Comparing Values In Environmental Policies: Moral Issues And Moral Arguments

National Research Council (US) Steering Committee on Valuing Health Risks, Costs, and Benefits for Environmental Decisions; P. Brett Hammond; Rob Coppock

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Research Council (US) Steering Committee on Valuing Health Risks, Costs, and Benefits for Environmental Decisions; Hammond PB, Coppock R, editors. Valuing Health Risks, Costs, and Benefits for Environmental Decision Making: Report of a Conference. Washington (DC): National Academies Press (US); 1990.

Cover of Valuing Health Risks, Costs, and Benefits for Environmental Decision Making

Valuing Health Risks, Costs, and Benefits for Environmental Decision Making: Report of a Conference.

Show details

< Prev Next >

5Comparing Values In Environmental Policies: Moral Issues And Moral Arguments

Douglas E. Maclean

Three kinds of problems seem particularly pervasive in administering environmental policy. The first is political. Policy makers work within a framework of environmental laws that are notoriously vague, apparently contradictory, and often otherwise flawed. Those charged with the responsibility to protect the environment must follow procedures that were designed in part to protect political interests; they must fight off relentless pressure while remaining accountable to politicians; and they must enact policies with an eye to withstanding the litigation that will surely follow. This process may be a model of democracy at work, but it does not encourage thoughtful responses to complex issues or regulatory improvement by "fine tuning" in the light of new information and greater experience.

The second problem in administering environmental policy is technical. Uncertainty is pervasive in this area, and the demand for precision is often greater than science or practicality permits. Under pressure to act now, decision makers grope for policies in ignorance, especially if the risks involved in various alternatives have latency periods between exposure and the onset of irreversible effects. The costs of always acting conservatively in the face of uncertainty can be prohibitively high; and it is not unusual for the uncertainties to be so great—ranging over six or seven orders or magnitude—that the risk analyses are useless to policy makers.

The third problem is one of comparability. It is impossible to eliminate environmental risks; policy makers must thus decide when to put more resources into reducing them, when to control other risks instead, and when to stop trying to reduce risks altogether and use resources for other purposes. To make these decisions, a broad range of values must be weighed and compared. What kind of a problem is this? Insofar as it concerns the allocation of scarce resources, it is a classic economics problem, a general problem in rational choice or decision theory. Why should it be singled out as an especially central problem for environmental policies? It is surely not as immediate a concern for the policy maker as the political and technical problems he or she faces. The reason to single out this problem for environmental policies has to do with the nature of the benefits or values that are involved.

Environmental decisions typically require the comparison of different benefits (e.g., the preservation of human life, health, clean air and water, wilderness, endangered species, money, and consumer products). The economics problem might be described as finding the correct weights or rates of exchange among these benefits. That our society considers each of them valuable is obvious, but it is far from clear that there is a satisfactory metric for weighing them together and trading them off. All of the general methods proposed so far seem notoriously controversial. The underlying issue is the nature and value of different benefits, which is not strictly an economic issue but rather a philosophical one. It is a subject of ethics.

Some of the benefits that must be weighed and compared are even more difficult to evaluate than those already mentioned. For example, it is taken for granted that public policies must be fair and must consider the distribution of benefits, risks, and costs across locations or populations and across time. An important goal of policies is to protect the health and environment of future generations. Yet risks can be distributed differently in various ways, all of which may seem morally relevant Some statistically certain number of annual deaths spread across a population may have to be compared with a small risk of catastrophic consequence. The average individual risks involved may be identical, but they seem to involve different values. How should they be compared? And how should these different distributions be compared with the other benefits and costs involved? These, too, are moral questions.

The philosophical issues involved in promulgating environmental policies, however, extend even beyond this. Many people believe it is simply wrong even to attempt to solve these measurement problems and to find a common and applicable rate of exchange between all of these benefits. One popular and persistent view holds that it is morally wrong to set forth the problem in this way. It is frequently alleged that the value of such benefits as human life or the nation's environmental heritage cannot be equated with money, or other resources, at any rate of exchange. To assign this kind of exchange value to such benefits is to treat them as commodities when they really have a different kind of value—a sacred value perhaps—and should be regarded as such. This objection suggests that it is necessary not only to find a method for weighing different goods but also to find appropriate ways of expressing or regarding different values. Part of the problem, it seems then, is procedural.

The philosophical issues at the core of this third problem are therefore quite basic. They involve not only the question of how to weigh different benefits but also the more general question of how to compare different values. The objections suggest that the benefits or values may be incommensurable, and yet decisions based on comparisons must be made.

This paper will address these moral issues, which are at the heart of many environmental controversies. Moral views drive debates over proposed legislation and are further reflected in subsequent litigation. The appropriateness of discounting the value of future lives, the application of benefit-cost analysis as a method of setting environmental policies more generally, and other issues that remain central and contentious in the environmental policy arena are essentially moral disputes.

