Washington’s Dangerous New Consensus on China
Don’t Start Another Cold War
"The Limits to Growth" is a brief, forceful, easily read polemic which has already generated many times its own weight in enthusiastic encomia and equally strong condemnations.[i] It advances a familiar, indeed fashionable, thesis. The goals and institutions of our present world society stimulate population growth and production increase at a rate that cannot be sustained. Further, and perhaps less familiarly, we are now about a generation from the point of no return, after which the world must suffer a catastrophic drop in numbers and wealth, no matter what is then done to restrain further growth. The argument is presented with a sufficient panoply of graphs, flow diagrams, references to the World Model and the new discipline of System Dynamics, and invocations of the computer to produce an aura of scientific authority for the conclusions. They have the additional weight of the endorsement of a prestigious private international group of respected businessmen, officials and academics, The Club of Rome, in a commentary appended to the study and signed by its executive committee. It is my contention that the authors' analysis is gravely deficient and many of their strongest and most striking conclusions unwarranted. None the less, it draws attention to a number of difficult and important problems which must be faced, including the question of whether its whole approach is helpful or harmful in dealing with these real problems.
The backbone of the argument of "Limits" is simple, and requires little elaborate intellectual machinery to develop. Many significant variables that characterize our global society, in particular population and industrial production, have been growing exponentially over the last century, that is, at a constant percentage rate, and thus showing a greater and greater absolute increment each year. The processes that determine this persistent growth at constant (roughly) percentage rates lie deep in the structure of our social order, and unless we deliberately make drastic changes in it, they may be expected to persist and continue to generate exponential growth in the future. Many important physical aspects of the world, however, are finite, and their finiteness implies that exponential growth cannot go on indefinitely, without, so to speak, bumping into the limits. In particular, supplies of cultivable land, reserves of mineral resources and the capacity of the earth to "absorb" pollution are finite, and one or another of these (or some combination of them) sets a ceiling level for population and industrial output.
What is more important, when one of the exponentially growing variables reaches the ceiling, it does not simply remain at the limit value, but rather moves sharply down to a much lower level in a process of catastrophic decline. Thus when industrial production, for example, reaches a ceiling level set by limits on mineral resources, it does not simply remain there but plunges from a wealth- to a poverty-level in a short space of time. It is this proposition, together with some of the characteristic time dimensions of the process that both constitute the core of novelty in the book and justify its urgent call for rapid and drastic action.
This characteristic sharp shift from growth to decline in turn reflects two features of the formal model which underly the computations and arguments presented in the book.[ii] The first is that the several variables and limits are all interrelated in a system in which growth in each of the main variables is reinforced by growth in the others. The second is that changes in some elements of the system have their effects on others only after a long lag. Thus, for example, a fall in the birth rate affects the demand for food fully only after a lag determined by the average length of life.
The question of how the system behaves when it reaches or approaches a limit is the central question of interest, and it is worth repeating that the kind of behavior which the authors find characteristic of their system is what gives their argument both its interest and its compelling quality. The fact that some limits exist, that the earth is in principle finite, is hard to deny, but does not in itself lead to any very interesting conclusions. Examples of growth systems are known that display quite different behavior as they approach their natural limits than the sharp reversals portrayed in "Limits." For instance, a system in which the rate of growth of the major variables was proportional to their distance from their limits would show a smooth, gradual, stable adaptation to its growth ceiling.
Further, the response times of the system the authors present to changes in some of the key variables are such that we must anticipate the possibility of castastrophe by half a generation or more, in order to have time to act and avert it. By the time we see the whites of their eyes, our guns will no longer fire. Thus the book's chief conclusion, endorsed by its sponsors in The Club of Rome, is that we must planfully, radically reorganize the fundamental institutions of our social world soon or face an unmanageable crisis not so late. To do so, we must now recognize the need, and begin to devise the means.
The analysis supporting these conclusions is unconvincing. It contains at least three kinds of flaws, each of which alone would justify a skeptical view of the result. Further, the first two are deficiencies of principle, which operate at the same level of simplification, approximation and qualitative generality that the authors attribute to their analysis. The most important question concerns the nature of the limits that enforce the growth ceiling in the model. Basically, there are two: arable land and the supply of exhaustible minerals. The first operates primarily on population, the second on industrial production. In order to demonstrate the ineluctability of the limits, and unimportance of the precise magnitudes assigned to them, the authors show that doubling the productivity of agricultural land, or doubling the reserves of natural resources, leads to no qualitative change in the behavior of the system, and only a relatively brief postponement of the moment of catastrophe. Pollution operates as a limit too, but somewhat more indirectly, through its effect on length of life and thus on population. Making pollution control more effective is seen as possible only with sharply increasing costs; thus an economic limit is built into the model in respect to pollution control that functions in the same way as the physical limits on agricultural land and mineral resources. The various alternative assumptions the authors work into the model always rely on one or more of these limits to bring about the characteristic crisis of the system. Even the variant of the model described as "utilizing a technological policy in every sector of the world model to circumvent in some way the various limits to growth" (p. 141) in fact incorporates all three limits-though they operate in a more distant future than in other variants, and the onset of catastrophic decline in population occurs only at the end of the twenty-first century.[iii]
The notion that such limits must exist gains plausibility from the use of physical terms to indicate the relevant quantities-acres of arable land, tons of chrome ore reserves-implicitly invoking the physical finiteness of the earth as the ultimate bound. But this is fundamentally misleading. Resources are properly measured in economic, not physical, terms. New land can be created by new investment, as when arid lands are irrigated, swamps drained, forests cleared. Similarly, new mineral resources can be created by investment in exploration and discovery. These processes of adding to the supplies of "fixed" resources have been going on steadily throughout human history. Indeed, the authors themselves in effect recognize this when they describe the pollution limit not in physical terms, but in terms of the increasing costs of achieving higher and higher degrees of pollution control.
However, once the problem is recognized as one of cost limits, not physical limits, it appears in a different light. The force of rising costs as mines go deeper or exploit thinner veins, or as drier and more distant lands need more water brought from farther sources and the like, meets the force of advancing technology, which brings down the costs of using existing resources and literally creates new resources by bringing within the bounds of cost feasibility materials or methods which formerly lay outside it. Thus, for example, the Hall process for reducing aluminum oxide by bringing the costs of the metal down to a level that made it an industrially usable material rather than a jeweller's curiosity, literally added hundreds of millions of tons to our reserves of metal ores. New ways of locating oil pools and new ways of exploiting them have combined to keep oil reserves- measured in terms of annual consumption-about constant over the past generation, though the actual rate of consumption has been growing exponentially. In general, the relative prices of mineral raw materials and agricultural products have not been rising, and the share of minerals (even allowing for imports) and agricultural output in total production have been falling fairly steadily over a long period in the United States. This is also true in other developed countries for which we have good evidence. While comparably good quantitative evidence for the whole world is not available, and such evidence as there is has not been assembled and analyzed, the best guess is that for the world as a whole, the share of extractive industries in output has been falling over the long period.[iv]
In sum, the advance of technology, like the growth of population and industry, has also been proceeding exponentially. In the United States- again the society for which the best data are available over a long period of time-the average annual rate of technological growth over the last half- century for the private economy as a whole has been in the neighborhood of two percent. Broadly speaking, this means that a representative bundle of inputs-labor, capital, raw materials, land-of constant value (in constant prices) will each year yield two percent more output than the year before. As "Limits" points out in urging the force of exponential growth, a two percent annual growth rate corresponds to a 35-year doubling time. Thus, technical progress over the life of a generation has made it possible for our children to get twice as much output from the same bundle of inputs as their parents. There is even some evidence that the rate of technological advance in the United States has speeded up in recent years, but it is not conclusive. Other industrial countries also show exponential growth in technology; some, such as Germany and Japan in recent years, at higher rates than the United States but the data pertaining to them cover only a short recent period.
Once an exponentially improving technology is admitted into the model, along with exponentially growing population and production, the nature of its outcomes changes sharply. The inevitability of crisis when a limit is reached disappears, since the "limits" themselves are no longer fixed, but grow exponentially too. The qualitative character of the results then depends on the fine details of the model, and, in particular, on the differences between the growth rates of the most important variables. Catastrophes need no longer be the rule, and more stable outcomes, in particular continuing growth at low rates, now become possible.
The second major flaw in the authors' analysis lies in the total absence of adjustment mechanisms of any kind in the model. Certain behavioral relations among the major variables are laid down, the magnitudes of their parameters determined by average behavior over the past, and then the relations projected unchangingly into the future. That is not how real social mechanisms work. Especially in the workings of the economy, adjustment mechanisms play a crucial role. The most important of these is price: as a resource becomes scarce, the consequent rise in price leads to savings in use, to efforts to increase supply, and to technical innovation to offset the scarcity. All economists know that these adjustment mechanisms are far from perfect and smoothly functioning. Yet they are and have historically been sufficiently powerful to mediate very large shifts in use of resources location, of population and patterns of consumption. Prices play no significant role in the basic logical structure that supports the argument of "Limits," although it is precisely their function to make smooth transitions possible as scarcities and demands change. Their absence is not unrelated to the characteristically unstable responses the model system of "Limits" displays. Only the effort of constructing another and much more complex model could show in detail what kind of stabilizing influence the incorporation of price changes and responses to them would exert. It is, however, well known that dynamic models structurally similar to those employed in "Limits," that characteristically display various forms of unstable behavior in the absence of prices as variables, are stabilized by the incorporation of prices and normal responses to price changes.
The third defect of the analysis is of a quite different order, one of detail rather than of principle. It is simply the failure of the authors to use available knowledge fully, effectively, or in some cases, at all. No one detail is of great importance, but together, they weaken seriously the claim of the work to respect. The most important single example is the authors' treatment of the determinants of population growth. Nowhere in their discussion do they acknowledge the great fact of demographic history in the Western world: the adjustment of birth rates to death rates. Our understanding of this "demographic transition" is far from complete; even if the underdeveloped countries repeated the same pattern over the same (relative) time period, they and the world would not be free of appropriate concern over the magnitude of population growth. But what should we think of a model of a process in which population growth plays a crucial role that simply ignores this central, elementary and familiar fact? Or to take another example of much less significance to the central argument, the discussion of equality and economic growth (p. 42-44) closes with an italicized warning that "the process of economic growth, as it is occurring today, is inexorably widening the absolute gap between the rich and the poor nations of the world." The "absolute gap," i.e. the difference in dollars between average per capita income in the United States and, say, Peru, is growing and, given their present levels, will probably continue to do so for a very long time. But is that interesting or important? The relative gap between average income in many of the poorer countries and the industrial West is narrowing, and that is what is relevant to the question of equality. Economic history shows that, after the early stages of urbanization and the development of commerce, economic growth has tended to greater equality of incomes, both within nations and between them. A complete syllabus of errors would be tediously long; perhaps the length of the list is the natural result of the process of reinventing economics, demography and much else as System Dynamics.
So much for the analysis. Can the major conclusion stand alone on its intuitive (or counter-intuitive?) merits without the analytic underpinnings? Is there merit in the proposition that we must seek now to move as rapidly as possible to the state of "global equilibrium" defined by stability of both population and capital, and that failure to do so invites catastrophe? After all, this proposition is now frequently advanced on the basis of much simpler arguments than those we have examined. Briefly, and simply, the answer is "No." There are no credible reasons for believing that the world as a whole cannot maintain a fairly high rate of economic growth (though not necessarily the present one) over a long period of time into the future. Further, if it becomes necessary, for whatever reason, to slow down the growth rate, a relatively smooth transition from higher to lower rates will be perfectly possible, and not achievable only through the mechanism of catastrophe. Moreover, whatever is done to slow down the rate of population growth, population will continue to grow, especially in the poorer countries, for a long time. Only art increased rate of economic growth in those countries will make it at all possible for them to deal with their unavoidable population increases without catastrophe. The large poor countries contain in aggregate a substantial share of the world's people, and thus increased growth for them will have some reflection in world totals. Further, it is difficult or even impossible to conceive of continued substantial economic growth in the poor countries in general taking place in a context of economic stagnation in the industrialized world. Thus, seen both in terms of need and of feasibility, the prospect for the foreseeable future is continued long-term economic growth, perhaps at rates lower than those currently observed, and with quite a different distribution of rates as among countries.
