“The Pretence of Knowledge”: how economists come to make astrologers look good


(Update: This post is attracting significant traffic, so I have indicated three Austrian-School primers throughout the text. Please click the images for the publications.)

It seems everyone is talking about swine flu and the financial crisis, but they seem to be most persistently interested in what has gone wrong with the economy. After I explained the key points of this Nobel Prize Lecture by F A Hayek to my hairdresser this morning, she had confirmed in her mind what she always knew: that economists cannot discover all they need to make accurate predictions.

This is something we all need to understand.

What follows is a precis of that 1974 lecture (sometimes quoted verbatim). It explains why and how the economic policies of the time contributed to inflation and unemployment and it points the way out of our present and coming difficulties.

This is its message:

  • Physical scientists can observe and measure the things that drive the sytems they are studying.
  • Society, and therefore the economy, is not like a physical system: many of the most important factors cannot be seen or measured. Consider the thoughts and intended actions of millions of people at different times, for example.
  • Economists and other social scientists, in their attempt to be scientific, ignore what they cannot measure.
  • Therefore, many of the most important factors affecting the economy are not considered, while some of those factors which can be measured are deliberately controlled.
  • The results are incorrect predictions and actions which positively harm society.

In a sentence: society is not a machine to be controlled, but a garden to be cultivated.

By the way, the remark about making astrologers look good is attributable to the Keynesian economist J K Galbraith:

The only function of economic forecasting is to make astrology look respectable.

The precis:

A Precis of “The Pretence of Knowledge”

The topic of this lecture is the chief practical problem of economists, who must now explain how to stop the accelerating inflation they have caused. In imitating the techniques of the physical sciences, economists have made grave errors of economic policy.

The assertion guided policy that there exists a positive correlation between total employment and the aggregate demand for goods and services. It implies we can permanently ensure full employment by maintaining total expenditure. This is fundamentally false and very harmful.

Economics deals with fundamentally complex phenomena[1]: we cannot obtain all the data and we may not be able to find the important ones. The physical sciences reasonably assume the important factors will be measurable but, in the market, which depends upon the actions of many individuals, the circumstances determining the outcomes will hardly ever be fully known or measurable. The social sciences often treat as important only that which is measurable, sometimes in order to fulfill a demand to state theories in those terms.

Such a demand limits the facts which may be admitted as causes of real events. It is a fiction that only the measurable factors are relevant; to think that this is scientific is naive.

The relation between aggregate demand and total employment, though approximate, is accepted as the only relation which counts because it is the only relation for which there is quantitative data. On this standard, false theories may be accepted because they are seen as more scientific and valid explanations rejected for lack of quantitative evidence.

This can be illustrated by sketching the chief cause of unemployment which appears to be the existence of discrepancies between what people want to buy and what people have been assigned to produce. We have a fairly good understanding of the forces which bring about an alignment of supply and demand, of the necessary conditions, and of the obstacles. An account of this understanding relies on the facts of everyday experience and few would disagree with the logical conclusions. We have good reason to believe that unemployment indicates that the structure of relative prices and wages has been distorted and that to restore balance requires some transfers of labour.

But when we are asked for quantitative evidence for the structure of prices and wages that would bring ‘equilibrium’, we must admit we do not know. We know the general conditions for ‘equilibrium’, but we cannot know the particular prices and wages which would be established. Ours is not a theory which can provide predictions of prices and wages.

Why should economists have to plead ignorance of the sort of facts physical scientists would have to explain? The social sciences, unlike the physical sciences, must deal with structures of essential complexity, that is, models made up of large numbers of variables.

Some fields can overcome this by using probability, but this is valid only for systems of unorganized complexity. The organized complexity dealt with by the social sciences relies on the interdependence between individual elements, not just their properties and frequency of occurrence. Statistical information about individual elements in isolation is therefore not enough: we require full information about each element in the system. Without such information, we can only predict patterns, not specifics.

This is particularly true of relative prices and wages. We cannot observe and measure the subjective information known by each person in the market, nor can any one mind fully comprehend it. This is the essential superiority of the market system. We cannot know what particular structure of prices and wages would bring about an equilibrium of supply and demand nor can we measure deviations from that order or statistically test our theory that it is those deviations which make it impossible to sell some products and services at their offer price.

The mathematical method is a great advantage in describing the general character of patterns but it has led to the illusion that we can determine and predict specific numerical values. This has led to a vain search for constants. While the founders of modern mathematical economics provided equations that could in principle predict prices and quantities, they were under no illusion that it would be possible. Pareto said it would be “absurd” to assume we could ascertain all the data. It has been historically well known that the price depends on so many factors it can never be predicted: this is not a young branch of research.