Methods Of Reasoning About Morality

Most policy analysts would agree that the problems I have been describing are pervasive and at the core of many environmental disputes. My aim so far has been simply to suggest that many of these moral concerns raise a common general question about whether and how different kinds of benefits and values may be compared.

I wish to consider some of these moral issues directly, but it would be useful first to discuss briefly the nature of moral or philosophical reasoning. The approach I will take to these issues—the approach I find most natural and useful—is to inquire directly into the nature and implications of some of the values involved. What is the value of human life? What are current society's obligations to future generations? This process is not an empirical inquiry into what most people happen to believe but rather a normative inquiry. The question is, what is it reasonable to believe? A successful argument must appeal to what can be found to be reasonable and not merely what is persuasive. If different and incompatible views pass the test of reasonableness, then the moral inquiry turns to more general principles and procedures for making balanced or fair choices among these different individual views.

This kind of inquiry inevitably appeals to moral intuitions and to arguments that try to link such intuitions to other intuitions and to more general principles. Moral reasoning, as I would characterize it, is in this respect similar to scientific reasoning. Instead of reasoning back and forth between hypotheses and laws, on the one hand, and observations and experiments on the other, in moral reasoning, one reasons between cases and considered intuitions on one hand and more general concepts and beliefs on the other. Both kinds of reasoning are objective to the extent that their claims are susceptible to rational assessment.

One difference between scientific and moral reasoning, of course, is that moral principles are not meant to predict but to prescribe; but this difference is not crucial to the nature of the reasoning involved. If the nature of reasoning in science and ethics is similar, however, why is it so much harder to reach consensus in ethics than to reach consensus in science? One reason is that methods of reasoning are more highly developed and agreed upon in the sciences than they are in the field of ethics. Scientists can also design experiments explicitly to test hypotheses and resolve disputes. In ethics, one uses actual and hypothetical situations to test one's intuitions, but these "thought experiments" are much less likely to be convincing than are the experiments of science.

Such an approach to moral inquiry is common within the discipline of philosophy, but it appears to be alien and objectionable to policy analysts who are trained in other disciplines. They seem to favor other approaches to resolving fundamental moral issues. One of these approaches is empirical. The way to resolve conflicts that arise among analysts, it is sometimes thought, is to find out what people actually believe and prefer. This process will reveal what the nation's policies ought to be, for in democratic societies, the people are sovereign. I call this the low road to moral inquiry.

Another approach appeals to moral theories or general normative doctrines to resolve disputes and explain more particular values. Philosophers, of course, are also interested in moral theories and normative doctrines, but their interests often have very little to do with shedding light on particular moral or policy questions. Nevertheless, it is not uncommon to find policy analysts suggesting that this is the approach that philosophers might use most effectively to shed light on the moral aspects of environmental issues. According to this way of thinking, what is needed is a relatively comprehensive survey of the application of moral philosophy and ethics to problems of valuing risks to life and health, not a restricted set of specific arguments using a specific mode of philosophical analysis:

Whereas no one could fault a philosopher concluding with an argument in favor of one perspective (e.g., Kantian) over others (utilitarian, rights libertarian, contractarian, etc.), what seems essential is to lay out what the alternative approaches are, what specific ethical and moral issues they address, and where the approaches diverge, as well as how one might assess the relative applicability and relevance of each.¹

Let us call this the high road to moral inquiry.

The kind of direct moral inquiry I favor—using a restricted set of specific arguments—is neither merely empirical nor overly general or theoretical. It is a middle road. Let us briefly consider, then, the weaknesses in the high- and low-road approaches.²

Empirical Approaches To Moral Issues

Empirical approaches to these questions are usually attempts to uncover individual preferences. What do most people think are current society's obligations to future generations? How would most people trade off longer life expectancies against improved health, health against welfare, and so on? Revealed preference theory attempts to uncover these attitudes by looking at actual behavior and, in particular, at individual consumer choices in cases in which people are free and informed. The problem with this particular approach, of course, is that people are not always free or informed; even when they are, their consumer behavior does not always indicate their considered preferences, let alone their reflective moral judgments.

An alternative empirical approach is expressed preference theory or contingent valuation methods. These methods generally involve surveys in which people's preferences, their willingness to pay, and so on can be measured directly. Methodological difficulties aside (e.g., the reliability of surveys), a number of problems arise in applying the results of such surveys to settle moral disputes. The first is that people are normally asked in these studies what they would prefer, which is not at all the same as inquiring empirically into their moral beliefs. It is both consistent and common for people to have certain preferences that it would be wrong or unfair for policies to satisfy directly. Setting policies that attempt to satisfy the maximization of preferences would ensure, for example, that toxic wastes and hazardous technologies are always sited in the least populated areas simply because there are fewer people to object. Such a policy would obviously be unfair, and most people would agree that it is unfair, despite their personal preferences.

