Home > Uncategorized > Post-Keynesian Response

Post-Keynesian Response

from Asad Zaman

This continues from previous post on New Directions in Macroeconomics. Among the heterodox responses to the crisis in economic theory created by the Global Financial Crisis 2007, we will briefly discuss the following:  Post-Keynesian Economics, Modern Monetary Theory, Political Economy, Evolution of Global Finance, Ecological Economics, Complexity Economics, Islamic Economics. This post is about Post-Keynesian Economics.

The response of mainstream macroeconomists to this crisis has been disappointing; see for example Antara Haldar (2018) “Economics: The Discipline that refuses to change”. The failure of classical economics in the Great Depression of 1929 led Keynes to the create the field of macroeconomics, which was revolutionary many different ways. Unfortunately, as Romer remarks, the profession went backwards, losing hard-won insights. All of the revolutionary Keynesian insights (discussed in greater detail below) have since been rejected by the orthodoxy.  Similarly, there has been little or no response to the demonstrated failure of macroeconomic models following the Global Financial Crisis. read more

  1. Yoshinori Shiozawa
    September 5, 2020 at 3:32 am

    Post Keynesian economics is full of good insight, but unfortunately lacks theoretical / microeoconomic foundations. Consequently, it remains as a set of disparate theories based on observed or intuitive tendencies. This state of the art produces different strands such as Fundamentalist PK, Kaleckian, Minskian and Sraffian economics, each of which believes that others are no authentic successor of Keynes. It is required that more arguments on theoretical and unifying foundations of Post Keynesian economics. We should go beyond mere methodological / philosophical arguments. Post Keynesian economics needs a theoretical core that can compete with neoclassical economics.

    Please join the discussion page in the ResearchGate:
    Does Post Keynesian Economics need no theoretical foundations?

    See also: Marc Lavoie’s book review on our attempt.

    • September 5, 2020 at 7:01 am

      I’ve repeatedly pointed out that “the state of the art”s theoretical core based on the Newtonian Maxwell/Grassman theory of unrestrained radio power transmission has bemused itself by not taking account of the directional constraints of telephony and Shannon’s definitions of information and noise. It seems Hayek’s speculations on information (1946) preceded Shannon (1948), but Keynes wrote after Newton and his theoretical core – however intuitively expressed – was consistent with both the PID mathematical distinctions – proportional, integral differential – in Newtonian mathematics, and the error-correcting rather than knowledge-seeking purpose of information science, as in the cybernetic model of eliminating error.

      Economics professors like Yoshinori, wishing to advance the “state of the art”, need to spend some time studying how DC telephony, AM, FM and Digital radio work; how negative feedback is able to reduce noise (familiar from Dolby hi-fi) and how positive feedback (reaction) first strips the information from any signal it is applied to, and more strongly applied, generates first oscillation then noise (a pre-war practical demonstration of complexity theory).

      What I can offer as new to me are the proposals of Agent-based Computational Economics (ACE),. “Roughly defined, ACE is the computational modeling of economic processes (including whole economies) as open-ended dynamic systems of interacting agents. The stress on open-ended dynamics fundamentally distinguishes ACE from all other currently mainstream economic modeling approaches”. This might be about Shannon’s theory:

      “One key issue for ACE normative design is the extent to which resulting outcomes are efficient, fair, and orderly, despite possible attempts by strategic decision-making agents to game the design for personal advantage. A second key issue is a cautionary concern for adverse unintended consequences. Optimal design might not always be a realistic goal, especially for large complex systems; but ACE models can facilitate robust design for increased system resiliency, a goal that is both feasible and highly desirable”. So Shannon’s ‘redundancy’ (as in duplicate equipment) is used to increase reliability in space travel,.


    • Yoshinori Shiozawa
      September 6, 2020 at 3:58 am

      Erratum: Yoshinori Shiozawa on September 5, 2020 at 3:32 am

      It is required that more arguments on theoretical and unifying foundations of Post Keynesian economics. >> It is required that more arguments on theoretical and unifying foundations of Post Keynesian economics are made.

  2. Yoshinori Shiozawa
    September 5, 2020 at 2:28 pm

    ACE (Agent-based Computational Economics) is an attempt to understand economic processes using multi-agent simulation. The latter is a powerful tool in pursuing what happens in economy step by step. So, it is one of indispensable tools of analysis for non-equilibrium economics. However, there are many models based on simulation such as Computatble General Equilibrium models that are simple repetition of mainstream economics. It is dangerous to judge whether an economic theory or subdiscipline is good or not, or neoclassical or not. It is necessary to judge by its contents.

    Please read my paper: A Guided Tour of the Backside of Agent-Based Simulation, Chapter 1, pp.3-50 in Kita, Taniguchi and Nakajima (eds.) Realistic Simulation of Financial Markets: Analyzing Market Behaviors by the Third Mode of Science, Springer-Japan, Tokyo, 2016. The subtitle of the book means that we are considering computer simulations as the third mode of science after theory (classic Greece) and experiments (observations) (Renaissance). It is my belief that a good synthesis of three modes of science is necessary in order to succeed in attacking complex entity as the economy.