The superstition that only measurable magnitudes can be important has created few instances of positive harm in economics. However, in inflation and unemployment, it has produced policies which have made matters worse. It is preferable to have true yet imperfect knowledge instead of a pretence of exact knowledge which is likely to be false. Unwarranted credit for seemingly simple but false theories has grave consequences.

In fact, the very measures recommended as a remedy for unemployment — raising aggregate demand — have become the source of a large-scale misallocation of resources which is likely to make later large-scale unemployment inevitable. Continuously injecting additional amounts of money where it creates temporary demand, together with an expectation of continuously rising prices, draws labour and resources into use in areas which will last only as long as the supply of new money. These policies bring about not so much a raise in the level of employment, but a distribution of employment which cannot last and which eventually can only be maintained by ruinous levels of inflation. The position is precarious, in which we cannot prevent substantial unemployment from reappearing. Unemployment is not deliberately brought about to combat inflation; it is a deeply regrettable but inescapable consequence of mistaken policies as soon as inflation ceases to accelerate.

Returning to errors concerning the nature of social science, there is reason to be apprehensive about the long-run danger of the uncritical acceptance of assertions which appear to be scientific: what superficially looks like the most scientific procedure is often the most unscientific. In these fields, there are definite limits to what we can expect science to achieve. To entrust to science, or deliberately to control scientifically, more than the scientific method can achieve, may have deplorable effects. The progress of natural science has been such that there are those who will resist the insight that our increasing power of prediction and control cannot be used to mould society to our liking. It sometimes almost seems as if the techniques of science were more easily learnt than the thinking that shows us what the problems are and how to approach them.

The conflict between public hopes for the achievements of science and what is in its power is serious. Even if true scientists recognize the limits of what can be achieved in human affairs, there will always be some who wish to meet the public hope. Distinguishing between legitimate and illegitimate scientific claims can be difficult or impossible. The silence of the media on the critiques of The Limits to Growth in contrast to the report itself is a a source of apprehension. Economics is not the only field in which far-reaching claims are made on behalf of a more scientific direction of all human activities in place of spontaneous processes.

If we are to safeguard the reputation of science, we must debunk the arrogations of those who make specious claims for what science can achieve in the fields of social science. In these fields, there are absolute obstacles to the prediction of specific events; to act as if we possessed transcendent scientific knowledge may hinder the advance of the human intellect.

The chief point to remember is that physical science advanced rapidly in fields in which explanation and prediction could be based on functions of relatively few variables. To derive a prediction from a large number of observed facts, we must obtain those facts. Given modern computers, deriving predictions from those facts should be easy enough. The real difficulty, to which science has little to contribute, is the determination of those facts.

Consider a ball game played by a few people of approximately equal skill. If we knew some additional facts, such as the state of their attention and their physiology, at each moment in the game, we could probably predict the outcome. But of course we cannot obtain those facts and the result of the game is outside the scientifically predictable. That does not mean we can make no predictions at all, but our capacity to predict will be general and not specific.

As we advance further into systems of organized complexity, we find increasingly that we cannot ascertain all of the facts which determine the outcome of a given process. Often all we can predict is some abstract characteristic of the pattern that will appear, yet we still have predictions which can be falsified and which are therefore significant.

Of course compared to the predictions of physical sciences, this sort of pattern prediction is a second best with which one does not wish to be content. Yet the danger is the belief that to have a claim accepted as scientific, one must do more. To act as if we have the knowledge and the power to shape society to our liking, when in fact we do not possess these, is likely to do us much harm. In the physical sciences, attempting the impossible may be beneficial but in the social field, the erroneous belief that some power would have beneficial consequences is likely to lead to new coercive authorities. Even if such power is not bad in itself, its exercise is likely to impede those spontaneous processes on which we rely.

If man is not to do more harm than good in his efforts to improve the social order, he will have to learn that he cannot acquire the full knowledge which would make mastery of events possible. He will therefore have to use what knowledge he can achieve not to control but to cultivate in an appropriate environment, as a gardener does for his plants. The recognition of the insuperable limits to his knowledge ought to teach the student of society a lesson of humility which should guard him against becoming an accomplice in men’s fatal striving to control society – a striving which makes him not only a tyrant over his fellows, but which may well make him the destroyer of a civilization which no brain has designed but which has grown from the free efforts of millions of individuals.


[1] — The point is not lost on modern sociology. Developing the theme of chaos theory — that certain systems are highly sensitive to initial conditions which cannot be perfectly known — complexity theory has been applied to social science since the 1990’s.

Tags: , , , , ,

Comments & Responses

Comments are closed.