On Growth

19. William D. Nordhaus

Professor Nordhaus was born in Albuquerque, New Mexico, in 1941. He studied economics at Yale University and obtained his Ph.D. at the Massachusetts Institute of Technology. At present he is associate professor of economics at Yale University and staff member of the Cowles Foundation for Research in Economics, as well as on the staff of the National Bureau of Economic Research.
Professor Nordhaus published Invention, Growth and Welfare: A Theoretical Treatment of Technological Change (1969).

Why are economists in the United States - contrary to, for instance, Tinbergen¹ in Rotterdam - hacking The Limits to Growth so ferociously to pieces?

1 See conversation no. 3.

[p. 120]

I think the contribution of the MIT team (Forrester, Meadows, et al.) is that they attempted to make explicit and quantitative what has been impressionistic and qualitative in the public's mind up to now. They are really the first ones to set out a global model (as in World Dynamics)² and make quantitative predictions. They predict that by the year 2050, the standard of living will be lower than it is now. By the year 2070, we will get either a pollution catastrophe or everybody will starve. More than anything else, it is the concreteness that upset and disturbed the public.

Here is where economists have had a violent allergic reaction. It is not that social scientists deny that population, pollution, and resources are important problems. These are clearly among the most difficult issues of theory and policy in political economy. It is the approach rather than the subject of Limits that has been criticized. A close look at the model has led many competent independent analysts to conclude that the underlying assumptions are founded in pure fantasy. World Dynamics did not refer to a single scientific study. Rather it is the computer saying that in the year 2050 the world is going to come to an end, and the computer somehow lends a dignity that would be lacking if someone just said that in a hundred years we will all be dead. Nobody would listen to a foolish man, but we worship at the temple of the foolish computer.

An analogy might be helpful. In this country we have the serious medical problem of cancer. Everyone recognizes that this is an extremely serious medical problem. I compare the Limits to Growth group to a group of people who suddenly claim to come up with the cure to cancer. When one asks them, ‘How do you know it is curable?’ they reply, ‘Well, we have some research going, we'll let you know about it. We have a computer. We put a model of the disease in the computer, and the computer says that cancer and heart disease will be wiped out within fifty years, you can see the computer output right here.’ Then one asks them for information. ‘We don't have any data, we didn't do any research really, we just have a theoretical model.’ Where did they get the model? ‘We just thought about the problem.’

This is the scientific approach behind World Dynamics and the Limits to Growth, but it is an approach quite at variance with Western traditions of scientific inquiry. The methodology that was used by the MIT team is that when you have a situation that you are trying to evaluate, you sit down in an armchair without any knowledge of the situation except what common sense tells you. Let me give a crucial specific example in both the Forrester and Meadows books. It is assumed that as per capita

2 See bibliography, conversation no. 34.

[p. 121]

income or per capita consumption rises, the rate of population growth increases. There is then an iron law: The only long-run state is one with everybody on the borderline of starvation. This iron law has been known, and periodically rediscovered, since the time of Malthus. It has been rediscovered again by the Limits people. They have been aptly called the new Malthusians.

The point is that almost every empirical study that has ever been made found exactly the contrary, that in fact population growth declines with affluence. Now, if in fact population growth declines with affluence, not the other way around, then the results of both the Forrester and, I think, the Meadows model are completely different. Population stabilization or decline - such as the US is approaching - will alleviate all the growth pains in the Limits model. Unfortunately, Forrester's hypothetical model did not in fact refer to empirical studies that have been done in the different fields. As a result, the conclusions of the model are unsupported by empirical evidence.

The point is that over the last few hundred years there's developed a methodology in scientific research. In medical research you do experiments, you do controlled experiments usually with animal subjects before you claim that a cure for a certain disease has been found. In the social sciences you discuss the matter before you proclaim a theory to be true about the world. It is customary to look at some data, to look at the experience of the countries involved. Forrester did not do that in his study. There was no attempt to try to verify it, and as a result, it is not at all clear what this model refers to. It could refer to an ant colony. It could refer to life on Mars. It could refer to fruit fly populations in the Amazon. But until any of these have either been verified or refuted, the model refers to nothing.

Let's now look at some particular shortcomings of the model. I am going to be referring mainly to World Dynamics because that is the only model which has been made generally available. There are three basic problems. The first one I've referred to already, is the problem of population. Forrester assumes that population growth rises with affluence, while in fact it declines. In his model if you change only that assumption, then the results are entirely different.

Really entirely different?

Absolutely different. For example, in some work I did in a review of his book, I used a simplified version of his model. His model is very

[p. 122]

complicated. It's virtually impossible to understand - I'll talk about that in a second. I used a simplified model to try to perform some sensitivity analysis to see whether, if you changed a few of the assumptions toward greater realism (as in the case of population growth), the results change. And the one thing that came out loud and clear is that if, instead of population growth at a very rapid rate, it grows something like either zero population growth, or a small decline of population growth, then in fact you have a greater affluence than we have ever seen before and greater economic growth than has ever been possible.

A second assumption in the Forrester model is the assumption that there is no technological change.

