition from Friedman to Lucas, etc. The line of reasoning can be summarized along the line from Hume’s thesis to the rethinking of the role of the government.

Chapter 3. This is the core of the book: here I will introduce the methodological framework to be used for evaluating the Friedman-Lucas transition. This is hard core philosophy of science, but customized for economists – so it is going to be an easy read hopefully. In this chapter, the chosen framework for a structuralist analysis of modern business-cycle theory is considered. In order to respond to the debates around the realist and instrumentalist interpretations of the works of Friedman I will provide a framework in which the tension between the opposite standpoints can be overcome. Based on epistemic structural realism, first, I will identify Friedman’s stance as a mix of agent-level instrumentalism and macro-level causal realism and, second, I will scrutinize the consistency of this epistemic strategy. I will argue that causal realism requires one to be realist regarding the assumptions defining agents as well.

Within the framework of ESR the problem I can solve is whether the representation of relations can be separated from the representation of entities. I will argue that structural realism cannot do without entity realism. In other words, a structuralism detached from the representation of entities (i.e. from the approximately true descriptions of agents) is not a viable option in economics. Structure and causality are of primary importance in creating scientific knowledge. In these terms, the only problem is that knowledge of a structure cannot be separated from the knowledge of entities. Their causal properties and dispositions for behaviour are of crucial importance in forming their (causal) relations. Entities having other properties stand in other relations, thus properties of entities are not neutral in terms of relations. This is the reason why the micro-foundations are so important in establishing the macro-level causal connections.

Chapter 4. This chapter is devoted to an assessment. One strength is the extensive reading of Lucas’ unpublished manuscripts with lots of word-to-word citations. Here I will provide an interpretation based on Max Weber’s traditional neoclassical methodology that can help us to find the adequate territory of abstract-idealized mainstream models. Special attention is paid to the intellectual connection between Weber and the subsequent business-cycle theorists. It is argued here that scrutinizing the relationship between models and reality should reach beyond the oversimplifying approach of empirical tests. Simultaneously, this assessment gives us a refined set of aspects by the application of which to contemporary macroeconomics some methodological breaks can also be highlighted. These breaks would not have been revealed by a linear narrative.

Here Max Weber’s social scientific methodology is introduced as an adequate strategy to establish the macro-level causal connections on realistic agent-level properties.

Chapter 5. In this concluding chapter I will consider the chances of a new theoretical consensus placed upon methodological grounds. I will argue that the complementarity of the approaches applied in macroeconomics and the related fields makes it possible to reconcile the competing approaches with one another without modifying the research traditions.

Reading Charles I. Plosser’s ‘Understanding real business cycles’

Charles Plosser’s review paper entitled ‘Understanding real business cycles‘ is a must for all those interested in the modern theories of the business cycles. A major advantage of the paper that it is written in plain English, it has a well-organized structure and summarizes by giving a clear vision of future development. Some of the most important suggestions of the paper:

  1. Keynesian macro is not a theory judged by the modern choice-theoretic standards. Keynesian models had significant empirical success but failed to trace the macro-level back to solid microfoundations.
  2. The aim of general equilibrium modelling is to help us to understand the role played by optimization and equilibrating tendencies at the agent-level. It makes no sense to postulate a disequilibrium framework or market failures before understanding the contribution of a general macroeconomic equilibrium.  Large-scale fluctuations may be triggered even in the absence of market failures.
  3. RBC-theory has common roots with Lucasian macro with the very nature of the stochastic shocks as a major difference. Shocks, be they technological disturbances or monetary noises, exhort the agents to dynamically adjust to the new situation. Business cycles emerge as the outcome of the equilibrating-optimizing responses of rational agents.
  4. RBC-theory is the unified theory envisioned by Lucas in his ‘Understanding business cycles‘. As Lucas puts it:

Though there is absolutely no theoretical reason to anticipate it, one is led by the facts to conclude that, with respect to the qualitative behavior of co-movements among series, business cycles are all alike. To theoretically inclined economists, this conclusion should be attractive and challenging, for it suggests the possibility of a unified explanation of business cycles, grounded in the general laws governing market economies, rather than in political or institutional characteristics specific to particular countries or periods.

