Leontief and the sorry state of economics

from Lars Syll

The core assumption of 'modern' macro — totally FUBAR | LARS P. SYLLPage after page of professional economic journals are filled with mathematical formulas leading the reader from sets of more or less plausible but entirely arbitrary assumptions to precisely stated but irrelevant theoretical conclusions …

Year after year economic theorists continue to produce scores of mathematical models and to explore in great detail their formal properties; and the econometricians fit algebraic functions of all possible shapes to essentially the same sets of data without being able to advance, in any perceptible way, a systematic understanding of the structure and the operations of a real economic system.

Wassily Leontief

Mainstream economics is, as noted by Leontief, hopelessly irrelevant to the understanding of the real world, and the main reason for this irrelevance is the failure of economists to match their methods with their subject of study. The fixation on constructing models showing the certainty of logical entailment has been detrimental to the development of relevant and realist economics. Insisting on formalistic-mathematical modeling forces the economist to give up on realism and real-world relevance.

It is — sad to say — a fact that within mainstream economics internal validity is everything and external validity next to nothing. Why anyone should be interested in that kinds of theories and models — as long as one does not come up with export licenses for the theories and models to the real world in which we live — is beyond comprehension.

  1. Garrett Connelly
    November 2, 2022 at 12:38 pm

    Thank you. Now I remember why I liked him when I was young.

  2. Reply
    November 2, 2022 at 2:09 pm

    Tibor Scitovsky was a contemporary of Wassily Leontief although located across the country at Stanford instead of Harvard. See his book The Joyless Economy as a useful companion to Leontief’s works.

  3. yoshinorishiozawa
    November 2, 2022 at 3:23 pm

    Lars Syll posted the same text two years ago on December the
    first, 2020:
    except that he has omitted the last sentence “Stupid models are of no help at all in understanding the real world” of the 2020 post in this new version.

    There were 20 comments on the 2020. How about the new post? I hope many commentators are not tired of Syll’s arguments which are almost always reactive.

    • November 6, 2022 at 5:09 pm

      A similar critique to the one Yoshinori puts forward here about not being “reactive” and that as long as one cannot come up with some own alternative model to the failing mainstream models, one shouldn’t expect people to pay attention, is in at least one way fundamentally wrong-headed.

      Why? Since it utterly misunderstands the role of philosophy and methodology of economics! What philosophy and methodology can contribute to economics is clearing obstacles to science by clarifying the limits and consequences of choosing specific modelling strategies, assumptions, and ontologies.

      People like Yoshinori don’t seem to get some of the basic rules of division of labour. By (pretending) not to see the need also in academia for specialisation — some of us are methodologists and others are model and theory builders — they conflate important differences. It’s somewhat like criticizing literary critics such as Sainte-Beuve or Bloom for not being literary writers such as Ibsen or Shakespeare. I think we have to accept that different specialisations are vital and necessary if we want to make scientific (and literary) progress!

      • Meta Capitalism
        November 6, 2022 at 8:38 pm

        History and philosophy, contrary to those who are so philosophically naive and/or willfully ignorant as to claim that neither helps us understand economics (or biology, physics, or for that matter any other scientific field) are ironically engaging in both history and philosophy themselves, albeit unconsciously without any awareness of their unexamined historical and/or philosophical presuppositions. Frequently, those who are so philosophically naïve as to dismiss history are also slaves to their own unexamined philosophical presuppositions; intellectual parrots to old dead men’s ideas.

    • yoshinorishiozawa
      November 10, 2022 at 5:14 pm

      Perhaps Lars became so furious that he forgot his cool head.

      I do not deny the usefulness and significance of philosophy and methodology of economics. My second book (in Japanese) was titled:
      Reflections on modern economics (Kindai keizaigaku no hansei, Nikkei Shibun, 1983). The introductory chapter had the title:
      Toward psychoanalysis of the science,
      where the “science” meant economics. The title of the first chapter was About equilibrium, the second On distribution, and the final third chapter The economic (L’économique in French, an inquiry on what is deemed “economic”).

      Rightly this book was criticized as too philosophical and methodological. Despite this criticism, I do not deny that this book was important for me. It helped me to guide my later research in economics in a right direction. My recent work with my colleagues Microfoundations of evolutionary economics (2019) adopts a causal and processual framework. Intentionally rejecting all kind of equilibrium analysis in this book, I was loyal to my young-days contentions.

      What I am complaining about Lars’s comments is that (1) he throws too categorical rejection of models, mathematics, theories, and even economics as a whole without considering any differences of contents of various strands in economics, and (2) he agues irrelevance of this and that theories and concepts of mainstream economics without hinting any solution or alternatives. As long as he continues to argue like this, it is rather reasonable to call his attitude reactive, if not reactionary.