Recently, however, a number of contingent valuation studies have surveyed people's moral attitudes directly—in particular, their intuitions about procedural fairness and distributive justice (Kahneman et al., 1986). The information from these studies is interesting and perhaps quite useful, but there is a certain absurdity involved in thinking that these data can settle moral disputes. This absurdity is easy to show. Either moral truths can be reduced to individual preferences or they cannot. If they can be determine what is fair or right—is the wrong question. Either people will answer according to which alternative they think is morally justified or best supported by moral reasons (which by hypothesis rests on an incorrect view of morality) or they will respond by saying what they think other people believe is right (if they have the "correct" moral views). Actually, to avoid relying on the false moral view one step removed, they would have to respond by saying what they think other people think that other people think that—and so on. If moral truths can be reduced to individual preferences, then soliciting opinions must either be ultimately self-defeating (if people express what they believe is morally justified) or land in an infinite regress.³ If moral truths cannot be reduced to individual preferences, however, there is no apparent reason for surveying opinions to advance the understanding of moral truth and justification, except perhaps to help uncover reasons or arguments that might not be immediately apparent.

The foregoing is an argument against the low road to moral inquiry, but there are other reasons to respect popular sovereignty and in so doing uncover public opinion for use as a basis for policy decisions. One reason, of course, is that some comparisons may not essentially involve moral reasoning or moral problems. One might believe that moral philosophy really has little to say about how improvements in welfare should be traded off against improvements in health or life expectancy at the margin, when either improvement will fall to the same people (see Schelling, 1984). These choices may simply be matters of personal preference, but they are only some of the more troubling choices environmental decision makers must face. Many of the comparisons they make also involve distributional issues: improvements to one population must be balanced against risks to another, the price must be paid now to protect future generations, and so on. These, at least, are certainly moral issues.

Although some issues are simply matters of preference, there is a further argument for citizen or consumer sovereignty, based on the character of democracy, that applies to public policies. Marglin writes (1963:97), "Whatever else democratic theory may or may not imply, I consider it axiomatic that a democratic government reflects only the preferences of the individuals who are presently members of the body politic." The function of government, however, is not simply to reflect current preferences. Government also has an ennobling and educational role to play, even in a democracy. The U.S. Constitution protects many values, even those that are socially unpopular.

More specifically, it has frequently been suggested that government ought to have greater concern for the welfare of future generations than is expressed by individuals in their own choices (Pigou, 1932). Individuals is expressed by individuals in their own choices (Pigou, 1932). Individuals die, after all, but the society continues, and there can be no objection to the government looking after the interests of future generations. This goal is explicitly stated in the National Environmental Protection Act and so should be a central concern of the Environmental Protection Agency (EPA).

For these reasons, it is proper for environmental policy makers to take moral arguments seriously, even when those arguments run counter to popular opinions. Some might object to this degree of ''moralism" on the grounds that it is paternalistic. The argument against taking the low road, however, is moralistic only to the extent that it insists that moral conflicts be resolved through moral reasoning and not by empirical means. This approach is not necessarily paternalistic. Paternalism means restricting a person's freedom of choice or overriding personal preferences for the individual's own good. The argument against the low road does not call for replacing an individual's preferences for his or her own welfare with the decisions of "moral experts." It says only that moral decisions should be made that are justified by moral arguments.

Arguing From Theory or Basic Doctrines

Let us consider now the high-road approach to moral inquiry as it relates to environmental policies. This approach looks to moral theory or basic normative doctrines and attempts to apply them to particular policy questions.

Let us first consider libertarianism, contractarianism, and utilitarianism—which (as I indicated above) are sometimes thought to be different representative moral perspectives—in order to illustrate the difference between moral theories and basic normative doctrines. Libertarianism (or "rights libertarianism"), for example, usually appears in political theories that claim that individual rights are fundamental to determining political or moral obligations or claims. Even in the more comprehensive discussions of libertarianism (e.g., Nozick, 1974), libertarianism is at most a normative doctrine that is important in discussions of political theory.

