On Growth

11. Carl Kaysen

Carl Kaysen has been director of the Institute of Advanced Study at Princeton, New Jersey, since 1966.
He was born in Philadelphia, Pennsylvania, in 1920. He studied economics, first at Columbia University, and obtained a Ph.D. from Harvard in 1954.
During World War II he served in the Office of Strategic Services in Washington, D.C., with an accent on problems concerning the German war economy. After entering the Air Force he served in Europe as intelligence officer.
From 1947 to 1968 he was consultant to the Rand Corporation in Santa Monica, California. From 1953 to 1954 he acted as consultant to the Weapons System Evaluation Group in the office of the Secretary of Defense. In 1961 he worked as White House staff member to President John F. Kennedy in the office of the special assistant for national security affairs. From 1961 to 1963 he became deputy special assistant to the President for national security affairs. From 1963 to 1965 he acted as special consultant to President Lyndon B. Johnson.

[p. 66]

Why are some of the world's top economists so critical of the MIT study Limits to Growth?

Because the model around which the argument is built, ignores some elementary economic propositions. The most important elementary economic proposition which the model ignores is that there are adjustment mechanisms through prices that when things become scarcer, that prices rise, change, that change in prices in turn includes further adjustment and demand on the one side, lessening the supply and on the other hand increasing it, calling forth new technologies, new substitutes. These adjustments have in the past led to increasing supplies of resources. The fact they are nowhere recognized or discussed makes economists suspicious of the model.

Could resources be measured at all in economic or physical terms, by computer simulation or otherwise?

Resources can certainly be measured in physical terms and they conventionally are. One speaks of acres of land, or tons of coal reserves and millions of oil reserves and so on. On the other hand, I think one gets a deeper insight into resources by measuring them or thinking of them in economic terms. There are acres of land which are easily accessible and easily operable. More acres of land can be added at a higher price by irrigation, or by reclamation or by swamp draining and so forth. Similarly, if one thinks of a mineral resource, there are easily available supplies, there are less available supplies - at higher prices of recovery, more is available.

I was speaking only the other day with an executive of one of the larger aluminum companies, who remarked that they were moving from bauxite, which is the most easily worked aluminum ore, to a somewhat less easily worked aluminum ore never before used but which has enormous reserves, and potential. This may mean that a ton of aluminum will cost a few more cents, but at a few more cents reserves increase, so that one of the problems of the notion of limits to growth is that the measurement of resources in physical terms obscures the important fact that they look different when measured in economics.

But more land, more of this new aluminum source, it all will need more money. How would poor countries, the Third World, raise the money in order to develop new land, irrigate it, cultivate it?

[p. 67]

There is no other way the problems of poverty in the Third World can be solved except by general world economic growth. Without general world economic growth the problem of the Third World will get worse, not better. I think it is totally unrealistic to expect that there can be a radical redistribution of wealth and income without growth in the present. First of all, a higher standard of living in the poor countries implies by the very fact economic growth in those countries. Since they include nearly two-thirds of the world population, economic growth in those countries implies economic growth in the world as a whole. Further, economic growth in the poor countries depends very heavily on trade with and investments from the richer countries. It's again really impossible to conceive that they will be increasing trade and increasing investments when incomes in the richer countries are going down, not up.

Would you say there are signs that technology also grows exponentially?

We have a good deal of history on that. Of course, predicting the future is a chancy business. The only way we can do it is to look at the experience of the past. The area in which this has been best studied in quantitative terms is in the United States. Over the last century approximately, technology in the United States - the value of output that can be got from a given bundle of input - has been growing at about two percent per year, in other words exponentially. We have information for rather shorter periods on some European countries and on Japan. They also show exponential growth of technology in the same sense. There is a very faint indication - which I myself find too inconclusive to base a reliable judgment on - that the rate of technological growth has increased in recent years. I think one will have to wait another decade or two to see whether this is in fact the case.

Bringing this increase of technological growth into the model, the MIT planetary model, that would of course change the outcome of the results.

It would be a very fundamental change in the model. This particular model can be described rather crudely in the following way: You have quantities, population, industrial production and the like, which are growing exponentially. You have a fixed ceiling. When these quantities bump into the ceiling, they not only cannot go through it, they must bounce back with a great crash. That great crash is the downturn in world

[p. 68]

population, world wealth, a catastrophe which the MIT model predicts.

Once we introduce the exponential growth of technology, what we see is that the ceiling is also growing. The question then becomes how fast is one growing with relation to one another. Further, at this point we can then go back to examine what is the adjustment mechanism which connects, so to speak, the level of the ceiling or its rate of growing with the rate of growth of these quantities like population, industrial production, use of mineral inputs and so on.

Adjustment mechanisms are vital?

No useful economic model is complete without some account of adjustment mechanisms. The very important adjustment mechanism which is central to the whole economic idea is the mechanism of the price system. The MIT model nowhere recognizes that this mechanism comes into account.