      I repeated (1) and (2) points for these perhaps almost seven to eight years. I admit that Lars became more cautious about point (1). For example, he had a custom to name simply “economics” without adding adjective “mainstream.” Without such adjectives, Lars’s blames apply to all economics research including all heterodox economics. Blames are sometimes correct but sometimes not for some strands of heterodox economics. As far as I read his recent comments, there is no big amelioration with regards to point (2).

  4. Jan Milch
    November 2, 2022 at 9:33 pm

    Alfred Marshall on Mathematics:

    “Balliol Croft, Cambridge
    27. II. 06

    My dear Bowley,

    I have not been able to lay my hands on any notes on
    Mathematical Economics that would be of any use to you:
    and I have a very vague recollection of what I used to
    think on the subject. I never read maths now: in fact I’ve
    even forgotten how to integrate a lot of things.

    But I know that I had a growing feeling in the later years
    of my
    work on the subject that a good mathematical theorem
    dealing with economic hypotheses was very unlikely to be
    good economics: and I went more and more on the rules—

    (1) Use mathematics as a short-hand language, rather than
    as an inquiry engine.

    (2) Stick to them until you’ve done it.

    (3) Translate into English.

    (4) Then illustrate with examples that are important in real life.

    (5) Burn the math.

    (6) If you don’t succeed in 4, burn 3. The last one I often did…

    I believe in Newton’s Principia methods, because they carry
    with them so much of common sense. Mathematics used in
    a Fellowship thesis by a man who is not a mathematician by
    nature—and I have come across a great deal of it—seems
    to me an unalloyed evil.

    And I think you should do everything you can to prevent people from using math in cases where the English language is as short as the math…”

  5. yoshinorishiozawa
    November 3, 2022 at 4:11 pm

    I agree with Marshall on points (3) and (4). But I do not agree with him on point (1). It should be revised as

    (1′) Mathematics can be used as a short-hand language, but it is not all. It can be an engine of inquiry and research.

    Point (2) is ambiguous. What did he mean by “Stick to them” (or Keep to them)? Does this mean it is necessary to stick to mathematics until one arrives to understand the hidden structure under an apparent phenomenon? If he did, it is quite important. Probably Lars Syll and Jan Milch do not have an experience to have sticked to mathematics util they would discover a deep meaning under the confused appearance.

    I even doubt if Marshall was a good mathematician himself. See many figures in Part II of The Early Economic Writings of Alfred Marshall, 1967-1890 edited by J. K. Whitaker (McMillan for the Royal Economic Society, 1975). An economy is a large complex entity. Mathematics sometimes (although quite rarely) is useful in detecting a covered structure under complicated relations. Marshall does not show his talent for cutting Gordian knots. Please also note that Marshall was one of founding fathers of actual mainstream economics.

  6. yoshinorishiozawa
    November 6, 2022 at 9:59 am

    It would be important to know what is argued between two paragraphs that Lars Syll cited in the above post. Here is an abbreviated citation:

    Nothing reveals the aversion of the great majority of the present-day academic economists for systematic empirical inquiry more than the methodological devices that they employ to avoid or cut short the use of concrete factual information. Instead of constructing theoretical models capable of preserving the identity of hundreds, even thousands, of variables needed for the concrete description and analysis of a modern economy, they first of all resort to “aggregation.” The primary information, however detailed, is packaged in a relatively small number of bundles labeled “Capital,” “Labor,” “Raw Materials,” “Intermediate Goods,” “General Price Level,” and so on. These bundles are then usually fitted into a “model,” that is, a small system of equations describing the entire economy in terms of a small number of corresponding “aggregative” variables. The fitting, as a rule, is accomplished by means of “least squares” or another similar curve-fitting procedure.

    A typical example of a theoretical “production function” intended to describe …

    (illustration of a production function by an algebraic expression)

    Hence, while the labels attached to symbolic variables and parameters of the theoretical equations tend to suggest that they could be identified with those directly observable in the real world, any attempt to do so is bound to fail: the problem of “identification” of aggregative equations after they have been reduced-that is, transformed, as they often are-for purposes of the curve-fitting process, was raised many years ago but still has not found a satisfactory solution. …

    Leontief is not rejecting “constructing theoretical models capable of preserving the identity of hundreds, even thousands, of variables needed for the concrete description and analysis of a modern economy”. On the contrary, he recommends such research.

    Lars Syll ignores these details and assumes that Leontief concluded that “Insisting on formalistic-mathematical modeling forces the economist to give up on realism and real-world relevance.” How can a model, without mathematics, be capable of preserving the identity of hundreds or thousand of variables for the concrete description and analysis of a modern economy?

