Economics as a Pretension to Science

Aug 18
17:03

2005

Sam Vaknin

Sam Vaknin

  • Share this article on Facebook
  • Share this article on Twitter
  • Share this article on Linkedin

Economics is not an exact science, nor can it ever be. This is because its "raw material" (humans and their behaviour as individuals and en masse) is not exact.

mediaimage

"It is impossible to describe any human action if one does not refer to the meaning the actor sees in the stimulus as well as in the end his response is aiming at."
Ludwig von Mises


I. Introduction

Storytelling has been with us since the days of campfire and besieging wild animals. It served a number of important functions: amelioration of fears,Economics as a Pretension to Science Articles communication of vital information (regarding survival tactics and the characteristics of animals, for instance), the satisfaction of a sense of order (predictability and justice), the development of the ability to hypothesize, predict and introduce theories and so on.

We are all endowed with a sense of wonder. The world around us in inexplicable, baffling in its diversity and myriad forms. We experience an urge to organize it, to "explain the wonder away", to order it so that we know what to expect next (predict). These are the essentials of survival. But while we have been successful at imposing our mind on the outside world – we have been much less successful when we tried to explain and comprehend our internal universe and our behaviour.

Economics is not an exact science, nor can it ever be. This is because its "raw material" (humans and their behaviour as individuals and en masse) is not exact. It will never yield natural laws or universal constants (like physics). Rather, it is a branch of the psychology of masses. It deals with the decisions humans make. Richard Thaler, the prominent economist, argues that a model of human cognition should lie at the heart of every economic theory. In other words he regards economics to be an extension of psychology.

II. Philosophical Considerations - The Issue of Mind (Psychology)

The relationships between the structure and functioning of our (ephemeral) mind, the structure and modes of operation of our (physical) bodies and the structure and conduct of social collectives have been the matter of heated debate for millennia.

There are those who, for all practical purposes, identify the mind with its product (mass behaviour). Some of them postulate the existence of a lattice of preconceived, born, categorical knowledge about the universe – the vessels into which we pour our experience and which mould it. Others have regarded the mind as a black box. While it is possible in principle to know its input and output, it is impossible, again in principle, to understand its internal functioning and management of information.

The other camp is more "scientific" and "positivist". It speculated that the mind (whether a physical entity, an epiphenomenon, a non-physical principle of organization, or the result of introspection) – has a structure and a limited set of functions. They argue that a "user's manual" can be composed, replete with engineering and maintenance instructions. The most prominent of these "psychodynamists" was, of course, Freud. Though his disciples (Jung, Adler, Horney, the object-relations lot) diverged wildly from his initial theories – they all shared his belief in the need to "scientify" and objectify psychology. Freud – a medical doctor by profession (Neurologist) and Josef Breuer before him – came with a theory regarding the structure of the mind and its mechanics: (suppressed) energies and (reactive) forces. Flow charts were provided together with a method of analysis, a mathematical physics of the mind.

Yet, dismal reality is that psychological theories of the mind are metaphors of the mind. They are fables and myths, narratives, stories, hypotheses, conjunctures. They play (exceedingly) important roles in the psychotherapeutic setting – but not in the laboratory. Their form is artistic, not rigorous, not testable, less structured than theories in the natural sciences. The language used is polyvalent, rich, effusive, and fuzzy – in short, metaphorical. They are suffused with value judgements, preferences, fears, post facto and ad hoc constructions. None of this has methodological, systematic, analytic and predictive merits.

Still, the theories in psychology are powerful instruments, admirable constructs of the mind. As such, they probably satisfy some needs. Their very existence proves it.

The attainment of peace of mind, for instance, is a need, which was neglected by Maslow in his famous model. People often sacrifice material wealth and welfare, forgo temptations, ignore opportunities and put their lives in danger – just to reach this bliss of tranquility. There is, in other words, a preference of inner equilibrium over homeostasis. It is the fulfilment of this overriding need that psychological treatment modalities cater to. In this, they are no different to other collective narratives (myths, for instance).

But, psychology is desperately trying to link up to reality and to scientific discipline by employing observation and measurement and by organizing the results and presenting them using the language of mathematics (rather, statistics). This does not atone for its primordial "sin": that its subject matter (humans) is ever-changing and its internal states are inaccessible and incommunicable. Still, it lends an air of credibility and rigorousness to it.