  3. Yoshinori Shiozawa
    September 7, 2020 at 1:53 pm

    Stability-instability problems in economics are much more complicated than the standard cybernetics problems, because the question is the stability-instability of high-dimensioned vectors.

    The main problem that we (Shiozawa, Taniguchi and Morioka) started to work with was just the stability-instability problem of quantity adjustment process of an economy as a whole. In a 1983 paper, I formulated the problem in detail such that the process contains no bootstrappings; in other words, any agent in the process does not use the information before it is decided by some others. I assumed that each firm decides the production quantities of its products one period earlier before the demands for products are known. I also assumed that the firm assumes that the demand of the product obeys a simple tendency law (that can be obtained by extending the known movement by some rules of prediction). The whole process is described by a very large matrix that has (M+3)N-dimension, where M is the number of periods we use as known time series of past demands and N is the number of products that exist in the economy. N is normally extremely large. In a modern economy, it goes up to 10 to 100 millions. This complication only represents the input-output interactions between producers. Post Keynesian economics such as Fundamentalist Post Keynesian economics and Kaleckian economics both neglect these interactions and simply assume this process proceeds without any hindrances.

    My result was not new in the sense that it was a part of well known result in the name of Dual Stability-Instability problems in the Leontief system. See Jorgenson (1960) Dual Stability Theorem, Econometrica 28: 892-899. Needless to say, this result was not only destructive to Post Keynesian economics, but all mainstream macroeconomics based on one-good assumptions. Before asking rational expectation’s validity, in a many-good economy, there was no assurance that coordination of many-good economy works. The last resort of mainstream economics was Arrow-Debreu (1954) Existence of a Competitive Equilibrium, Econometrica 22: 265-290. As all sincere economists know, it was based on many unrealistic assumptions such as unbounded rationality, infinite sight (or information collection) and simultaneous adjustment of all variables (general equilibrium). Introduction of bounded rationality and others did not change the situation essentially, because those awkward modifications of basic assumptions could not produce the theory how modern industrial economy works as a whole. For this crucial point, it was necessary to appeal to Arrow-Debreu parable. But, of course, it was a sheer consolation for mainstream economists. Nothing was solved by neglecting and forgetting crucial defects, although this is the usual custom for them. Capital controversy was only one of them.

    Several years later, a light came by the computer simulation by Taniguchi (1991). Taniguchi demonstrated that, if firms take (moving) average of past demands, the whole system converges to a constant quantity system when the demand vector stays constant. Soon later, it was Morioka who proved that Taniguchi’s result can be mathematically proved in a linear case. This was really an astonishing result, because he has to prove all eigenvalues of a large matrix of dimension (M+3)N have the absolute value smaller than 1. Morioka had finally solved dual stability-instability problem by showing that short-term quantity adjustment process converges or can follow the slow change of demand flows. This does not mean that Taniguch’s results were all absorbed in the Morioka’s theorem, because Morioka method only applies to linear process when the firms do not face any stock-out situation for their products or for their input materials. Taniguchi’s result covers this difficult situation, although by numerical experiments we can obtain no definite propositions.

    Our book (2019) Microfoundations of Evolutionary Economics contains of course many other points of arguments but the main contention is Taniguch-Morioka’s results. (See also Marc Lavoie’s book review, although there is no mention on the theory of international values). A substantial development since 1991 was the discovery of the theory of international values. By this development, the classical theory of value in the tradition of cost-of-production theory of value became a theory that can treats international trade situation in general forms. Arrow-Debreu theorem can be applied to international trade situation, but, as it has no tool to analyse corner solutions, the price theory that was developed in our book is not only far realistic but also covers far general situation. There are many international trade theories (they are called international microeconomics) but no one is a general theory of input trade. By this reason, mainstream economics could not present a theory of Gloval Value Chains (GVCs). This conspicuous phenomenon in the globalized economy has been studied until now only by historians, sociologists and geographers, and not by economists. The new theory of international values offers a possibility to study GVCs from economics point of view. If we succeed, we can really surpass mainstream economics in this important actual problem.

    Another implication of our book is that it is totally free from marginalist arguments, including marginal theory of distribution and aggregate production functions. The concept of “capital” remains as an ever-changing form of a money fund that takes forms of capital money, capital goods, materials and wages in production process, products in the sales points (or in distribution network) and finally an amount of money normally increased than before the cycle, as it is observed in Das Kapital by Marx. Of course, there is no need to argue in terms of labor value, because we have a good price theory with which we can explain how the capital increases it value in a normal state of the economy.

  4. Gerald Holtham
    September 7, 2020 at 3:38 pm

    I am an admirer of the work of Y.Shiozawa and his associates. I think it is a useful model of the industrial economy. I worry, though that it deals with a shrinking part of the economy as a whole. Service industries, which have risen in importance relative to production industries, do not have inventories. Their equivalent adjustment mechanism is queuing. My impression is that queuing leads to price variation quite quickly in a way that inventory variations need not. People in a queue try to jump it by bidding more. Doesn’t that sever in the short run the association between price and cost of production? Many services also rely on skills and implicit knowledge embodied in people. These are not easily standardised. So a great footballer gets paid thousands of times what an average one receives. A talented hairdresser likewise. There is no uniform cost of production in those circumstances.
    We know that agricultural industries face diminishing returns as the most productive land is used up so the cost of production depends on the level of demand. it seems demand also determines price in many service sectors. That means a classical theory of value, which depends on constant returns to scale, is not sufficient to explain economic value in those sectors. A theory of demand is also required.
    This is not to criticize what YS and colleagues have done. It is a valuable building block but it cannot be the whole story.