    Leontief proposed to change the mode of building mathematical models, not rejecting them. He criticized general equilibrium model (before Syll’s citation). He criticized the macoeconomic custom to use, think and argue by aggregated variables, in other words, he criticized the way of trying to understand an economy by small number of bundles labeled “Capital,” “Labor,” “Raw Materials,” “Intermediate Goods,” “General Price Level,” and so on (first paragraph in my citation). The worst example is the aggregate production function. See Felipe and McCombie 2013 The Aggregate Production Function and the Measurement of Technical Change / ‘Not Even Wrong’. Aggregate production functions are ‘not even wrong’.

    Simple literary studies are often insufficient. Mathematics is sometimes useful to find an underlying structure of the the large and complex system as the economy is. (Of course, there is no automatic method to achieve this.)

    An example is my own experience. I had a conjecture but could not solve it affirmatively. It was Masashi Morioka who solved the problem very beautifully by estimating the largest absolute value of eigenvalues of a matrix of a possibly very large dimension. With his result, we could present a new vision how the modern industrial economy works, not by price adjustment (which is at the core of mainstream economics), but by quantity adjustment. The latter has been assumed for long time by various heterodox economists including Post Keynesian economics but without no real examination of the process. The analysis was difficult, because it involved the adjustment of input-output relations for all goods that may count millions of differentiated products. Without mathematics, this kind of achievement must be impossible.

    Our research can change the research scene of Post Keynesian economics. Conventionally, it assumed that we can argue correctly (perhaps helped by our intuition) by assuming an economy that is composed of a single (or seldom of two or three) goods. Now we need no such ridiculous and absurd assumptions, because we now know that the large production system complicated with input-output relations can follow the slow movement of average final demands as a vector (on some reasonable conditions).

    Keynes made a similar arguments, for example, to explain his central theory, i.e. the principle of effective demand. At his time, there is no other method. See my paper 2021 in the Review of Keynesian Studies (open access).

    It is now no time to continue criticizing mainstream economics. It is time to supersede it. Reactive mode of research is now démodé.

    • rsm
      November 7, 2022 at 9:07 pm

      Weather models include thousands of inputs, yet how accurate are forecasts?


      Inches of lowland snow were predicted by today, but have I seen only rain?

      And in an economy where the most influential prices are unobservable because they are embedded in secret private contracts, how can you hope to get readings as accurate as the sensor inputs to weather models?

      • rsm
        November 10, 2022 at 10:41 pm

        yoshinorishi, can irrational psychological thoughts cause real economic things such as prices?

      • yoshinorishiozawa
        November 11, 2022 at 3:21 am

        Dear rms,
        you are misunderstanding how the prices are formed. Probably you are thinking that prices are determined by the equilibrium point of demand and supply functions, which is the core idea of neoclassical economics. If we base on this neoclassical theory (the standard and mainstream idea), it is necessary to consider how irrational psychological factors influence prices.

        If you are thinking of the prices of paintings of past famous painters, you may be right. However, prices of modern industrial products (which occupies absolute majority of GNP) are not determined in this way. They are regulated by the “full-cost principle”, i.e. the unit cost + profit margin (a markup rate times the unit cost). Psychology of consumers (if they are rational or irrational) is only relevant in deciding how much a product is sold.

        If you wonder how firms can produce their products as much as they are demanded under the constraints of complex input-output relations, pleas read for example Section III of my paper:
        The principle of effective demand: a new formulation, which is a free access paper.

    • yoshinorishiozawa
      November 8, 2022 at 10:17 am

      Dear RSM,

      increasing the number of variables does not ensure a more accurate prediction. There was a famous project called Project Link lead by L. Klein. It tried to link almost all countries predictive models into a single system. It is said that it counted more than thousand variables. But the accuracy did not increased.

      Economics is very far retarded than meteorology. It is quite hard to expect an accurate prediction even for the economy of a near future. Do you believe Milton Freedman? Prediction must not be the unique criterion for economics to be scientific. In my opinion, economics should pay much more efforts to clarify causal relations between events. This includes algebraic theory that cannot have any predictive powers. See the second paragraph of the first block citation of Hayek in my comment on complexity.

      A simple consequence is that constructing predictive models should have lesser importance than we pay to them actually.

      Economics is not a fortunetelling science.

      • November 11, 2022 at 2:12 am

        I agree with you, yoshshinorishi,

        Although satellite pictures and sample measurements are helpful for examining the next five or six days of weather, long term social prediction and control is more like a snake oil pipe dream promoted by gamblers casting around for someone to fleece.

        I see the complexity of a search for causal social relationships being a dynamic search for patterns in something like predicting the rainbow color patterns of a thin oil film spreading outward on water in sunlight. Large data sets measuring something like this are created for centralized rainbow chasing.

        Economics as a someday science examines as a student, not a doctor; that is one vector. I do not know how many vectors there are, yet there is one other to start a list, example; Helping coax emergency full-speed self-examination by society as a whole. This is similar to coaxing optimum growth of a healthful crop.