In the legend, there were in the end, real wolves. In the world today, there are real and difficult problems attendant on economic growth as we now experience it. The social-economic system is not self-correcting or self-managing; sustained, self-conscious efforts are necessary to deal with the problems, and they often must be maintained against strong resistance. Two of the authors' three central concerns, population growth and pollution, do indeed present genuinely urgent and difficult problems. A third equally important and difficult one, mentioned in "Limits," but only in passing, is the assessment of the indirect consequences of technical change, the unanticipated "side effects" that can sometimes outweigh the benefits. Present social mechanisms are not adequate for coping with any of the three, and the kinds of changes required to do so more effectively meet strong opposition at every level, from that of the individual family to organized interest groups and governments. From one point of view, all three problems can be seen as examples of "external effects," where costs and benefits of particular actions are not borne by the primary actors and thus fall outside the reach of the price system as it usually functions and the control of the incentives and adjustment mechanisms it provides.
In each case, the problem is to find a set of supplementary adjustment mechanisms and incentive systems which can guide the relevant actors to socially more desirable choices, a proposition easy to state in the abstract and difficult to realize in the concrete. In many situations we lack knowledge of the likely consequences of specific actions; in many, those who benefit from present arrangements or think they do resist change, while those who might benefit from change may lack both knowledge and power. In many situations we lack reliable indicators of what is desirable in an overall sense, and the machinery for resolving conflicting judgments is inadequate. Determined effort to deal with these problems is important. Failure to pay proper attention to them might well result in serious troubles, though they are unlikely to be of a kind which can properly be termed catastrophic. And, though there is widespread discussion of many of these problems and considerable social effort at dealing with some of them, it can be plausibly asserted that it falls far short of what is required.
Finally, therefore, how much does "crying Wolf" help to direct social energies toward improving our responses to these problems? In principle, it is not only useful, but indispensable. The social mechanism is made up of human beings moved by passion far more than by reason. The mobilization of feeling that is the necessary prelude to all but the most routine social action requires some stimulus stronger than a sound argument. But to be effective, the cry must be well directed: the wolves must be imminent and they must indeed be wolves. On this score we can give only a moderate grade to "Limits," or more properly, to its sponsors in The Club of Rome. The problems they call us to attend are real and pressing. But none are of the degree of immediacy that can rightly command the urgency they feel. Indeed, at least two problems of worldwide consequence outside the scope of this work seem to be more urgent than any it deals with: the creation of an international order stable enough to remove the threat of nuclear war, and the diminution of the staggering inequalities in the international distribution of wealth. A good sentry does not cry up tomorrow's wolves and ignore today's tigers.
[i] "The Limits to Growth," by D. H. Meadows, D. L. Meadows, J. Randers and W. W. Behrens III. New York: Universe Books (A Potomac Associates Book), 1972.
[ii] The details of the model are not given in the present volume, but are developed in a series of technical papers listed in its appendix, and in the book, "World Dynamics," by Jay Forrester (Cambridge, Mass.: Wright- Allen Press, 1971). Forrester is the intellectual father of System Dynamics.
[iii] The plot of this model (fig. 42, p. 140) shows an inexplicable and incredible rise in food consumption per capita, although its timing does not suggest that the population has overeaten to the point of extinction.
[iv] At this point, the reader probably feels uneasily that there must be some flaw in the argument. Surely the earth is finite, and even the wonders of technology must have some limit. The earth is finite, to be sure, and without broaching the larger question of whether the universe is or is not, it can be shown that the finiteness of the earth does not in itself set limits to what technology might accomplish that are relevant to the time horizons of the kind of argument with which we are concerned. I owe to Professor Robert Socolow of Princeton University a calculation that shows that in terms of physical limits alone, i.e. available matter and energy, the earth could support a population at least 1,000 times the present one at the current U.S. per capita income level.