Knowledge that Transforms

Two observations are in order before I discuss the topic suggested by the title. First, interest in it forces us to be future oriented, and oriented towards the longerterm future at that. This interest, while arising from a variety of sources, generally involves predictions of growing scarcity of natural resources and environmental carrying capacity, worries about increasing dependence on foreign supplies and concerns about loss of local amenities due to the rapid influx of population into limited regions. It is not sufficient to study past trends when the overriding concern is that the future will not be like the past. Otherwise, we are subject to the criticism of those like E. F. Schumacher who recently suggested that economists spend most of their time optimizing the arrangement of deck chairs on the Titanic. Second, we cannot discuss the effects of population independent of other factors that can cause resource and environmental pressures. While it makes sense to ask about the relative importance of population in causing these pressures, we cannot completely ignore tastes, technology, institutions, policies, and international relations. These and other determinants strongly condition the resource and amenity pressures generated by population growth; they often change and affect the situation more than do demographic changes within specific time periods; and they often interact with each other and with demographic variables. In particular, we cannot set aside changes in per capita output since so many of the problems we are concerned with here are related to the movement of materials through the total economy. These two considerations have led me to focus on three questions: Are there limits to economic imposed by resource exhaustion and environmental pressures? If so, how close are we to these limits? How important is population as opposed to other factors in moving us towards these limits? In discussing these questions I will consider primarily the United States. The first question has significance only if the word growth is taken to mean material economic or in population and material economic output, but not in GNP or national income. Since the latter are artifacts in large part measuring the value placed on the goods and services produced by an economy, they could conceivably continue growing even without any in the amount of materials embodied in these goods and services. Defined in this way, is quite definitely subject to limits. First, there are ultimate or natural limits imposed by the finite character of the earth. Ruling out the possibility of importing materials and energy from outside the earth on an everincreasing scale, the second law of thermodynamics, the entropy law, makes the existence of this limit quite certain (see Nicholas Georgescu-Roegen). Scientific * Program director, Resources for the Future, Inc.

How does the consumer react to uncertainty in his future inconme receipts from human and nonhuman wealth? In answering this question, it is helpful to replace his stochastic life cycle problem with a certaintyequivalent problem in which expectations of future income receipts are adjusted by the subtraction of risk premiums, as a manifestation of risk aversion on the part of the individual. A recent paper by Keizo Nagatani utilizes this idea. B1v supposing the consumer to maximize a lifetime utility function subject to an intertemporal budget constraint in which incomes are replaced by their certaintv equivalents, he shows that with the unfolding of his y-ears, the individual's planned consumption at anv' future date will shift up or down. Thus suppose the consumer is now aged v years and is planning consumption for age t>v. Then Cv(t) will in general differ from Ct(t). However, the risk premiums on uncertain future inconme (from hunman wealth) that Nagatani uses are arbitrarily imposed, or exogenotus to the maximizing problem. This seems wrong. Such premiiiums should reflect not onl the variability of future incomes, but also the individual's reaction to such variability; or in other words, the nature of his intertemporal utility function. Further, there is a prior question of uniqueness; are the risk premliums in each period uniquely defined? To see that this might prove potentially troublesome, imagine a decision problem: