Small Is Beautiful: A Study of Economics as if People Mattered (27 page)

Planners, of course, proceed on the assumption that the future is not

'already here', that they are not dealing with a predetermined - and therefore predictable - system, that they can determine things by their own free will, and that their plans will make the future different from what it would have been had there been no plan. And yet it is the planners, more than perhaps anyone else, who would like nothing better than to have a machine to foretell the future. Do they ever wonder whether the machine might incidentally also foretell their own plans before they have been conceived?

Need for Semantics

However this may be, it is clear that the question of predictability is not only important but also somewhat involved. We talk happily about estimating, planning, forecasting, budgeting, about surveys, programmes, targets, and so forth, and we tend to use these terms as if they were freely interchangeable and as if everybody would automatically know what was meant. The result is a great deal of confusion, because it is in fact necessary to make a number of fundamental distinctions. The terms we use may refer to the past or to the future; they may refer to acts or to events: and they may signify certainty or uncertainty. The number of combinations possible where there are three pairs of this kind is 2, or 8, and we really ought to have eight different terms to be quite certain of what we are talking about. Our language, however, is not as perfect as that. The most important distinction is generally that between acts and events. The eight possible cases may there fore be ordered as follows:

1 Act Past Certain

2 Act Future Certain

3 Act Past Uncertain

4 Act Future Uncertain

5 Event Past Certain

6 Event Future Certain

7 Event Past Uncertain

8 Event Future Uncertain

The distinction between acts and events is as basic as that between active and passive or between 'within my control' or 'outside my control'. To apply the word 'planning' to matters outside the planner's control is absurd. Events, as far as the planner is concerned, simply happen. He may be able to forecast them and this may well influence his plan; but they cannot possibly be p art of the plan. The distinction between the past and the future proved to be necessary for our purpose, because, in fact, words like 'plan' or 'estimate' are being used to refer to either. If I say: 'I shall not visit Paris without a plan,'

this can mean: 'I shall arm myself with a street plan for orientation' and would therefore refer to case 5. Or it can mean: 'I shall arm myself with a plan which outlines in advance where I am going to go and how I am going to spend my time and money' - case 2 or 4. If someone claims that 'to have a plan is indispensable', it is not without interest to find out whether he means the former or the latter. The two are essentially different.

Similarly, the word 'estimate', which denotes uncertainty, may apply to the past or to the future. In an ideal world, it would not be necessary to make estimates about things that had already happened. But in the actual world, there is much uncertainty even about matters which, in principle, could be fully ascertained. Cases 3, 4, 7, and 8 represent four different types of estimates. Case 3 relates to something I have done in the past; case 7, to something that has happened in the past. Case 4 relates to some- thing I plan to do in the future, while case 8 relates to something I expect to happen in the future. Case 8, in fact, is a forecast in the proper sense of the term and has nothing whatever to do with 'planning'. How often, however, are forecasts presented as if they were plans - and vice versa! The British

'National Plan' of 1965 provides an outstanding example and, not surprisingly, came to nothing.

Can we ever speak of future acts or events as certain (cases 2 and 6)? If I have made a plan with full knowledge of all the relevant facts, being inflexibly resolve to carry it through - case 2 - I may, in this respect, consider my future actions as certain. Similarly, in laboratory science, dealing with carefully isolated deterministic systems, future 2vents may he described as certain. The real world, however, is not a deterministic system; we may be able to talk with certainty about acts or events of the past - cases 1 or 5 - but we can do so about future events only on the basis of assumptions. In other words, we can formulate conditional statements about the future, such as: ‘If such and such a trend of events continued for another x years, this is where it would take us.' This is not a forecast or prediction, which must always be uncertain in the real world, but an exploratory calculation, which, being conditional, has the virtue of mathematical certainty.

Endless confusion results from the semantic muddle in which we find ourselves today. As mentioned before, 'plans' are put forward which upon inspection turn out to relate to events totally outside the control of the planner. 'Forecasts' are offered which upon inspection turn out to be conditional sentences, in other words, exploratory calculations. The latter are misinterpreted as if they were forecasts or predictions. 'Estimates' are put forward which upon inspection turn out to be plans. And so on and so forth.

Our academic teachers would perform a most necessary and really helpful task if they taught their students to make the distinctions discussed above and developed a terminology which fixed them in words.

Predictability Let us now return to our main subject - predictability. Is prediction or forecasting - the two terms would seem to be interchangeable -

possible at all? The future does not exist; how could there be knowledge about something non-existent? This question is only too well justified. In the strict sense of the word, knowledge can only be about the past. The future is always in the making, but it is being made largely out of existing material, about which a great deal can be known. The future, there, is largely predictable, if we have solid and extensive knowledge of the past, Largely, but by no means wholly, for into the making of the future there enters that mysterious and irrepressible factor called human freedom. It is the freedom of a being of which it has been said that it was made in the image of God the Creator: the freedom of creativity.

Strange to say, under the influence of laboratory science many people today seem to use their freedom only for the purpose of denying its existence Men and women of great gifts find their purest delight in magnifying every mechanism', every inevitability', everything where human freedom does not enter or does not appear to enter A great shout of triumph goes up whenever anybody has found some further evidence - in physiology or psychology or sociology or economics or politics - of unfreedom, some further indication that people cannot help being what they are and doing what they are doing, no matter how inhuman their actions might be. The denial of freedom, of course, is a denial of responsibility: there are no acts, but only events; everything simply happens; no-one is responsible. And this is no doubt the main cause of the semantic confusion to which I have referred above. It is also the cause for the belief that we shall soon have a machine to foretell the future.