        Economic reasoning is now chained to finance. Even so, It has the capacity to shift to a more active and fun maximum creativity measure than money for a human society that heals Earth. That threshold is where economics becomes science. Then almost all humanity will learn to think and communicate as scientists about matters solvable by socially thought out logical cooperative human effort.

  7. rsm
    November 14, 2022 at 11:22 pm


    May I start off my price theory with Fischer Black, in “Noise”?

    “we might define an efficient market as one in which price is within a factor of 2 of value, i.e., the price is more than half of value and less than twice value. The factor of 2 is arbitrary, of course. Intuitively, though, it seems reasonable to me, in the light of sources of uncertainty about value and the strength of the forces tending to cause price to return to value. By this definition, I think almost all markets are efficient almost all of the time. “Almost all” means at least 90%.”

    In other words, how arbitrary is the “markup” variable in your formula?

    Wikipedia says:

    “Black argues that econometrics is filled with [noise] in the forms of unobservables and mis-measurement. No matter how many variables one puts into a model, there are always more to add but can’t (ones you can’t observe) and the ones you have will always have error. This is how noise manifests in econometrics (as well as poor interpretation of regressions, such as assuming correlation means causation).”

    Can I argue that the same problem extends to the physical sciences? What is the error margin on the rainbow predictions mentioned by Garrett Connelly? Does the weather prediction failure I personally experienced provide evidence that observable variables alone can fail to account for underlying inconsistencies in actual nature?

    • yoshinorishiozawa
      November 15, 2022 at 3:01 pm


      I understand now that we were talking of different things with the same name: price.

      I am talking of the prices of those products that are produced in a modern industry. You are talking about financial market, for example, the price of options that must obey a stochastic differential equation.

      Both two have the same name but their behavior is totally different and obey totally different laws.

      When you said that “the most influential prices are unobservable because they are embedded in secret private contracts” (in the post of
      November 7, 2022 at 9:07 pm Reply), I did not imagined that you were talking about financial markets. I am interested in financial economy (FIance, Real estate and Insurance), but it is my opinion that FIRE economy cannot be “the most influential prices” and must not be so. But, this is only my own opinion, I have no intention to persuade you to adopt the same opinion with me. You must have your own evaluation and opinion.

      That is agreed. Then, what do you want to say?

      I questioned if Lars Syll’s summery of Leontief was right. He resumed it as follows:

      the main reason for this irrelevance [of the mainstream economics] is the failure of economists to match their methods with their subject of study. The fixation on constructing models showing the certainty of logical entailment has been detrimental to the development of relevant and realist economics. Insisting on formalistic-mathematical modeling forces the economist to give up on realism and real-world relevance.

      Lars Syll may be right when he said that “the failure of economists to match their methods with their subject of study.” But, is it right to summarize “Insisting on formalistic-mathematical modeling forces the economist to give up on realism and real-world relevance”? I believe Leontief did not claimed such rough contention. He objected to the common use of aggregated variables such as Capital, Labor, Raw Materials, Intermediate Goods, General Price Level,” and others.

      There are infinite possibility for formalistic-mathematical models. We should not reject formalistic arguments and mathematical models, that must include simulation models. Although actual simulation models are in a very primitive state, I believe it has a possibility that gives us a new method of understanding how our economy works.

      See my paper: Guided tour of the backside of agent-based simulation.

  8. rsm
    November 15, 2022 at 11:47 pm


    In your simulation model, do you have “Fed” that produces the world’s most-demanded money at its whim? Is there an “OPEC” agent engineering gamma squeezes by buying long oil calls, thus driving up spot (which drives what one pays at the gas pump), giving them a free lunch?

  9. yoshinorishiozawa
    November 16, 2022 at 6:07 am

    Dear rms

    do you have “Fed” that produces the world’s most-demanded money at its whim? Is there an “OPEC” agent engineering gamma squeezes by buying long oil calls, thus driving up spot (which drives what one pays at the gas pump), giving them a free lunch?

    Of course, not! Don’t be surprised. We have not arrived to the stage to make good predictions by means of simulation models. In that point, I agree with Lars Syll. Economists had been too ambitious to imagine as if they were all mighty. They failed to “to match their methods with their subject of study.”

    We have worked for a simulation model named U-Mart. This is a simulation model of a futures market, in which both human and machine agents can participate in the same simulation model.

    See the book Realistic Simulation of Financial Markets: Analyzing Market Behaviors by the Third Mode of Science” 2016. See for more technical features of the U-MART this article.

    We are not thinking that simulation model can describe the behavior of the whole economy that may include “Fed” or “OPEC”. As I have mentioned in my above post, I am thinking that Project Link was a failure. We have no intention tor repeat such silliness. We must know the appropriate object of research so that a simulation model can effectively increase our understanding how it works.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: