Home > Uncategorized > Do unrealistic economic models explain real-world phenomena?

Do unrealistic economic models explain real-world phenomena?

from Lars Syll

When applying deductivist thinking to economics, neoclassical economists usually set up ‘as if’ models based on a set of tight axiomatic assumptions from which consistent and precise inferences are made. The beauty of this procedure is, of course, that if the axiomatic premises are true, the conclusions necessarily follow. idealization-in-cognitive-and-generative-linguistics-6-728The snag is that if the models are to be relevant, we also have to argue that their precision and rigour still holds when they are applied to real-world situations. They often don’t. When addressing real economies, the idealizations and abstractions necessary for the deductivist machinery to work simply don’t hold.

If the real world is fuzzy, vague and indeterminate, then why should our models build upon a desire to describe it as precise and predictable? The logic of idealization is a marvellous tool in mathematics and axiomatic-deductivist systems, but a poor guide for action in real-world systems, in which concepts and entities are without clear boundaries and continually interact and overlap.

As Hans Albert has it: 

Clearly, it is possible to interpret the ‘presuppositions’ of a theoretical system … not as hypotheses, but simply as limitations to the area of application of the system in question. Since a relationship to reality is usually ensured by the language used in economic statements, in this case the impression is generated that a content-laden statement about reality is being made, although the system is fully immunized and thus without content. In my view that is often a source of self-deception in pure economic thought …

200px-Hans_Albert_2005-2A further possibility for immunizing theories consists in simply leaving open the area of application of the constructed model so that it is impossible to refute it with counter examples. This of course is usually done without a complete knowledge of the fatal consequences of such methodological strategies for the usefulness of the theoretical conception in question, but with the view that this is a characteristic of especially highly developed economic procedures: the thinking in models, which, however, among those theoreticians who cultivate neoclassical thought, in essence amounts to a new form of Platonism.

Seen from a deductive-nomological perspective, typical economic models (M) usually consist of a theory (T) – a set of more or less general (typically universal) law-like hypotheses (H) – and a set of (typically spatio-temporal) auxiliary assumptions (A). The auxiliary assumptions give ‘boundary’ descriptions such that it is possible to deduce logically (meeting the standard of validity) a conclusion (explanandum) from the premises T & A. Using this kind of model economists are (portrayed as) trying to explain (predict) facts by subsuming them under T, given A.

An obvious problem with the formal-logical requirements of what counts as H is the often severely restricted reach of the ‘law.’ In the worst case, it may not be applicable to any real, empirical, relevant situation at all. And if A is not true, then M doesn’t really explain (although it may predict) at all. Deductive arguments should be sound – valid and with true premises – so that we are assured of having true conclusions. Constructing, e.g., models assuming ‘rational’ expectations, says nothing of situations where expectations are ‘non-rational.’

Economic theories and models have to be compared to the situations they are supposed to represent/explain/predict. There is no way of getting around questions of realism and real-world relevance. Building theories and models that are ‘true’ in their own very limited ‘idealized’ domain is of limited value if we can’t supply bridges to the real world. Economic ‘laws’ that only apply in specific ‘idealized’ circumstances —  in ‘nomological machines’ — are not the stuff that real science is built of. Results derived in mainstream economic models

Results deduced in a ‘closed world’ mainstream economic model is obtained only because the model (machine) was built for that purpose. Outside the machine — in the real world — we know that most of the assumptions, including the typical ceteris paribus condition, do not apply.

Most mainstream economic models are abstract, unrealistic and presenting mostly non-testable hypotheses. How then are they supposed to tell us anything about the world we live in?

When confronted with the massive empirical refutations of almost every theory and model they have set up, mainstream economists usually react by saying that these refutations only hit A (the Lakatosian ‘protective belt’), and that by ‘successive approximations’ it is possible to make the theories and models less abstract and more realistic, and – eventually — more readily testable and predictably accurate. Even if T & A1 doesn’t have much of empirical content, if by successive approximation we reach, say, T & A25, we are to believe that we can finally reach robust and true predictions and explanations.

There are grave problems with this modelling view. What Hans Albert most forcefully is arguing with his ‘Model Platonism’ critique of mainstream economics, is that there is a strong tendency for modellers to use the method of successive approximations as a kind of ‘immunization,’ taking for granted that there can never be any faults with the theory. Explanatory and predictive failures hinge solely on the auxiliary assumptions. That the kind of theories and models used by mainstream economics should all be held non-defeasibly corroborated, seems, however — to say the least — rather unwarranted.