To be sure, if everything simply happened, if there were no element of freedom, choice, human creativity and responsibility, everything would be perfectly predictable, subject only to accidental and temporary limitations of knowledge. The absence of freedom would make human affairs suitable for study by the natural sciences or at least by their methods, and reliable results would no doubt quickly follow the systematic observation of facts. Professor Phelps Brown, in his presidential address to the Royal Economic Society, appears to adopt precisely this point of view when talking about 'The Underdevelopment of Economics'. 'Our own science,' he says, 'has hardly yet reached its seventeenth century. Believing that economics is metaphysically the same as physics, he quotes another economist, Professor Morgenstern, approvingly as follows:

'The decisive break which came in physics in the seventeenth century, specifically in the field of mechanics, was possible only because of previous developments in astronomy. It was backed by several millennia of systematic, scientific. astronomical observation.... Nothing of this sort has occurred in economic science. It would have been absurd in physics to have expected Kepler and Newton without Tycho - and there is no reason to hope for an easier development in economics.' Professor Phelps Brown concludes therefore that we need many. many more years of observations of behaviour,

'Until then, our mathematisation is premature.’

It is the intrusion of human freedom and responsibility that makes economics metaphysically different from physics and makes human affairs largely unpredictable. We obtain predictability, of course, when we or others are acting according to a plan. But this is so precisely because a plan is the result of an exercise in the freedom of choice: the choice has been made: all alternatives have been eliminated. If people stick to their plan, their behaviour is predictable simply because they have chosen to surrender their freedom to act otherwise than prescribed in the plan.

In principle, everything which is immune to the intrusion of human freedom, like the movements of the stars, is predictable, and everything subject to this intrusion is unpredictable. Does that mean that all human actions are unpredictable? No, because most people, most of the time, make no use of their freedom and act purely mechanically. Experience shows that when we are dealing with large numbers of people many aspects of their behaviour are indeed predictable; for out of a large number, at any one time, only a tiny minority are using their power of freedom, and they often do not significantly affect the total outcome. Yet all really important innovations and changes normally start from tiny minorities of people who do use their creative freedom.

It is true that social phenomena acquire a certain steadiness and predictability from the non-use of freedom, which means that the great majority of people responds to a given situation in a way that does not alter greatly in time, unless there are really overpowering new causes.

We can therefore distinguish as follows:

(a) Full predictability (in principle) exists only in the absence of human freedom, i.e. in 'sub-human' nature. The limitations of predictability are purely limitations of knowledge and technique.

(b) Relative predictability exists with regard to the behaviour pattern of very large numbers of people doing 'normal' things (routine).

(C) Relatively full predictability exists with regard to human' actions controlled by a plan which eliminates freedom, e.g. railway timetable.

(d) Individual decisions by individuals are in principle unpredictable.

Short-Term Forecasts

In practice all prediction is simply extrapolation, modified by known

'plans'. But how do you extrapolate? How many years do you go back?

Assuming there is a record of growth, what precisely do you extrapolate -

the average rate of growth. or the increase in the rate of growth, or the annual increment in absolute terms? As a matter of fact. there are no rules:*

it is just a matter of 'feel' or judgment.

It is good to know of all the different possibilities of using the same time series for extrapolations with very different results. Such knowledge will prevent us from putting undue faith in any extrapolation. At the same time, and by the same token, the development of (what purport to be) better forecasting techniques can become a vice. In short-term forecasting, say, for next year, a refined technique rarely produces significantly different results from those of a crude technique. After a year of growth - what can you predict?

(a) that we have reached a (temporary) ceiling; (b) that growth will continue at the same, or a slower, or a faster rate; (c) that there will be a decline.

Now, it seems clear that the choice between these three basic alternative predictions cannot be made by 'forecasting technique but only by informed judgment. It depends, of course, on what you are dealing with. Which you have something that is normally growing very fast, like the consumption of electricity, your threefold choice is between the came rate of growth, a faster rate, or a slower rate.

It is not so much forecasting technique, as a full understanding of the current situation shat can help in the formation of a sound judgment for the future. If the present level of performance (or rate of growth) is known to be influenced by quite abnormal factors which are unlikely to apply in the coming year, it is, of course, necessary to take these into account. The forecast, 'same as last year', may imply a 'real' growth or a 'real' decline on account of exceptional factors being present this year, and this, of course, must be made explicit by the forecaster;

I believe, therefore, that all effort needs to be put into understanding the current situation, to identify and, if need be, eliminate 'abnormal' and non-recurrent factors from the current picture. This having been done, the method of forecasting can hardly be crude enough. No amount of refinement will help one come to the fundamental judgment - is next year going to be the same as last year, or better, or worse?

At this point, it may be objected that there ought to be great possibilities of short-term forecasting with the help of electronic computers, because they can very easily and quickly handle a great mass of data and fit to them some kind of mathematical expression. By means of 'feedback' the mathematical expression can be kept up to date almost instantaneously. And once you have a really good mathematics then the machine can predict the future.

Once again, we need to have a look at the metaphysical basis of such claims. What is the meaning of a 'good mathematical fit'? Simply that a sequence of quantitative changes in the past has been elegantly described in precise mathematical language. But the fact that I - or the machine - have been able to describe this sequence so exactly by no means establishes a presumption that the pattern will continue. It could continue only if (a) there were no human freedom and (b) there was no possibility of any change in the causes that have given rise to the observed pattern, I should accept the claim that a very clear and very strongly established pattern !of stability, growth, or decline) can be expected to continue for a little longer, unless there is definite knowledge of the arrival of new factors likely to change it. But I suggest that for the detection of such clear, strong and persistent patterns the non-electronic human brain is normally cheaper, faster, and more reliable than its electronic rival. Or to put it the other way round: if it is really necessary to apply such highly refined methods of mathematical analysis for the detection of a pattern that one needs an electronic computer, the pattern is too weak and too obscure to be a suitable basis for extrapolation in real life.