RBC-theory is a highly flexible framework that can be tailored to the needs of a researcher, according to the nature of the problem under scrutiny. Early-stage RBC-models were constructed to analyze the effects of technological shocks, but it is easy to modify the general framework in order to scrutinize the effects of seemengly all kinds of macroeconomic shocks: modifications in tax policy, changes in consumer taste, changes in monetary or fiscal policy and the like. RBC-theory seems to signify the end of macroeconomics: the time of the grand theories is over – once and for all?


Absorbed in James Forder’s Phillips curve myth

An outstanding book.

It was my first thought after reading James Forder’s seminal work ‘Macroeconomics and the Phillips curve myth‘. Dr. Forder breaks down a plethora of age-old fallacies written in the historiography of economics and reshapes the 20th century history of our discipline.

Personally, I am not a researcher of the Phillips curve per se. Only two chapters of this long story are most interesting for me, Friedman’s and Lucas’ Phillips curves, which are highly influential episodes of the story. I clearly remember, I was afraid to start reading the book a couple of weeks ago, because Dr. Forder promises his readers to get the facts straight. This is his very program: to read everything that belongs to the Phillips curve literature and correct all the errors we inherited from the previous decades. And this is the program Dr. Forder accomplishes.

For me personally, the most interesting part is the chapters in which the author clears up the true role of Milton Friedman‘s Nobel lecture in the distortion of the history of the Phillips curve and how Friedman came up with a brand new narrative. This was a shocking experience to realize that the commonly believed story is Friedman’s invention. In Friedman’s reading the Phillips curve was a high-ranked constituent of both theoretical economics and economic policy discussions from the early post-war years. In Friedman’s account, the Phillips curve was the most important analytical tool of Keynesian economics and Friedman places the blame on the Phillips curve (and the believers) for all the allaged mistakes of theoretical economics and economic policy. In a nutshell, it is failure of the Phillips curve that economics derailed – and Friedman’s ambition was to put economics back on the right track. His soultion was the inclusion of expectations in order to call attention to the fact that there is no stable trade-off between inflation and unemployment. On the contrary, Phillips’ contribution was only one element in the vast literature on the trade-off between inflation and employment/unemployment (and what is more, not the most important or most inspiring one) and the Phillips-curve was not widely used in Keynesian economics to argue for inflationary politics in favour of lower unemployment rates. Friedman was unable to change the game in terms of the Phillips curve because such a game had hardly existed. However, thanks to Friedman, after his reinterpretation the Phillips-curve really came to the fore: it became the common language, the common denominator through which monetary-interested authors could express their views. In this term the Friedman-Lucas transition is particularly interesting. Setting aside the famous Marshall-Walras divide, the central feature of the debate between Friedman and Lucas was the focus on the shape of the Phillips curve and whether there is a distinction between the short-run curve(s) and the long-run vertical line.

From time to time we need to face the fallacies of our thinking. Some courageous authors like James Forder are not afraid to rebuild our science from the ultimate grounds. To be perfectly honest, sometimes I feel tempted to think there is no real progress in our discipline. Oftentimes it seems as if the cornerstones of economics and the history of economic thought are established once and for all. But some ambitious books such as Forder’s Phillips curve myth can be successful in reopening our most important questions. This is the reason why Forder’s book is a must: it can teach us how reshape our knowledge through sound reasoning.

Here is a review from Kevin D. Hoover, and another one from Michel De Vroey. Both are worth reading.



A structuralist approach to the methodology of modern macroeconomics

I am proud to say that a new book project of mine has started. In the forthcoming book to be published at Elsevier I am studying the relationship between agent-level assumptions and the possibilities of an adequate macro-level causal analysis. Following a realist philosophy of science I am interested in demonstrating that we are to distinguish causal adequacy and descriptive accurace when it comes to discuss the performance of abstract models along the Friedman-Lucas-RBC line.