Confronted with the massive empirical failures of their models and theories, mainstream economists often retreat into looking upon their models and theories as some kind of ‘conceptual exploration,’ and give up any hopes/pretences whatsoever of relating their theories and models to the real world. Instead of trying to bridge the gap between models and the world, one decides to look the other way.

To me, this kind of scientific defeatism is equivalent to surrendering our search for understanding the world we live in. It can’t be enough to prove or deduce things in a model world. If theories and models do not directly or indirectly tell us anything of the world we live in – then why should we waste any of our precious time on them?

  1. October 26, 2017 at 3:03 am

    This is how things started in the physical sciences. Statics in mechanics for example preceeded dynamic analysis and still does in teaching mechanics. Also linear models of dynamic systems preceeded non-linear dynamic modeling by centuries. Linearity guarantees uniqueness of solutions so if one can find one solution you’re done looking by uniqueness. What is the implication of multiple solutions to models of real world problems? One idea is that multiple solutions may be possible in reality but only one is stable to infinitesimal perturbation so one needs to find the one solution most stable and so the one to which other solutions will evolve to. The teaching example is the inverted pendulum versus non inverted pendulum. Both solutions satisfy the steady equilibrium configuration per the equations of motion for all time. However the inverted pendulum, if perturbed, will leave this configuration and evolve toward a non inverted condition that is stable to perturbation. With nonlinearity comes also the possibility of chaos whereby many solutions exist that are stable but each depends on infinitesimal variations in the initial data. Thus a butterfly flaps a wing which induces a hurricane to evolve. As Steve Keen points most existing economists have not paid their dues studying any of this despite great practical strides enabled by use of these tools in
    Math Physics. This learning/knowledge deficit is so wide and deep that bridging it seems hopeless. It’s like contemplating bringing an aged and ignorant rice farmer up to speed in using perturbation method to linearize the wave equation as the first step toward understanding and applying quantum mechanics. The phrase: Pack a lunch, comes to mind.

  2. October 26, 2017 at 6:42 am

    It is true that we should defeat “scientific defeatism.” To avoid falling in it is to change our whole scientific research program (SRP). We have to discover a new strategy in re-building (or restructuring) economics into a better science. This must be a change from metaphysical stage of science to an experimental. stage. (“Experimental” simply means here “truly empirical.” It does not imply the meaning that is understood in experimental economics.)

    My proposal is to abandon Walrasian SRP. Economy (at least modern economy) is an extremely complex entity and it is impossible to understand its total mechanism in a stroke. It is much better to return to Marshalian SRP. In other words, let us try to discover fields and phenomena (even if they are very narrow and restricted) and establish a firm relation (or relations) between small numbers of variables. This is the way that modern science started in its starting period, i.e. at the time of Galileo and Kepler.

    See my comment on Robert Delorme’s keynote paper to the WEA conference ECONOMIC PHILOSOPHY: COMPLEXITIES IN ECONOMICS

    Peterblogdanovich (who?) does not correctly present the history of physical science. Dynamics (Newton) had not evolved from statics (Archimedes). It needed a scientific revolution. Simply put, modern physical science started from Galileo and Kepler. They searched firm established relations, either by experiments or by observations. Kepler struggled with 8-minute errors. Do any of fundamental premises (axioms or assumptions) of economics have such qualities either in constancy, exactness, observability, or repeatability? Galileo and Kepler were well conscious that their theories or models hold to a very specific (and narrow) situation or object and did not pretend that they have discovered a universal law. They knew the range of validity of their theories.

    Mainstream economics has forgotten that scientific statements normally have a range of validity. Modern physics arrived at some of universal laws only at the time of Newton. Even at that time, Newton knew practically nothing about electromagnetic mechanics. His theory was not a theory of everything. Mainstream economists pretend that they have a theory of everything economic. We economists must be more humble.

    Economics has discovered no such universal laws. And yet, mainstream economists pretend that a combination of quite dubious assumptions can produce firm knowledge on the whole economy. A reasonable person understands that a combination of blurred propositions multiplies blurredness of the results. Many economists think in opposition that they can arrive at a good prediction by combining such blurred laws. Such a “science” is simply a metaphysical discipline which is only useful in screening young bureaucrats in business and public administrations.

  3. Nick
    October 26, 2017 at 8:23 am

    i think the problem is viewing the distortions needed to save the theory – (and not the phenomena – deductivist-nomothetic, see Descartes and the Geometric method of axiomatic reasoning (Euclid) and Hume – nothing in the mind that was not before in the senses and Kant’s ‘copernican revolution’ beyond this impasse of Rationalism(Dogmatism) vs Empiricism(Scepticism) and the failure of transcendentalism….)…as an idealist account of intellectual history…free from the blood guts and sweat of real history. You will not ‘defeat’ neolib divorce from reality by arguments or meta-critiques of methods…it is not about ‘may the best model win’ – its about maintaining the status-quo in terms of money and power…

  4. Frank Salter
    October 26, 2017 at 9:49 am

    I am in total agreement with the sentiments expressed by Lars Syll, Peterblogdanovich and Yoshinori Shiozawa.

