Robert U Ayres - Turning point: an end to the growth paradigm

Before plunging headlong into controversy, I want to make some (hopefully) non-controversial remarks about forecasting. It is common for writers who disparage the activity to note that prediction is difficult, especially as concerns the future. Isaac Asimovs Foundation novels feature a psychohistorian genius named Hari Seldon, who has developed a probabilistic theory of social systems. An early scene has Seldon demonstrating to a young research assistant that there is a probability of 92.5 per cent that the galactic empire will collapse into anarchy within 500 years (Asimov, 1951). The notion that any such prediction could be made from a set of mathematical symbols has always tickled my funny bone.

Anybody who has wrestled with forecasting models will share my amusement at the navet of the notion that future history could ever be reduced to a set of equations. We can take it for granted that 100 per cent accuracy in forecasting is impossible. Nevertheless, some important things can be said. One of them is that not all imaginable futures are possible. Some imaginable futures are simply inconsistent with the laws of nature, as far as we know them. (For instance, most of the gadgets and plots of Star Trek violate several of these laws. Unfortunately, it is not only on TV that impossible futures are taken seriously by some people.)

Of course, some possible futures are far more plausible than others. It is imaginable, but very implausible, that all the nations and peoples of the earth would decide spontaneously to disarm, live in peace with each other, and decide all disputes by referendum and a world court. There are, however, still a great many plausible but different futures. It is impossible to predict which among the many plausible ones will come to pass. On the other hand, it is the legitimate business of forecasting to select the most plausible among them for consideration, particularly as a guide to policy. In this context, there are two very different types of forecast. For symbolic reasons, I will call them alpha and omega.

The first type is the alpha forecast or contingent extrapolation. It is a description of the situation that will arise if present social or demographic trends, or policies, continue as currently. An important point to note is that not all contingent extrapolations lead to desirable outcomes. Some outcomes are flatly impossible. Thus the contingent extrapolation may point to a need for policy or behavioral change. In such a case, the forecast is intended to be self-denying.

An interesting example comes from the history of technology. In the early 1940s a vice-president of Bell Telephone Laboratories extrapolated the growth of demand for telephone services for several decades. Then he extrapolated the need for automated central telephone exchanges. Finally he calculated the amount of electric power needed to serve those exchanges (at the time, telephone exchanges were based on electromechanical switches). According to the story, he then noticed that demand for electric power for electromechanical telephone switching would exceed the then current forecasts of total demand for electric power for all purposes.

This was a self-denying forecast. The Bell Laboratories man saw that continued growth of the telephone system would require a new kind of switching technology that would use much less electric power than the (then) current technology. As luck would have it, there were some young physicists at Bell Laboratories who thought that a strange phenomenon in certain transition metals and alloys called semi-conductivity might point the way. Their names were William Shockley, John Bardeen and Walter Brattain. They were assigned to work on the project. The end result was the invention of the transistor, surely one of the dozen or so most important inventions since the wheel. Needless to say, demand for telecommunications has multiplied hundreds of times since then, but demand for electric power to run the telephone system is now almost unnoticeable.

The second type of forecast (omega) is an attempt to describe a desirable distant future and, working back, to delineate possible pathways to that future. It is sometimes called a normative forecast. Such a forecast also has policy implications. Needless to say, an omega forecast is intended to be self-fulfilling, rather than self-denying. An example might be John Kennedys 1962 forecast that an American would land on the moon before the end of the decade. Kennedys forecast was a call to action. It galvanized the Congress to create NASA and to provide the funds needed. The moon landing was achieved, on time, in 1969.