The problem in analysing assumptions in economics

The common problem in discussing the design of assumptions in economics is rooted in merging two independent aspects into each other. One is descriptive accuracy: how our assumptions can provide true or realistic descriptions of reality. It is a heritage of neoclassiccal economics to abandon this purpose. The more “unrealistic” a model, the more useful it is – this is the summary of the widely believed tenet. However, this aspect tells us nothing about the causal adequacy of a model: highly abstract models with highly unrealistic assumptions (unrealistic is descriptive terms) can do a great deal about causal understanding. So this is the second aspect: causal adequacy. We are to check whether our descriptively unrealistc and false assumptions are adequate in causal terms. And they can be. The only thing to do is to highlight relevant and causally active entity properties. Unrealistic models can have unrealistic assumptions (again: unrealistic in descriptive terms) which are realistic in causal terms. This is the most important consequence of the structuralist reconstruction of business-cycle models and the famous microfoundations project along Anjan Chakravartty’s philosophy of science exerted in his A Metaphysics for Scientific Realism (Cambridge University Press, 2007).


Considering Weber’s social scientific realism

I spent the last couple of days reading. I am making efforts to reword Weber’s methodological stance through concepts I am familiar with. Here I would love to refer to entity realism and structural/causal realism. I tried to concieve what Weber would think and say in these terms.

Entity realism: this comes from the philosophy of physics, currently applied to social sciences. At the bottom line this idea is very simple: our scientific terms refer to entities that are real. There are electrons in reality and they have the properties we attribute to them. This problematic, I know. No theoretical description is full, so when an entity realist characterize electrons, he can portray only ’a slice of reality’ at best. An entity realist is unable to seize all the features of electron, but – and this is important – he attributes features to the electron that it really has. I think Weber would be an entity realist. I found a very interesting paragraph in Ringer’s Max Weber’s methodology. Referring to Mommsen, on page 118 the author writes: ideal types cannot be constructed without regard for realities. So, and this is very essential to me, Weberian abstract-theoretical ideal types are constructed (and should be constructed) so that the researcher examines reality and seize those features of the entities that are important to him as for his research interest. As I concieve ideal types, an ideal type is constructed from real features. Rationality of the homo oeconomicus is the rationality of everyday human beings – but in reality this rationality is corrupted or disturbed by a lot of mental, psychological and social (and so forth) factors. But rationality is ’out there’, it is a part of our complex picture. Following Weber we are interested in partial truths, but the elements we exaggerate in ideal types are real features of our fellow agents.

Structural/causal realism: this also comes from the philosophy of physics. To simply put, the causal/structural connections we depict in theories are real. If I tried to conceive neoclassical economics in a Weberian fashion and according to structural/causal realism, I would say: Weber is a structural realist. In these terms, however, we have only partial truths again: treading Weber’s footsteps, we seize some causal mechanisms of the complex real causal structure – but these mechanisms are real and effective causal mechanisms. I think Weber would say: in the Marshallian cross, an increase in the price would really reduce demand. That is, the mechanism underlying this basic model of market exchange is about an effectice cause-and-effect connection – much more than an empirical regularity. When I read Whimster and Bruun’s brilliant foreword to the collected methodological writings, I found a passage that seems to support my idea: on page xxvi they trace Weber’s causal notion back to Radbruch and place Weber’s causality in the context of legal responsibility. So, I think, Weber is interested in real and objective causal connections.

As far as the realism-instrumentalism debate is considered, I believe Weber is important part of the context. Let’s have Friedman’s instrumentalism: with his parable in F53 on the leaves acting as if they were able to rationally consider their positions, Friedman gives up entity realism. He argues for assumptions that can be set up neglecting the real features of entities he ’portrays’. Thus, Friedman abandons the idea of entity realism (even as partial truths). It is questionable whether Friedman was a causal realist. Accordig to Kevin Hoover, he was, but I don’t think so (I think entity realism and causal realism is a package, causall realism cannot be achieved without entity realism). To simply put, with his emphasis on predictive performance of theories, he pays no attention to the causal connections behind macro-social phenomena. I think the whole F53 paper is devoted to the idea that researchers should be uninterested in the causal connections, and it is only predictive performance that matters. In these terms, Hausman rightly debates Friedman’s methodological stance.

To have an example: there is a widely cited anecdote in statics about a ’model’ in which the number of storks is the independent and the number of newborn babies is the dependent variable. For Friedman, such a model is a useful model as long as it provides good predictions. For Weber, such a model, I think, would be unacceptable for such a model would say nothing about the real causal connections.

Is this hairrising…?