On testing and learning in a non-repetitive world

28 Feb, 2018 at 18:05 | Posted in Economics | 8 Comments

marquesThe incorporation of new information makes sense only if the future is to be similar to the past. Any kind of empirical test, whatever form it adopts, will not make sense, however, if the world is uncertain because in such a world induction does not work. Past experience is not a useful guide to guess the future in these conditions (it only serves when the future, somehow, is already implicit in the present) … I believe the only way to use past experience is to assume that the world is repetitive. In a non-repetitive world in which relevant novelties unexpectedly arise testing is irrelevant …

These considerations are applicable to decisions in conditions of radical uncertainty. If the actions that I undertake in t0 will have very different consequences according to the eventual state of the world in t1, it is crucial to gather reliable knowledge about these states. But how could I evaluate in t0 my beliefs about the state of the world in t1? If the world were repetitive (governed by immutable laws) and these laws were known, I could assume that what I find out about the present state is relevant to determine how the future state (the one that will prevail) will be. It would make then sense to apply a strategy for gathering empirical evidence (a sequence of actions to collect new data). But if the world is not repetitive, what makes me think that the new information may be at all useful regarding future events? …

Conceiving economic processes like sequences of events in which uncertainty reigns, where consequently there are “no laws”, nor “invariants” or “mechanisms” to discover, the kind of learning that experiments or last experience provide is of no use for the future, because it eliminates innovation and creativity and does not take into account the arboreal character and the open-ended nature of the economic process … However, as said before, we can gather precise information, restricted in space and time (data). But, what is the purpose of obtaining this sort of information if uncertainty about future events prevails? … The problem is that taking uncertainty seriously puts in question the relevance the data obtained by means of testing or experimentation has for future situations.

Marqués’ book is a serious challenge to much of mainstream economic thinking and its methodological and philosophical underpinnings. A must-read for anyone interested in the foundations of economic theory.

To yours truly, Marqués’ book is especially important since it shows how far-reaching the effects of taking Keynes’ concept of genuine uncertainty really are.

treatprobAlmost a hundred years after John Maynard Keynes wrote his seminal A Treatise on Probability (1921), it is still very difficult to find economics textbooks that seriously try to incorporate his far-reaching and incisive analysis of uncertainty, inductive inference and evidential weight.

The standard view in economics and statistics — and the axiomatic probability theory underlying it — is to a large extent based on the rather simplistic idea that ‘more is better.’ But as Keynes argues – ‘more of the same’ is not what is important when making inductive inferences. It’s rather a question of ‘more but different.’

Variation, not replication, is at the core of induction. Finding that p(x|y) = p(x|y & w) doesn’t make w ‘irrelevant.’ Knowing that the probability is unchanged when w is present gives p(x|y & w) another evidential weight (‘weight of argument’). Running 10 replicative experiments do not make you as ‘sure’ of your inductions as when running 10 000 varied experiments – even if the probability values happen to be the same.

According to Keynes we live in a world permeated by unmeasurable uncertainty – not quantifiable stochastic risk – which often forces us to make decisions based on anything but ‘rational expectations.’ Keynes rather thinks that we base our expectations on the confidence or ‘weight’ we put on different events and alternatives. To Keynes, expectations are a question of weighing probabilities by ‘degrees of belief,’ beliefs that often have preciously little to do with the kind of stochastic probabilistic calculations made by the rational agents as modelled by ‘modern’ social sciences. And often we “simply do not know.” As Keynes writes in Treatise:

If different wholes were subject to different laws qua wholes and not simply on account of and in proportion to the differences of their parts, knowledge of a part could not lead, it would seem, even to presumptive or probable knowledge as to its association with other parts … In my judgment, the practical usefulness of those modes of inference … on which the boasted knowledge of modern science depends, can only exist … if the universe of phenomena does in fact present those peculiar characteristics of atomism and limited variety which appears more and more clearly as the ultimate result to which material science is tending.

Science according to Keynes should help us penetrate to “the true process of causation lying behind current events” and disclose “the causal forces behind the apparent facts.” Models can never be more than a starting point in that endeavour. He further argued that it was inadmissible to project history on the future. Consequently, we cannot presuppose that what has worked before, will continue to do so in the future. That statistical models can get hold of correlations between different ‘variables’ is not enough. If they cannot get at the causal structure that generated the data, they are not really ‘identified.’

How strange that writers of economics textbooks do not even touch upon these aspects of scientific methodology that seems to be so fundamental and important for anyone trying to understand how we learn and orient ourselves in an uncertain world. An educated guess on why this is a fact would be that Keynes concepts are not possible to squeeze into a single calculable numerical ‘probability.’ In the quest for quantities one puts a blind eye to qualities and looks the other way — but Keynes ideas keep creeping out from under the carpet.