Libertarianism is not a moral theory, became it does not take the kind of external view of normative claims that moral theories inevitably take. Discussions of libertarianism typically do not attempt to explain the foundation of its basic moral principles or relate them to other basic moral issues—for example, the nature of the subject matter of morality, the nature of moral reasoning, or the connection between moral principles and rational motivation or will. Moral theories address these more abstract issues, and they frequently are not very closely related to more particular normative claims or doctrines, unlike virtually all discussions of libertarianism.⁴

Contractarianism, in contrast, is a moral theory, at least as it has recently been developed (e.g., Rawls, 1971; Scanlon, 1982). Rawls's theory of justice as fairness, which is perhaps the best recent example of moral theory of any kind, does address these more fundamental and abstract issues. Although Rawls derives principles of justice (e.g., the principle of the priority of liberty, or the difference principle for justifying inequalities in the distribution of primary goods) that can be interpreted as normative doctrines, these principles are themselves intended to be applied only at a basic and rather abstract level. Rawls emphasizes that his principles of justice cannot be applied to immediate cases or particular policies.

It is important to emphasize this difference between moral theories and normative doctrines, because most moral theories appear to be compatible with many different moral principles or normative doctrines. A contractarian moral theorist, for example, might argue that a utilitarian principle is the normative doctrine at which a properly defined social contract would arrive.

In this respect, utilitarianism is relatively more complicated or confusing, because it has been defended both as a moral theory and as a normative doctrine or basic normative principle. As a moral theory, utilitarianism constitutes a particular view about the nature of moral goodness and moral justification or moral reasoning. Such a moral theory, however, may or may not prescribe the principle of utility as a normative principle. J.S. Mill, for example, did not regard utilitarianism as a moral principle that should be commonly adopted and applied directly to individual actions or to social policies. More recently, rule utilitarian and motive utilitarian theories have been suggested that defend other, commonsense moral principles and doctrines. Some of these theories even suggest that utilitarianism can be correct only as a moral theory, and that attempts to apply the principle of utility directly are likely to be self-defeating, from a utilitarian perspective (Scheffler, 1982).

Thus, if the high road means beginning with moral theory, it will require a long and philosophical journey before arriving—if at all—at recommendations for important policies and decisions. The high road would also require the resolution of disputes about which moral theory is correct. If these disputes are resolved, it will be in part only because one theory better explains all the moral phenomena, including more particular reflective moral judgments and the principles that unify them. In sum, it may be necessary to travel down what I am calling the middle road first in any case.

If following the high road means instead beginning with basic normative doctrines—for example, some basic utilitarian principle, or libertarianism, or some egalitarian principle, or even the difference principle (although this may go beyond anything that might be attributed to Rawls)—then the question is, which normative doctrine should be chosen? A further task is to determine what in particular each normative doctrine means. Does the principle of utility call for maximizing pleasure, as it did for Bentham, maximizing some more abstract good, as it did for Mill and G.E. Moore, or maximizing the satisfaction of preferences, as it has come to be interpreted frequently among economists? If one chooses libertarianism, what do rights entail, which rights are basic, and how should conflicts among rights be resolved?

Surely, the only plausible way to answer these questions is to move to the more abstract realm of moral theory and to the realm of concrete reflective moral judgments about cases, policies, and principles. If all of the moral doctrines imaginable imply the same result in some case, then that consensus would be the strongest moral argument available, although it is very unlikely that there will be many cases like this, and, where they exist, they will probably not be in dispute anyway. Every moral doctrine, for example, must imply that gratuitous torture is bad; otherwise, people would immediately reject the doctrine. In general, if policy makers reason carefully and critically about more concrete cases and policies, they will have more confidence in their judgments about them than in their judgments about more general and sweeping principles or doctrines. Again, the middle road seems the place to begin and, for many practical purposes, the only road that need be traveled.

Valuing And Discounting Lives In Environmental Policies

Having defended the approach to moral reasoning that I find most natural and useful to apply to moral issues in environmental decision making, I will now illustrate this approach by considering some of the prominent questions I described at the outset of this paper.

Policy makers are usually interested in analyzing the values of the consequences of alternative policies. It is well known that some analytic methods that quantify costs, benefits, and risks and that aim at maximizing net benefits have difficulties evaluating different distributions of these effects. These methods take a narrow view of consequences.

Because a broader view of consequences would include the distributional effects of alternative choices, a sensitive metric should also, in principle, be able to evaluate distractions. The problem is that distributions may be valued differently because of their effects. Distributional principles cannot simply be valued independent of the context of their application. Rather, different principles and weights must be applied to different decision problems. Distributional values do not challenge analytic principles per se, but they challenge the generality of their application.

In addition to valuing the consequences of policies, certain procedures or ways of making decisions are also chosen. Elections are valued as a way of selecting public officials, trials as a way of determining criminal guilt, random processes as a way of achieving fairness. Some procedures are valued intrinsically and not only for the instrumental reason that they are most likely to produce the best outcomes. (There is generally little reason to believe they will.)