Take a very simple reason for skepticism about the arguments that are made about the limits of mineral resources: It is not in general true that the share of minerals in the national output or the share of food in the national output, agricultural production, has been rising. Rather it has been falling over the long period. This is certainly true in the United States, where we have good lengthy-period figures. It seems to be true in Western Europe as well, though it can be that our information to this comparison is not as complete or does not stretch over a longer period of time. If the MIT thesis were correct, we would expect to see rising a relative share of output in minerals and more resources needed to supply these scarcer and scarcer minerals. We see the opposite, a gentle decline in the share of output.

Talking to Professor Nordhaus¹ at Yale, who constructed a model with Professor James Tobin of Yale, he said that the MIT model was far too complicated. He pointed to the need for a much simpler model than the MIT one.

It can be simpler in many dimensions, but I think it should be more complex in some others, in bringing in some price-adjusting mechanisms. I have not yet had a chance to see the Nordhaus-Tobin model.

You say no to the MIT model, on what grounds exactly?

Really on the ground that the basic argument that leads to the conclusion we must stop growth is simply wrong, intellectually unconvincing. Once we say the catastrophe isn't inevitably at hand, we have to ask

1 See conversation no. 19.

[p. 69]

what are the alternative choices. As I said earlier, if one looks first and foremost at the question of poverty in two-thirds of the world's population, it is impossible to see how that poverty can be alleviated without continuing economic growth.

In saying this I don't want to suggest either of two things: One, that the patterns of growth must be the same as in the past; perhaps they should change. Two, I don't want to suggest that the problems of pollution and of resource exhaustion are not real, aren't problems to which attention has to be paid; I think the problems of pollution both in the industrialized countries and in the industrializing countries are very urgent problems.

We haven't been doing the right thing about them. We are only beginning to learn what to do about them. It's quite important that we should do so. Part of the problem of pollution is to think of the choice of technology in ways which will make pollution less. Part of the problem is again to use the very simple advantages of the price mechanism and make polluters pay for the damage they inflict on society in general. If we make them pay for it, they will have strong economic incentives to minimize it. Under our present system, producing pollution, so to speak, is free; and since it is free, there are no economic incentives to avoid doing it. Similarly there is no question that there is a world population problem.

However, the discussion in the little book, Limits to Growth, of that problem is very shallow. There is no analysis, no distinction between desired family size and achieved family size. There is no discussion of what determines the capacity of the achieved desired family size, on the one hand, and what determines desired family size, on the other. My own judgment is that no matter what we do in the sphere of spreading knowledge of birth-control methods and pointing out to people the desirability of having smaller families - we will have to make people aware of the fact that with the higher survival rate, especially the lowering of the death rate, complete family composition will be attainable with a much smaller number of births and that the world population will increase very shortly in the next thirty years. We have to think more of ways of adapting to this increased standard than trying to wish that it didn't increase.

This striving for a global computer model in order to inventorize this planet's resources and problems, would you be basically in favor of such a thing?

Well, I think this is a question like other questions of intellectual enter-

[p. 70]

prise. They are things that people have to decide for themselves. If somebody thinks this is a very good idea and feels he knows how to do it, let him go ahead and do it. My own taste would not run in that direction. I think there is such an enormous information gap in many areas that the model as such will be really very crude, therefore the conclusions that one can draw from it are very slender. Furthermore, I think there is a great deal of computer mysticism which is terribly unnecessary. Many of the main ideas in the MIT model could be dealt with without elaborate computation. Elaborate computation is important when a model contains a great deal of detailed information and when it makes a real difference when a number has one magnitude rather than another. Given both the lack of knowledge of the structure of the relations involved in the MIT model and the very, very crude quantitative dimensions, it seems to me that there is a tremendous overemphasis on the role of computation.

But it was really meant as a tiny first step -

Yes, it was meant as a tiny first step, and perhaps an interesting first step. I have been more interested than many critics in the model. Perhaps the criticism will help the model makers make a better second step -

That's why I am here.

- and that I think is always a desirable direction. There is, let me repeat, the element of computer mysticism. Calculations made by the computer are not by that fact more correct than the calculations made by the man with pencil and paper or a calculation machine. When a model gets sufficiently complicated in the sense of having a right number of equations or relations in it, it is necessary to use a computer. I think it's all a matter of judgment.

Do you see society evolve towards a planetary management on a scale like IBM or General Motors? Do you think it will be desirable to move more towards a kind of practical pragmatic management of this planet?

I cannot see that far ahead. I would guess myself that some things will be managed on a large scale and some things will be managed on a smaller scale. As an American I have a kind of a Jeffersonian prejudice to believe that things should be managed on as small a scale as possible.

[p. 71]

A look at the current world scene does not suggest that we are moving toward more political unity, more political homogeneity with any great speed.

But could scientists influence politics toward, I would say, this most necessary kind of world management?

Let me perhaps answer this question indirectly by saying the most powerful political force in the world today seems to be nationalism, not rational calculations. I suspect we will have to contend with that for some long period of time. Very few things in the political world happen because the logical arguments show that they would be good things to do. I don't want to underprize the role of reason, in which I believe ultimately; sometimes it takes a very long time for the arguments of reason to prevail.