No exponential growth in technology -

Forrester shows no technological growth at all. How much output can you get per unit input? Inputs being labor, capital, land, and natural resources. Past studies by John Kendrick and Edward Denison find something like the following: For the last hundred years the amount of output you get per unit input has been rising between one and a half and three percent. In the Forrester and the Meadows study, it is assumed that if input stays the same, there is no way you can increase output. There are no inventions. Nobody invents the computer. Nobody invents fission processes for energy. Nobody invents more efficient engines. There is no solar energy. There are no jet planes. There are no tape recorders.

But it was established that technology increases slower than population or economic growth.

We cannot really say for sure what the critical rate of growth of technology is that will make the system viable. That's the first thing. Another is that we must recognize that there are some bad aspects as well as some good aspects about technological change. Many people feel, for example, that a lot of environmental problems today are due to certain kinds of technological change. Especially those that have occurred in the chemical industries, like the growth of new inorganic fertilizers. Technology is not an unmixed blessing. Most people feel that when you net out the costs from the gains, the technological change is a plus. I suppose all you have to do is to go and travel around the world to see the difference between high and low technology and see what material benefits and costs that has brought forth.

[p. 123]

There is a third area that I have found of particular concern and that is Forrester's view that human society is completely mechanistic, that people and institutions show no kind of adaptive behavior. When economists talk about adaptive behavior, they think of the use of the price system. For example, in Forrester's book you run out of natural resources. What happens to utilization of natural resources? Nothing happens. Prices don't go up; people don't get upset; firms don't try to economize, or recycle resources. People are just dumb lemmings and they just keep going over the cliff one after another without a thought. Just as there are no prices in the model, there is no political system in the model either. Prices can certainly fail us, as they have up to now for our atmospheric resources. Perhaps we can rely on our political system to help us out. But there is no political system in the Limits which is going to be of any use.

The Soviets have the same problem.

The problem of clear air and clean water is one which we've known about for many years. We cannot expect the price system to solve it for us, because these are public goods. On the other hand, if resources are appropriable (either by the state in a socialist society or by individuals in a private-property society), I think there is less cause for concern. For appropriable resources like coal and ores and oil, the storm signals of increased scarcity manifest themselves in the form of increases in price. We see these signals in areas of natural resources today, with price rises which are very large. The whole point about adaptive behavior in economic systems and political systems is that it allows the system to respond to scarcities in one area by moving to substituting plentiful materials and plentiful resources.

The reason this is so terribly important stems from another assumption in the Limits model, namely that there is a fixed pool of natural resources in the world. In the Forrester model there are four hundred years' worth of natural resources at the present consumption rate. In fact, there are not four hundred years of certain things. Natural gas is something that is extremely scarce right now. Petroleum is something that will probably run out long before four hundred years if we continue to use it at the current rate. But on the other hand, there are other materials which are virtually infinite in supply, like basic materials for several of the fusion processes; or like the inputs for some plastics. We are going to run out of something else long before we run out of energy if the fusion processes

[p. 124]

come along. This is an example of the misplaced concreteness in Limits to Growth.

Do you think there is a leeway for some economists like you to get together with the MIT team and look at some of these trends together? Why couldn't the Forrester model still be the basis for further study, as Tinbergen is doing in Rotterdam? Or could the work be carried out without a model of that kind?

You asked a number of questions. If you ask me whether there is room for collaboration between ecologists and economists and scientists and engineers, I certainly think there is and there has been a lot of collaboration in the past. If you are asking me, Is the Forrester model or the Meadows model the model that I would use to analyze the world economy, even as a first approximate, the answer is definitely not. Do I think systems analysis is the right way? If you mean by systems analysis building a quantitative, economic, social, political, demographic model, certainly that is what we have been building all along. But if you mean, Am I going to build a model in which I make assumptions without looking at data, a model in which I posit relations without talking to knowledgeable experts, a model with one hundred equations none of which have been confirmed by any empirical study, a model in which I am seriously going to predict what will happen a hundred or two hundred years from now, I think the answer to all those questions is no.

What has been your way of studying these problems?

Some time ago in fact, James Tobin³ and I worried about the problems of population, natural resources and economic welfare. We had a couple of models of this problem in which we tried to analyze such things as the factor of population growth, the role of fewer natural resources, the importance of a slower pace of technological change on future growth of consumption, all for the United States. This was only a very preliminary exercise. For the most part they were one or two equation models. We had data for the United States from which we fitted the functions that we used, and we did quite a bit of sensitivity analysis as to the proper form of the function. But the basic difference was that these were extremely small models, because one of the things you learn when you use these models is that when you use, as Forrester did, a model with forty-three nonlinear relations and a number of other variables, that you don't have a prayer of understanding what you've built.

3 A colleague of Dr. Nordhaus at Yale University.

[p. 125]

I think that the experience in economics has been that although big models look fancy and impress people, they're very difficult to build, very difficult to coordinate, and they don't really predict well. So as a methodological principle, I like to stay with simple models, so that I can understand what's going on. The second thing that I think is terribly important is that we continue to apply some of the work that has been going on in traditional economics and ecology to the global perspective. One of my objections to the work of the MIT team was their complete disregard of all past science. The new Malthusians profess to protect the earth and environment from the ravages of modern technology and desire to return to more traditional patterns of living. It is ironic that in returning to traditional economic standards they eschew the precious scientific tradition which has been so carefully established over the past four centuries.