Some would argue that procedures, like distributional effects, should also be regarded as part of the consequences of environmental policies (Keeney, 1984). This argument is a far more controversial extension of the concept of consequences. I can illustrate why this is problematic with a personal example. Not long ago, I suggested to my wife that, instead of buying her a birthday present this year, I would give her money and she could shop for her own present. I am not very good at picking out presents for her, and I do not enjoy shopping for them. I offered her $50, explaining that I would probably spend $25 for a present if I bought it, so $25 could be considered to cover the cost of her time. Besides, she enjoys shopping for gifts. My wife did not appreciate this suggestion. She valued the traditional procedure for increasing her stock of goods on her birthday, and she did not regard money as compensation for this loss.

If one insists that the value of procedures is commensurable with other consequences, then it might be difficult to regard procedures as having more than instrumental value. In any case, an evaluation of a procedure must include a full assessment of the effects of using it, which is more than an evaluation of its expected consequences.

Two issues are central to moral debates over environmental policies. The first is the social discount rate, which raises distributional issues about how to compare costs and benefits that are spread out in time. The central moral issue here is whether future lives should be valued differently than present lives. This question leads to the second issue, which is how to compare the value of saving lives or improving health with other values and the suggestion that the best way to do this is to assign monetary equivalents to all of these values. This solution raises the problem of putting monetary values on human life and health. The moral objection to this approach is that the value of human life is incommensurable with other economic values: one cannot put a price on human life. I will argue that when some of the confusion is removed from this objection, the moral core of the issue involves procedural values.

The Social Discount Rate

It is common in policy analysis to apply a discount rate to expected consequences as they occur further and further in the future. The reasons for discounting can appeal to the opportunity costs of capital, the reasons for wanting returns on investments sooner rather than later; or they can appeal to rates of time preference, the claim that people tend to care less about consumption in the future or about the remote effects of their actions and policies. These are very different kinds of justifications. Often, they are not distinguished, or they are not treated differently.

In a recent provocative study comparing the effectiveness of a large number of regulations, Morrall (1986) expresses a theoretically popular view about discounting. Explaining his own assessment techniques, he writes (p. 28):

For the sake of consistency, I adjusted these temporal variations using a uniform 10-percent discount rate for both benefits and costs.
Students of benefit-cost analysis will recognize an unavoidable imprecision in using a uniform discount rate, and a certain arbitrariness in using 10 percent rather than some other rate. Some regulatory costs displace investment and others displace consumption, and the two effects are not economically identical. Here as elsewhere, however, the analytical demands of tailoring a precise discount rate for each rule were impossibly large, and for comparative purposes the benefits of greater precision would have been small. Students of regulatory politics will recognize that discounting benefits as well as costs runs afoul of the policies of some regulatory agencies, not to mention the positions of some political representatives and op-ed writers. On this point my procedure is impeccable. Discounting costs but not benefits leads to absurd results, such as that a rule saving 100 lives a decade from now is more desirable than a rule of equal cost saving 99 lives right away, and that all rules yielding continuous benefits are worth any amount of immediate costs.

Morrall here suggests several different reasons why all benefits and costs should be discounted at the same rate.

Some of these issues arose recently with EPA's proposed asbestos rule because the detailed knowledge available about the health effects of asbestos allowed EPA to make relatively precise estimates of the latency periods between exposure and the onset of disease. Following its traditional practices, EPA proposed that it not discount the cancers averted for time, but the Office of Management and Budget (OMB) insisted on a 10-percent discount rate for all costs and benefits, which included discounting the latency periods between exposure to asbestos and the development of cancer. The issue is important for many other EPA concerns as well—for example, waste disposal regulations, an area in which much of the cost of regulating is borne initially but the possible costs of not regulating occur well into the future, and uranium mill tailing regulations, in which the yearly benefits of regulating are small but spread over a century or more.

Morrall does not explain why he chose 10 percent as the discount rate (10 percent is considered by many to be high), but in so choosing he follows recommendations urged by OMB since 1972 (OMB, 1972). The discount rate obviously has important political consequences. It has been the subject of considerable debate among practitioners and theorists who realize that the benefit-cost analyses of many development projects, as well as many regulatory proposals, are sensitive to small changes in the discount rate (Lind, 1982:1-18). I will try to disentangle some of the philosophical issues involved in discounting and treat them systematically.

Should Lives Be Discounted?

I have mentioned two justifications for discounting: opportunity costs and time preference. A third reason is uncertainty, which I will not consider here (see Parfit, 1983).

What justifies discounting lives saved or lost in the future? Clearly, lives do not have the properties of money: they cannot be invested or used to earn a rate of return. The idea that opportunity costs attach to the value of lives and not only to the resources invested in life-saving programs does not make sense.

The only justification for discounting lives would seem to be time preference or the discount rate on consumption. It must be this reason that leads Morrall to conclude that it is absurd to save 100 lives a decade from now rather than 99 lives right away. I wish to argue that it is not absurd to save more lives in the future than can be saved now, other things being equal, and that there are compelling moral reasons for doing so.