    • Yoshinori Shiozawa
      September 7, 2020 at 5:23 pm

      I agree with Gerald. The economy which we have treated with is “a shrinking part of the economy as a whole.” We do not claim that our theory covers all of economic problems. As I have noted, finance economy cannot be dealt with by our theory. Question of how wage rates are determined is still an open problem. This is the reason why I emphasize it is the core of reconstructed economics.

      In Chapter 2 of our book, I have intentionally adopted axiomatic method by making clear various postulates I assumed to deduce main theorems. (I know of course there are many people who denounce axiomatic construction of neoclassical economics, which is wrong not because it employs the axiomatic method or mathematical formulations, but it employs wrong formulations such as demand and supply functions and equilibrium.) If, for an industry, one of Postulates that are used in the proof is not valid, it means that (1) the industry is out of the range of validity, and (2) we need to build a new theory that may comprise the industry.

  5. Gerald Holtham
    September 7, 2020 at 3:47 pm

    PS I do not understand the relevance of “by not taking account of the directional constraints of telephony and Shannon’s definitions of information and noise.” Dave, can you spell out what that means? What are those constraints and how should we incorporate them? Noise-signal extraction problems are a common concern in some branches of economics, notably finance. Shiozawa and colleagues assume firms use moving average procedures in order not to be misled by random fluctuations in customer demand. In the real world firms do use fairly simple procedures like exponential smoothing so I can’t see what it is you want them to do. This is a sincere question so do me a favour and please don’t mention either Hume or Kant in your reply. Remember I’m simple soul who likes addressing one question at a time.

  6. September 10, 2020 at 9:05 am

    Apologies for this response being delayed. That had the advantage I had to re-read what Yoshinori said above on 7 Sep and your very pertinent comment about the difference between queuing and inventories. (Reduced to mathematics, that is the difference between continuous and counting numbers which I had to work with in the earliest days of digital computing).

    Apologies for my using philosopher’s names as a short-hand for what their arguments were about, as I had to when responding to A J Sutter 7 Sep at 5.18 am on the “Capital Controversies” thread. Put simply, that argument was about a can of beans. The one guy said we can’t know what’s in it: the best we can do is agree about what it says on the label. The other asks, how can you do that, if the word ‘beans’ isn’t in your vocabulary? My argument as to what you can know is the scientific one. The can of beans can be anywhere; if there is anything in it, the weight of it (Newton’s gravity) will exert a force on the table; knock it off and it will fall down accelerating. You can open it to see why (Baconian science), but if you empty it too much it will first float and eventually fly away (into chaos theory).

    When Yoshinori mentions high-dimensioned vectors and eigenvalues he is in my short-hand talking Grassman. The simple example of a high-dimensioned vector is a table of mileages between towns. The intersection of the row for one town with the column for another gives the mileage between them, but this is zero where row and column point to the same town. That is indeed stable, but it doesn’t tell us what happens in the towns, or even which direction to go to get from one to another. The method is no more informative if one substitutes commodities (or even industries) for towns. Sorry, YS.

    So what do I mean “by not taking account of the directional constraints of telephony and Shannon’s definitions of information and noise”? That broadcasting is directionally continuous, telephony involves queuing. That Shannon’s “information” is not a continuously variable meaning but a measure of “how digital”: how many there are in the queue. That Shannon’s “noise” is not transmitted but received, environmental noise having added its pennyworth to the message during its transmission. Not surprisingly, received noise has the same “entropy” measure; what is surprising (and useful) is that the information content of its form is as great as one can get, so that maximum transmission efficiency is found when multiple messages or cypher keys are so mixed up as to appear random, usually leaving spare capacity which can be used to provide parity checks (like recognising error when adding an odd and an even results in an odd). As a minimum, the constraints on telephony include one having to make a connection, and even electric lighting or broadcasting has to be switched on. (In economic terms, we need to be provided with sufficient income)! Directionally, distance and habit constrain which way we go to and home from work, shops or bank.

    • September 10, 2020 at 4:50 pm

      PS. Where I wrote about emptying the tin of beans, it would have been better to have said “try to get too many beans out of it”. The continuous (radio amplifier or Brian Arthur) version of this involves positive feedback.

      I see the pseudo-scientific Yoshinori is still rabbiting on trying to sell his book when defending small world modelling of a large world. Obviously he has not understood Edward Fulton’s book on valuing Market-Values as fractions of an evolving whole rather than more than the sum of incommensurable parts. Is not the fractional part of a decimal number format a small-scale inverse (Copernician) model of the whole number part?

  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 )

Google photo

You are commenting using your Google 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.