Robert Lucas once wrote — in Studies in Business-Cycle Theory — that “in cases of uncertainty, economic reasoning will be of no value.”  Now, if that was true, it would put us in a tough dilemma. If we have to consider — as Lucas — uncertainty incompatible with economics being a science, and we actually know for sure that there are several and deeply important situations in real-world contexts where we — both epistemologically and ontologically — face genuine uncertainty, well, then we actually would have to choose between reality and science.

That can’t be right. We all know we do not know very much about the future. We all know the future harbours lots of unknown unknowns. Those are ontological facts we just have to accept. But — I still think it possible we can go for both reality and science, and develop a realist, relevant, non-ergodic economic science.


  1. Comments on Syll´s remarks on my book
    I appreciate very much Syll’s gentle comments about my book. He correctly captures my claim that when uncertainty is taken seriously, there are problems for those that want to defend the possibility (and usefulness) of learning and testing in economics. Lars seems to think that only orthodox economists are in trouble. However, this is a problem for both, orthodox and heterodox economists. Even more, the problem seems to be more compelling for those economists that are committed to uncertainty. Those who are not, like Lucas, do not have to face this sort of incompatibility in their own theoretical system.
    I also underline that the notion of rationality should be rethought if uncertainty is accepted. This is perhaps a more serious challenge for economics than the two mentioned before.
    Part of my book is dedicated to point out these problems. But I go beyond that and propose some constructive ideas about how rationality could be conceived under uncertainty, and in which sense economic practice can be considered rational. On the other hand, elaborating on Sugden’s ideas about models, I suggest how some theoretical models could be successfully used in a realistic sense. These suggestions could serve as a starting point for a defensible realistic methodological analysis.
    Lars thinks that economics must (or at least could) be a science. I have no particular objection to his position on this point. In fact, what “science” means depends in part of our conventions in the use of the word. We may accept (it is a convention, not a fact of nature) that “science” does not entail “testing and learning”. Otherwise the consistency between good theoretical economics and these two requirements (that Sill does find problematic) should be shown. On the other hand, if I´m right, Lawson´s well known view of economics as science is also problematic when it is accepted that radical uncertainty entails no mechanisms of any kind. This remark is important because I suspect that Lars share Lawson’s position on this point.
    In spite of these remarks, in my book I pointed out that economics (orthodox or heterodox) can be science in a weaker sense, one that drops out the necessity of uniformities, laws or mechanisms and demands only accuracy and precision.
    Insisting too much on the scientific character of economics is miss-leading, because even if some (not necessarily scientific) knowledge is required to approach economic processes, considerations about lobby and politics (in the ample sense) may be of a greater importance. In opposition to a contemplative view of the economic processes, I have an activist view in which the economic phenomena are the result of the many interventions of the different economic actors striving each of them to reach their desired results.
    Lars correctly relates my view to that of Keynes. However, my view about the implications of assuming radical uncertainty is not just an expansion of Keynes own vision. I argued that Keynes´ view is not completely consistent and that two different (and opposing) visions coexist in his main texts. Particularly, his treatment of rationality and some of the “techniques” he suggested to face uncertainty seems to be inadequate when this situations prevails. In fact, his techniques are only defendable if a weakest concept (which I called “complexity”) is assumed.
    I hope my book be useful for those who want to analyze some of these problematic features of the practice of economics. A brief outline of my view has been published in Stuart Birks´ blog (see: https://www.worldeconomicsassociation.org/files/2018/01/Issue7-6.pdf). I feel deep indebted to both, Lars and Stuart.
    Gustavo Marqués

  2. ‘Robert Lucas once wrote…….. that “in cases of uncertainty, economic reasoning will be of no value.” ‘

    Solution – conveniently pretend uncertainty doesn’t exist. It won Lucas the Nobel prize. Can’t beat that.

  3. Economics is a social science. It is neither a branch of mathematics nor the study of nature. It is, instead, analysis of humans by humans…. Economics is difficult, because it does not study a demarcated sphere of behaviour. What we consider to be economic behaviour is but a part of the totality of human action. Psychologists, sociologists, anthropologists, geographers and historians also analyse the phenomena considered by economists. The assumption that it is possible to separate out economic behaviour and objectives from other forms of human behaviour and objectives is an heroic simplification and, like all such simplifications, it is fundamentally false. (Rethinking Economics: An Introduction to Pluralist Economics, p. xiii)

  4. “than in the mike’s attempt to answer questions about itself.” Naturally, one must ask, Who is Mike? Well, that is my fat finger always hitting the wrong key 😉 Meant to say, “than in the mind’s attempt to answer questions about itself.”