Suppose one knew that an action would produce two independent nonmonetary outcomes that have quite different values. The outcomes would occur several days or weeks apart, but it is not known which will occur first. Would that matter for an evaluation of the action? Except for considerations of anticipation and memory, which are additional, secondary effects, it would not. Mere differences in the timing of events do not matter.

The murder of an innocent 20-year-old woman is equally horrible if it happens today or tomorrow or 20 years from now. I take this to be nearly a self-evident truth. It is part of what it means to treat people as equals to regard these deaths as equally bad. I am not saying, however, that the murder of an innocent 20-year-old today or the murder of the same woman when she is 40 are equally bad. The difference in what she would lose is morally important. Yet when one compares the loss of a 20-year-old today to the loss of a similar 20-year-old in the future, the two losses are equivalent.

Other things being equal, therefore, if it is better to save 100 lives today than to save 99 lives today, then it is also better to save 100 lives in the future than to save 99 lives today. As a matter of psychological fact, many people might have a greater concern for the lives saved sooner, but this psychological fact is also morally insignificant. People suffer many prejudices, psychological biases, and weaknesses of imagination and will, but it is the point of moral reasoning to help them overcome these natural but unjustifiable inclinations. I believe Ramsey was right to dismiss these preferences as due merely to a lack of imagination (Ramsey, 1931). The situation might be different if there were other kinds of connections to the more proximate lives, for special relations can produce moral obligations. A person may be justified, for example, in devoting greater attention and resources to his or her own children than to strangers, but this position has nothing to do with time.

Policy theorists tend to be remarkably unpersuaded by this argument, except perhaps to regard the conclusion as a public attitude widely enough held that it must be taken into account. I will try, therefore, to respond to four objections to it.

Democracy And Consumer Sovereignty

Some have argued from the assumption of consumer sovereignty that government policies should include discount rates that reflect society's time preference. Earlier, I discussed the relevance of arguments about democracy to moral argument and so will add a few brief comments here.

Three different questions should be distinguished: (1) As an individual, am I morally justified in being less concerned about the welfare of people in the future than about the welfare of those alive today? (2) As a community, are people morally justified in being less concerned about the more remote effects of social policies than about the more immediate effects? (3) If most people are less concerned about the welfare of people in the future, ought the government to override this majority view?

Consumer sovereignty is relevant to the third question. The moral inquiry being conducted here focuses on the first two questions about the justification of people's values and normative beliefs (Parfit, 1983). The question, therefore, is whether consumer preferences or popular opinions constitute a promising starting point for moral argument or whether they should be taken by themselves to be a moral intuition with normative status.

The preferences and the more considered values of individuals may differ considerably. Individuals might show time preference in their consumer behavior and investment decisions but different values and a greater concern for the future in their political choices (Sagoff, 1986). Marglin (1963:97) considers this possibility to be schizophrenic and argues that, in any case, consumer behavior is a better indicator of a person's "real" preferences: "[S]ince deeds speak louder than words, one can argue that preferences revealed in the market place are more genuine and better considered."

I find it remarkable to dismiss political activity as merely words. I would guess that most people are more committed to their political beliefs as an expression of their values than to much of what they buy as consumers. Political activity takes many forms. It can consist of supporting environmental groups that lobby for preservation while consuming nonrenewable resources beyond one's strictest needs. It is not clear whether there is even an inconsistency involved in this pattern of behavior, but it is surely not schizophrenic. It might be true that the values people express in political contexts and the programs they support tend insufficiently to take economic considerations into account. People find it easy to support a worthy program when they do not know what it costs. Yet it is perhaps equally likely that people do not take their moral values sufficiently into account when they shop or invest. Even if there were inconsistencies between market-exhibited preferences and those that are politically revealed, it is far from obvious that they should all be resolved in one direction.

These issues demand a much fuller treatment than can be provided here. I am suggesting, however, that the connection between consumer sovereignty and democracy be viewed with some suspicion. Sovereignty in a democracy rests with the citizens—people—and people express their values in many ways.

Excessive Sacrifice

The remaining three objections do not challenge the moral argument directly but claim instead that certain absurdities follow if all benefits, including the value of human lives, are not discounted in benefit-cost analyses. Morrall points to one such absurdity when he suggests that "all rules yielding continuous benefits are worth any amount of immediate costs" (1986:28).

Situations in which this result might follow are familiar at EPA. For example, EPA's analysis of a uranium mill tailings standard estimated that the present costs would be $388 million and that the standard would save 4.9 lives per year perpetually. If lives were discounted at 10 percent, the present value would be $8 million per life saved, a figure that is probably too high to recommend. If lives are not discounted, the cost-effectiveness is $800,000 per life saved for a horizon of 100 years, $80,000 per life saved for a horizon of 1,000 years, and so on.