III. The Scientific Method

To qualify as science, an economic theory must satisfy the following conditions:

  1. All-inclusive (anamnetic) – It must encompass, integrate and incorporate all the facts known.
  1. Coherent – It must be chronological, structured and causal.
  1. Consistent – Self-consistent (its sub-"narratives" cannot contradict one another or go against the grain of the main "narrative") and consistent with the observed phenomena (both those related to the subject and those pertaining to the rest of the universe).
  1. Logically compatible – It must not violate the laws of logic both internally (the narrative must abide by some internally imposed logic) and externally (the Aristotelian logic which is applicable to the observable macro world).
  1. Insightful – It must inspire a sense of awe and astonishment, which is the result of seeing something familiar in a new light or the result of seeing a pattern emerging out of a big body of data ("data mining"). The insights must be the inevitable conclusion of the logic, the language and of the development of the narrative.
  1. Aesthetic – The narrative must be both plausible and "right", beautiful (aesthetic), not cumbersome, not awkward, not discontinuous, smooth and so on.
  1. Parsimonious – The narrative must employ the minimum number of assumptions and entities in order to satisfy all the above conditions.
  1. Explanatory – The narrative must explain the behaviour of economic actors, their decisions, why events develop the way they do.
  1. Predictive (prognostic) – The narrative must possess the ability to predict future events, the future behaviour of economic actors and of other meaningful figures and the inner emotional and cognitive dynamics of said actors.
  1. Prescriptive – With the power to induce change (whether it is for the better, is a matter of contemporary value judgements and fashions).
  1. Imposing – The narrative must be regarded by society as the preferable and guiding organizing principle.
  1. Elastic – The narrative must possess the intrinsic abilities to self organize, reorganize, give room to emerging order, accommodate new data comfortably, avoid rigidity in its modes of reaction to attacks from within and from without.

In some of these respects, current economic narratives are usually theories in disguise. But scientific theories must satisfy not only most of the above conditions. They must also pass the crucial hurdles of testability, verifiability, refutability, falsifiability, and repeatability – all failed by economic theories. Many economists argue that no experiments can be designed to test the statements of economic narratives, to establish their truth-value and, thus, to convert them to theorems.

There are five reasons to account for this shortcoming - the inability to test hypotheses in economics:

  1. Ethical – Experiments would have to involve humans. To achieve the necessary result, the subjects will have to be ignorant of the reasons for the experiments and their aims. Sometimes even the very performance of an experiment will have to remain a secret (double blind experiments). Some experiments may involve unpleasant experiences. This is ethically unacceptable.
  1. Design Problems - The design of experiments in economics is awkward and difficult. Mistakes are often inevitable, however careful and meticulous the designer of the experiment is.
  1. The Psychological Uncertainty Principle – The current position of a human subject can be (theoretically) fully known. But the passage of time and the experiment itself influence the subject and void this knowledge ("time inconsistencies"). The very processes of measurement and observation influence the subject and change him.
  1. Uniqueness – Experiments in economics, therefore, tend to be unique and cannot be replicated elsewhere and at other times even if they deal with the SAME subjects. The subjects (the tested humans) are never the same due to the aforementioned psychological uncertainty principle. Repeating the experiments with other subjects adversely affects the scientific value of the results.
  1. The undergeneration of testable hypotheses – Economics does not generate a sufficient number of hypotheses, which can be subjected to scientific testing. This has to do with the fabulous (=storytelling) nature of the discipline. In a way, Economics has affinity with some private languages. It is a form of art and, as such, is self-sufficient. If structural, internal constraints and requirements are met – a statement is deemed true even if it does not satisfy external (scientific) requirements. Thus, the standard theory of utility is considered valid in economics despite empirical evidence to the contrary - simply because it is aesthetic and mathematically convenient.

So, what are economic narratives good for?

Narratives in economics offer an organizing principle, a sense of order and ensuing justice, of an inexorable drive toward well defined (though, perhaps, hidden) goals, the ubiquity of meaning, being part of a whole. They strive to answer the "why’s" and "how’s". They are dialogic and prescriptive (=provide behavioural prescriptions). The client (let's say, a politician) asks: "Why am I (and here follows an economic problem or behaviour". Then, the narrative is spun: "The situation is like this not because the world is whimsically cruel but because...and if you were to do this or that the situation is bound to improve". The client is calmed by the very fact that there is an explanation to that which until now bothered him, that there is hope and - providing he follows the prescriptions - he cannot be held responsible for a possible failure, that there is who or what to blame (focussing diffused anger is a very policy instrument) and, that, therefore, his belief in order, justice and their administration by some supreme, transcendental principle is restored. This sense of "law and order" is further enhanced when the narrative yields predictions which come true (either because they are self-fulfilling or because some real "law"- really, a pattern - has been discovered).