  5. Reference List

    1. Polkinghorne, John. Science and Religion in Quest of Truth. London: Yale University Press; 2011; p. 23.

    The physical reductionist who claims that there is nothing but matter and energy, and no truth but the truth of science, is making a metaphysical statement as clearly as someone who looks at the world from a theistic perspective. The reductionists have not derived their belief from science alone. Everyone, implicitly or explicitly, has a metaphysics.

    Scientism is the metaphysical belief that science tells us all that can be known or is worth knowing. It must clearly be distinguished from science itself which, owing to its intrinsic limitation to only a certain kind of encounter with reality, is far from being in a position to make such an overblown claim for its explanatory power. Science has bracketed out too much (meaning, purpose, beauty) from its consideration for it to be the universal source of understanding.


    Ethics and morality were important to Adam Smith. Today economics needs a good dose of philosophy and ethics. Physics envy is a dead end, especially given the fact that modern physics has long ago moved beyond the clockwork universe of Newtonian physis that economics seem stuck within.

    • Neoliberalism is Adam Smith without the anxiety about morals …

  6. … if we want to evolve an sustainable economics …

  7. “In the quest for quantities one puts a blind eye to qualities and looks the other way…. If we have to consider — as Lucas — uncertainty incompatible with economics being a science, and we actually know for sure that there are several and deeply important situations in real-world contexts where we — both epistemologically and ontologically — face genuine uncertainty, well, then we actually would have to choose between reality and science.”

    Nobel laureate Sir Peter Medawar said, “Catastrophe apart, I believe it to be science’s greatest glory that there is no limit upon the power of science to answer questions of the kind science _can_ answer.” (P. B. Medawar, The Limits of Science, p. 87)

    What is science? Are there limits to what science can tell us? What is scientific “theory” and what role does it play in science?

    How we answer these questions depends greatly, in my view, upon our pre-existing beliefs (philosophical assumptions) about the nature of reality. Mechanistic materialists who believe mind is nothing more than an epiphenomena of matter approach science in fundamentally different ways than a scientist who accords mind an ontic level of reality. (See, for example, Mindful Universe: Quantum Mechanics and the Participating Observer (The Frontiers Collection) by Henry P. Stapp, http://a.co/aoKKBNB)

    Scientism is an “exaggerated an exaggerated and often distorted conception of what science can be expected to do or explain for us. One aspect of scientism is the idea that any question that can be answered at all can best be answered by science. This, in turn, is very often combined with a quite narrow conception of what it is for an answer, or a method of investigation, to be scientific. Specifically, it is supposed that canonical science must work by disclosing the physical or Chemical mechanisms that generate phenomena. Together these ideas imply a narrow and homogenous set of answers to the most diverse imaginable set of questions. Everywhere this implies a restriction to the powers of the human mind; but nowhere is this restriction more disastrous than in the mike’s attempt to answer questions about itself.” (Dupre 2002, Human Nature and the Limits of Science.)

    Science teaches man to speak the new language of mathematics and trains his thoughts along lines of exacting precision. And science also stabilizes philosophy through the elimination of error, while it purifies religion by the destruction of superstition.

    Science may be physical, but the mind of the truth-discerning scientist is at once supermaterial. Matter knows not truth, neither can it love mercy nor delight in spiritual realities. Moral convictions based on spiritual enlightenment and rooted in human experience are just as real and certain as mathematical deductions based on physical observations, but on another and higher level.

    True science can have no lasting quarrel with true religion. The “scientific method” is merely an intellectual yardstick wherewith to measure material adventures and physical achievements. But being material and wholly intellectual, it is utterly useless in the evaluation of spiritual realities and religious experiences.

    Philosophy is the mediatory between science and religion. Only a holistic view of reality — one that recognizes things, meanings, and values (fact, idea, relation) — can harmonize the divergent views of science and religion.

    Reducing human mind to the status of a mere nomological machine impoverishes our understanding of what it means to be fully human. Human interaction that results in economic behavior is based upon more (or at least it aught to be) than pecuniary self-interest. We are not only material beings, but intellectual and spiritual beings as well. And this is why I think we must take religion seriously, along with science and philosophy, if we want to shove an sustainable economics that address 21st century human needs. Nelson’s Economics as Religion is an important work that reveals the religious motives that once were explicit but are now hidden under a pseudo-science of secular materialism grounded in mechanistic materialism that refuses to accord reality to mind meanings and spiritual values.

    Reference List

    1. Corvellec, Hervé. What is Theory? [Answers From the Social and Cultural Sciences]. Copenhagen: Copenhagen Business Press; 2013; p. 9.

    Notes: “What is theory?” is a difficult one to answer…. [It] is not an innocuous question. Instead, it is a destabilizing question that can take academics off guard and unsettle even entrenched convictions.

Sorry, the comment form is closed at this time.

Blog at WordPress.com.
Entries and Comments feeds.