Even in this example, the result of not discounting lives is absurd only with the assumption of an infinite time horizon, and there are good reasons to avoid such an assumption. The uncertainties involved should probably lead to the rejection of even the analysis with the 1,000-year horizon; yet however a reasonable horizon is determined, there is no justification for discounting the values of the lives analysts believe can be saved.

If the suggestion of selecting a time horizon in such cases seems ad hoc, there is another way to avoid the absurdity, using a method that has a stronger conceptual foundation: limiting the sacrifice required of the present generation.

It is common to think that individuals have less responsibility to do good than to avoid doing harm because avoiding harm can often be accomplished by not doing anything, whereas doing good requires some effort. There are limits on the effort that can be demanded of people on behalf of morality. It would not be wrong for an individual to avoid saving even hundreds of lives if that act required the sacrifice of all other benefits in that person's own life. It would be extremely noble for a person to devote himself or herself to such a morally worthy cause, but society cannot require this great a sacrifice. Individual moral rights may be seen in part as an effort to protect individuals from the possibly excessive demands of morality. There are similar limits on the sacrifices that can be demanded of current generations for the benefit of future generations. Identifying these limits blocks the absurdity noted by Morrall. It should be no more difficult to identify a maximum level of acceptable sacrifice for a generation than it is to determine an optimal social rate of savings.

Indefinite Delay

Let us consider the following example.⁵ Certain social resources are available and can be used for either of two projects and in no other way. The projects have the following potential results:

Project A will reduce by 100 the number of fatal accidents occurring in the 10th year following its initiation.
Project B will reduce by 500 the number of fatal accidents occurring in the 40th year following its initiation.

It follows from the conclusion that the value of future lives should not be discounted that Project B should be chosen. Yet the final two objections I wish to consider claim that, if several plausible assumptions are added to this example, other absurd consequences will follow.

Let us first assume that the resources do not have to be invested immediately and that the identical projects will remain available in the future. Suppose the resources can be invested at a 6 percent rate of return. If policy makers decided to save for 10 years and then invest in Program B, the number of fatalities in the 50th year would be reduced by 900. Let us call this Program C. If discounting is not used, decision makers should prefer Program C to Programs A or B because it saves more lives. Then, however, decision makers should actually prefer Program D (which says wait 20 years and then choose Program B) to Program C, and so on. By this line of reasoning, investment would be delayed indefinitely.

This objection makes one plausible assumption—that there are usually alternative possible uses for resources—but it also makes some implausible assumptions. The resources would have to be invested and not consumed; then, they would have to be reinvested, with interest, in the life-saving program in the 10th year. Will programs be available at the same cost in the future? Will policy makers have the resolve to use the resources that have been set aside for this purpose? The half-life of political commitments is considerably shorter than the half-life of uranium mill tailings. Furthermore, how will science and technology have changed in this period? It is unreasonable to assume that the situation will remain the same indefinitely. Policy makers have good pragmatic reasons to do what they can in the present and to choose programs they can implement immediately, but these are not reasons to save lives immediately. It is not uncommon at EPA for analysts to develop programs that can be implemented now but that will save lives mostly in the future.

A Paradox

There should be discounting for opportunity costs with money invested at some rate of return. Therefore, monetary costs should be discounted. In deciding, however, whether a life-saving program is cost-effective or cost-beneficial, it is necessary to assign a monetary value to the lives saved. (I will discuss the moral issues involved in this process later.) These assumptions appear to generate a paradox: the value of lives saved should not be discounted but money should be discounted; nevertheless, the lives saved can be given monetary equivalents.⁶

Let us refer again to Projects A and B and consider now an alternative, Project C', which is to choose neither program and return the resources to the private sector where they will be consumed or invested according to individual savings preferences. Suppose the present value of these resources is $100 million and the discount rate for investment and consumption is 6 percent.

Suppose further that it is determined that a life saved is worth $2 million. Program A would confer benefits worth $200 million in the 10th year; Program B would confer benefits worth $1 billion in the 40th year. The present value of Program A is $112 million, and the present value of Program B is $97 million. Without Program C', Program B is preferable to Program A. With Program C', Program A is preferable to Program B, and Program A passes a benefit-cost test while Program B does not. This result seems to violate the independence of irrelevant alternatives.

One of our assumptions must go: either that the value of lives should not be discounted or that the value of a life has a monetary equivalent. It is the latter assumption that should be given up.