IV. Current Problems in Economics

Neo-classical economics has failed on several fronts simultaneously. This multiple failure led to despair and the re-examination of basic percepts and tenets:

1. The Treatment of Government

Government was accorded a special status and special treatment in economic theory (unlike other actors and agents). It was alternatively cast as a saint (seeking to selflessly maximize social welfare) - or as the villain (seeking to perpetuate and increase its power ruthlessly, as in public choice theories). Both views are caricatures of reality. Governments do seek to perpetuate and increase power but they use it mostly to redistribute income and not for self-enrichment.

2. Technology and Innovation

Economics failed to account for the role of innovation in growth and development. It also ignored the specific nature of knowledge industries (where returns increase rather than diminish and network effects prevail). Thus, current economic thinking is woefully inadequate to deal with information monopolies (such as Microsoft), path dependence and pervasive externalities.

3. Long Term Investment Horizons

Classic cost/benefit analyses fail to tackle very long term investment horizons (periods). Their underlying assumption (the opportunity cost of delayed consumption) fails beyond the investor's useful economic life expectancy. Put more plainly: investors care less about their grandchildren's future than about their own. This is because predictions concerned with the far future are highly uncertain and people refuse to base current decisions on fuzzy "what ifs". This is a problem because many current investments (example: the fight against global warming) are likely to yield results only in the decades ahead. There is no effective method of cost/benefit analysis applicable to such time horizons.

4. Homo Economicus

The economic actor is assumed to be constantly engaged in the rational pursuit of self interest. This is not a realistic model - merely a (useful) approximation. People don't repeat their mistakes systematically (=rationality in economics) and they seek to optimize their preferences (altruism can be such a preference, as well).

Still, many people are non-rational or only nearly rational in certain situations. And the definition of "self-interest" as the pursuit of the fulfilment of preferences is a tautology.

V. Consumer Choices

How are consumer choices influenced by advertising and by pricing? No one seems to have a clear answer. Advertising is both the dissemination of information and a signal sent to consumers that a certain product is useful and qualitative (otherwise, why would a manufacturer invest in advertising it)? But experiments show that consumer choices are influenced by more than these elements (for instance, by actual visual exposure to advertising).

VI. Experimental Economics

People do not behave in accordance with the predictions of basic economic theories (such as the standard theory of utility and the theory of general equilibrium). They change their preferences mysteriously and irrationally ("preference reversals"). Moreover, their preferences (as evidenced by their choices and decisions in experimental settings) are incompatible with each other. Either economics is not testable (no experiment to rigorously and validly test it can be designed) - or something is very flawed with the intellectual pillars and models of economics.

VII. Time Inconsistencies

People tend to lose control of their actions or procrastinate because they place greater importance (greater "weight") on the present and the near future than on the far future. This makes them both irrational and unpredictable.

VIII. Positivism versus Pragmatism

Should economics be about the construction and testing of of models, which are consistent with basic assumptions? Or should it revolve around the mining of data for emerging patterns (=rules, "laws")? On the one hand, patterns based on a limited set of data are, by definition, inconclusive and temporary and, therefore, cannot serve as a basis for any "science". On the other hand, models based on assumptions are also temporary because they can (and are bound to) be replaced by new models with new (better?) assumptions.

One way around this apparent quagmire is to put human cognition (=psychology) at the heart of economics. Assuming that the human is immutable and knowable - it should be amenable to scientific treatment. "Prospect theory", "bounded rationality theories" and the study of "hindsight bias" and other cognitive deficiencies are the fruits of this approach.

IX. Econometrics

Humans and their world are a multi-dimensional, hyper-complex universe. Mathematics (statistics, computational mathematics, information theory, etc.) is ill equipped to deal with such problems. Econometric models are either weak and lack predictive powers or fall into the traps of logical fallacies (such as the "omitted variable bias" or "reverse causality").