    However, Lars Syll states, ‘The logic of idealization is a marvellous tool in mathematics and axiomatic-deductivist systems, but a poor guide for action in real-world systems, in which concepts and entities are without clear boundaries and continually interact and overlap’. I believe this to be accepted by economists as a universal truth, but, in fact, it is not. The underlying axioms must be those of the real world. These are the laws of physics. Complexity comes from these laws. Just look around. Complexity is everywhere, but it is merely solutions to the underlying physical relationships. It is necessary, therefore, to start from known physical relationships and minimise the number of further assumptions – Occam’s razor! The major problem in economics is that entities do proliferate. Mathematical models with inappropriate assumptions are hypothesised merely to conform to the perceived shape of reality — entities proliferate!

    Peterblogdanovich states: ‘As Steve Keen points most existing economists have not paid their dues studying any of this despite great practical strides enabled by use of these tools in Math Physics. This learning/knowledge deficit is so wide and deep that bridging it seems hopeless.’ This summarises an essential problem. More time is required to study the fundamentals of science and engineering than could be incorporated into an economics degree course. It is difficult to see how this may be resolved.

    Yoshinori Shiozawa points out that: ‘Economics has discovered no such universal laws. And yet, mainstream economists pretend that a combination of quite dubious assumptions can produce firm knowledge on the whole economy.’ This is indubitably true, but some aspects of the economy can provide firm conclusions.

    It is only some three weeks since RWER81 was published. Salter (2017) is based only upon rigorous physical principles. It describes manufacturing industries. It is very likely to provide extremely difficult reading, but it does demonstrate that, from a single definition of productivity, what might be described as production theory follows — from the solution of both differential and algebraic equations. I contend that it satisfies the requirements of Occam’s razor. It, therefore, undermines most of the ideas of neoclassical analysis.

    Salter, Frank M. (2017). “Transient Development”. In: Real World Economic Review (81), pp. 135–167

  5. October 26, 2017 at 12:14 pm

    It seems to me that one reasonable induction one can make about both physics and economics is that theories are never complete: there are always neglected factors that could – under possible future circumstances – become significant for applications. Moreover, these factors are often overlooked or even denied until it is too late.

    Thus it seems to me that empirical theories are always contingent on unknown factors, hence one form of radical uncertainty. One can try to develop better theories. Indeed I have found it helpful to seek potential factors very widely, and then to analyse the extended theories to identify potential factors, to be monitored for signs of an insipient crisis. But one still has an irreducible possibility of being taken by surprise, hence the importance of general ‘horizon scanning’.

    For example, from game theory, the rise of Asia and the demise of the US would could lead to financial and instabilities. Similarly, control theory would suggest that rising inequality within nations could lead to instability. These kind of deductions seem to me valid, and suggest scenarios to inform horizon scanning and policy..

    To me, then, theories are far from useless: it is just that we make mistakes if we rely on them too much. (As Peterblogdanovich says, if economics wished to be a science like physics, there are plenty of mathematical models to draw on, as a corrective to the simplistic mainstream.)

    • Frank Salter
      October 26, 2017 at 2:59 pm

      What you say is essentially true. However, there is danger in the over interpretation of evidence. The scientific method is build upon iterative procedures. Mechanisms are examined to see if they provide appropriate descriptions of reality. They are accepted until further evidence reveals their failings.
      For over 1500 years, the Ptolemaic system provided excellent descriptions of planetary positions and eclipses — the latter embodied in the Antikythera mechanism. The earth-centric description was wrong but provided accurate predictions because the mathematics used was good enough for practical use. This theory was replaced by the Copernican heliocentric analysis, but Ptolmaic calculations still continue to provide the correct positions of planetary bodies. The calculations have not been invalidated.
      Newtonian mechanics are refined but not replaced by relativity. Relativity is not introduced into calculation of the orbits of space vehicles moving in the solar system. It is not necessary, but relativity is an essential fact for global positioning.
      The scientific method relies upon predicting and comparing the predictions with empirical fact. When predictions fail to correspond to reality, the theory is replaced or modified in the most appropriate manner. The process continues with further iterations. Refined analyses are compared with empirical evidence. There is no limit to the possible number of iterations.
      You state ‘we make mistakes if we rely on them too much’. With this I totally disagree! Any failure of a theory requires refinement or correction. Your continuation about mathematical models is a non sequitur. Mathematical models are not, of themselves, theories, they are applications of some theory. Their failings should trigger new thinking, not abandonment.

      • Frank Salter
        October 27, 2017 at 6:21 pm

        For Peterblogdanovich:
        Time continues in its normal relentless manner, no matter what the interest rate is. Negative rates imply, merely, that the flow has changed direction.

  6. Risk Analyst
    October 26, 2017 at 6:27 pm

    The neoclassicals recognize their models do not work and do adjust for their lack of realism. For example, at the Federal Reserve (at least when I knew what was going on there) they had this thing called add-factors and they will apply them liberally to generate the results they wanted. And even with the estimation of the parameters, they would “beat the data” until it confesses to the result they want. And in the private sector with high profile economist predictions at various banks, they will “stein estimate” their estimates using an average of their estimate and market expectations for their reported number.

    In addition, economic models can not be used as in physics. They change with the changing human behavior. If you don’t feel comfortable putting on 1970s bell bottom jeans and a skin tight big collared shirt and vest to hit the mall any more, why would you think economic behavior does not change. If one wants to forecast the interest rate yield curve one year in the future, not much in the current data is of help because this is a unique time (low rates, thin credit spreads, high debt, tax cuts, unknown potential administration policies, hostile politics, and so on).

  7. October 27, 2017 at 4:42 am

    Frank Salter is arguing like “add-factors” analysts that Risk Analyst cited as the most conventional method used in the “iterative procedure”.

    The actual problem of economics is not how to obtain gradual improvement of predictions or fitness. As I have argued elsewhere (see two of my “replies” cited the reference below), today’s economics is to be seen as Ptolemaic system. It can improve its fitness by adding factors or modifying their models. Ptolemaic system also did this: epicycles, eccentric deferent, equant point, and their combinations.

    The trouble with Ptolemaic system was not that it lacked accuracy. It was sufficiently exact for calendar making, for example. As I have written in my “reply” of September 22, “[d]espite its accuracy, Ptolemaic system was a wrong system. Further development of astronomy depended on the overthrow of the system.” In the same vein, economics requires its Copernican revolution: a complete paradigm shift from neoclassical economics.

    We have to restart our economics. How to achieve it is not yet visible. Salter’s challenge might be one of possibilities, but I do not believe that his physicalist approach will solve our problem. He is (intentionally?) ignoring the arguments presented by Shaik (1974) and Simon (1979).

    My proposal is to abandon demand and supply equilibrium framework (its supreme theory is Arrow and Debreu’s model) and retreat to a more primitive classical theory of value (Ricardian cost-of-production theory of value. See Shiozawa 2016). This theory is developed into a new theory of international values (Shiozawa 2017). The latter is for the moment the unique general theory that incorporates trade of input goods (intermediate goods). If you admit the emergence and importance of global supply chains (or global value chains if seen from value side), this new theory now supersedes four generations of neoclassical trade theory, because all these theories are excluding trade in input goods. They are (1) textbook Ricardian trade theory (à la J.S. Mil-Viner-Jones), (2) Heckscher-Ohlin-Samuelson model, (3) the New trade theory (à la Krugman) and (4) the New New trade theory (à la Melits).

    The third “reply” to Lars Syll’s September 16 post: Where modern macroeconomics went wrong
    See also the sixth “reply” to the same post.

    The eighth “reply” (my second post) to Lars Syll’s August 3 post: Why should we care about Sonnenschein-Mantel-Debreu?

    Shaik, A. 1974 Laws of production and laws of algebra: the humbug production function. Review of Economics and Statistics, 56: 115-120.

    Shiozawa, Y. 2016 The revival of classical theory of values. In Yokokawa et al. (eds.) The Rejuvenation of Political Economy. Routledge.

    Shiozawa, Y. 2017 The new theory of international values: an overview. In Shiozawa et al. (eds.) A New Construction of Ricardian Theory of International Values, Springer Nature.

    Simon H.A. 1979, On parsimonious explanations of production relations’, Scandinavian Journal of Economics, 81: 459-474.

    • Frank Salter
      October 27, 2017 at 8:04 am

      I appear to have explained my views with insufficient clarity. I am not arguing for more terms in curve fitting procedures, quite the opposite. I am in total agreement with your analysis about the failure of conventional economic analysis.
      You state: ‘We have to restart our economics. How to achieve it is not yet visible.’ This is no longer true. I was attempting to explain why.
      In my RWER81 paper, there is no mention of supply and demand. It is irrelevant! My first principles analysis, based on the definition of productivity, implies a ‘cost-of-production theory of value’. It arises quite naturally from the mathematics. On page 162, I prove that, for reasons of dimensional analysis, the mathematical statement of the neoclassical production function, is true only if capital is labour-time, which may be a shock to many neo-classicists.
      My analysis of manufacturing production theory introduces time with physical rigour. It explains why Kaldor’s stylised facts are essentially true. Its predictions are confirmed by empirical data. It makes a totally new prediction, previously unnoticed, about the value of the Verdoorn coefficient in manufacturing industries. As an example of the scientific method in action, these statements imply that the analysis is valid. It should therefore be tested against other theories. I would categorise it as transient classical theory.
      I contend that it meets the test of Occam’s razor. In scientific terms that would be sufficient for it to be seen as a demonstration of a valid theory. On this, I would appreciate all of your views.

      • October 27, 2017 at 4:57 pm

        This is important. Economists have long neglected the power of dimensional analysis; the constraint that any model of any real world phenomenon must be dimensionally homogeneous. I offer two examples:
        First, the chartalists have an aphorism that Economics is the science of confusing stocks with flows. The joke is an observation that theorists constantly blend and interchange quantities with units of $ (stocks) with quantities with units of $/time. It’s hilarious because it’s so true.
        Second, the natural dimensionless scaling/grouping for time is interest rate. This has profound implications. What are the implications of this as rates go to zero or negative? Dimensionless time stands still, or goes backwards? In the physical sciences dimensional analysis reduces the number of variables and provides simplification by scaling arguments. We don’t typically have models where physical parameters vanish or turn negative and when we do, eg the Lorentz space-time dilation as velocity approaches the speed of light, it’s a real head scratcher.

      • Frank Salter
        October 27, 2017 at 6:23 pm

        For Peterblogdanovich:
        Time continues in its normal relentless manner, no matter what the interest rate is. Negative rates imply, merely, that the flow has changed direction.

        Sorry about the earlier misplacement

      • October 27, 2017 at 6:50 pm

        Simple linear physical systems can run backward too. Our solar system is one example, where planets going backward breaks no rules and looks reasonable. Irreversibility however does creep into physical models through nonlinearity and the second law of thermodynamics. The scaled time variable cannot simply be allowed to run backward. For example once mixed there is no reverse fluid motion that will unmix cream from your coffee. Economics will be showing signs of progress when the same is true for models there. Steve Keen is the only guy I see working on this too.

  8. October 27, 2017 at 10:14 am

    Why do economists have difficulty recognizing, per the maxim “the elephant in the room?” The historian Thomas Hughes describes technology “as a creative process involving human ingenuity.’’ In my view, Hughes’ insight applies to more than just technology. It applies to everything we generally refer to as human culture, human society. Including economics. Economies, economic actions, economic institutions are the result of human imagination. If we want to build an economics discipline, study this!

  9. October 28, 2017 at 2:02 am

    Re Frank Salter’s reply to my post October 27, 2017 at 4:42.

    “Abandon demand and supply equilibrium framework” does not imply that we can (re-)construct economics without using demand and supply concepts. You can use them but you have to change the viewpoint on how economic system works. Most of economists, including heterodox economists, are still a victim of rationality bias. They believe that the system as big as the world of more than 7 billion people is regulated by rational power of human agents to dealing with information. We have to escape from this rationality bias.

    See my comments on Beata Stępień’s short paper in ResearchGate on methodology. I have posted four comments. They are not very visible and you have to try to find them.


    • Frank Salter
      October 28, 2017 at 8:35 am

      Supply and demand are useful parameters in applying laws of conservation. However, with only manufacturing industries demonstrating unconditional convergence (which must not to confused with markets that clear) it is difficult to perceive equation forms which might be deployed, in order to further our understanding.

      In developing any theory, it is necessary to find the significant abstractions which are relevant and necessary. It is easy to add many unnecessary parameters and curve fit, which appears to be the methodology being applied. (As we have discussed, Ptolemaic geocentric theory provides accurate predictions but can be seen to be totally wrong. This is a counterexample which proves that curve fitting demonstrates correlation but not causation. Standard methodology clearly confuses the two.) This reiterates your original comments, with which I agree totally. In your initial comments you continue, ‘… and establish a firm relation (or relations) between small numbers of variables.’ My paper does exactly this and provides relationships which allow alternative policies to be investigated, but it also demonstrates that the best courses of action are already being followed in the engineering of competitive manufacturing industries. The empirical data confirms all its predictions. However, this is merely the start of rewriting economic theory. There is still a long way to go.

  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.