There are no opportunity costs attached to saving lives later rather than sooner in the example. If these benefits are assigned monetary equivalents, they should not be discounted to a present value. One might then say that Program B confers undiscountable benefits worth $1 billion when they occur; thus, they are worth that much today, but even this formulation is very misleading. One should instead regard the alternatives as not investing $100 million worth of resources in life-saving programs, investing the same resources to save 100 lives, and investing them to save 500 lives. Program B would save lives as a cost of $200,000 per life—at today's rates, a bargain. Program B should be chosen.

Variations of these last three objections are common in the literature defending discounting for time preference in benefit-cost analysis. None of the authors in the various articles, however, gives a good reason for discounting the value of lives saved or lost in the future as a result of present policy choices and investments. Similar arguments would apply to the value of improved health and perhaps to other kinds of benefits as well. These objections show that benefit-cost analysis must be applied carefully; in addition, selectivity must be exercised in choosing what to discount, determining time horizons for the analysis, assigning monetary equivalents, and so on. I am arguing for selectivity and not for rejection of the method of analysis or discounting; that is, in cases in which discounting is appropriate.

What Should The Discount Rate Be?

I have argued that some costs and benefits should not be discounted. Other costs and benefits should be discounted, but at what rate? With a 10 percent discount rate, the present value of $1,000 of benefits 50 years hence is $8.52; at a 5 percent rate, the present value in 50 years is $87.20; at a 2 percent rate, it is $371.53. For a 50-year horizon, the difference between 10 percent and 5 percent is an order of magnitude; the difference between 10 percent and 2 percent is a factor of 40. The discount rate chosen can easily determine in many cases whether a project is cost-beneficial. Not surprisingly, therefore, an enormous technical literature argues for different rates.

I will not try to assess this literature here. I will instead suggest only why it might be wise to choose a social discount rate for health and safety regulations that is lower than the private sector rate; that is, lower than the discount rate individuals and firms apply to investment decisions based on rates of return and rates of interest prevailing in the market.

The principal reason given for similar social and private discount rates is the belief that social investments should meet the same economic standards that private investments must meet. This belief is sometimes reinforced by a currently popular political theory, which says that the proper role of government is to enforce rights and to act to correct market failures. Many people would reject this political theory and defend a more positive role for government. Indeed, the laws that give EPA its mandates make little reference to correcting market failures or to treating health and the environment as economic resources.

Even those who do not accept the market failure concept of government, however, ought to favor policies that make some economic sense, and this rationale is a sufficient reason to apply discount rates to some aspects of public investments. In some cases (e.g., development projects), the policy goals are economic; in some development projects, government investments will be competing with private sector investments. In these cases, there are better reasons for applying the same criteria private investors apply so as not to displace private investment.

This line of reasoning applies most clearly to water and energy projects. It does not apply to environmental, health, and safety regulations, areas in which there are both little evidence that the problems are caused by market inefficiencies, and independent moral arguments and political support for government action. If one also accepts that government has a special responsibility to protect the interests of future generations, a responsibility individuals and firms do not have and do not reflect in their economic decisions, then the reasons for applying different criteria to public and private investments are even stronger. The argument that regulations should be justified in economic terms is weaker; consequently, so is the argument that the social and private discount rates should be the same. Private investors do not worry that some resources are nonrenewable and that some costs and benefits are not replaceable. They do not worry that Americans have moral objections to treating certain benefits—such as human lives or perhaps wilderness areas—as resources to be exploited and invested. Citizens do worry about such things, however, and that is why laws are passed to establish regulatory agencies. It is natural to expect these agencies to operate with standards that are different from those applied in the marketplace.

There is another, related argument for lower social discount rates that is much discussed in the literature. This is the isolation argument (Marglin, 1963; Sen, 1967, 1982), which claims that the social discount rate should be lower than the private rate, even it the social rate is based strictly on individual preferences and all individuals have the same private rate of time preference.

The isolation argument assumes that individual savings decisions are based in part on an individual's altruistic preferences for future generations, perhaps especially for his or her own descendants. Yet individuals have limited control over the benefits these later generations will receive bemuse they cannot control the investment decisions of others. Thus, individuals might choose to save more, or to have a lower discount rate on consumption, if they could be assured that their greater savings would be matched by the greater savings of others. This case is a variation of a common public benefits problem, in which government policies for protecting the interests of future generations act as the coercive mechanism needed to secure the cooperation of others. Thus, a discount rate lower than that used by individuals in isolated decisions should be applied to national environmental policies.

As I have presented them here, these arguments for different private and social discount rates are obviously neither conclusive nor prescriptive (they do not say what the social discount rate should be). My point is only to suggest how moral arguments might be applied to determine not only which future costs and benefits should be discounted but also what discount rate should be applied.

Putting A Price On Life

Regulatory agencies do not set a price on human life; they merely try to uncover public preferences for risk reduction from data about consumer safety decisions, wage rate differentials for hazardous occupations, surveys, and contingent valuation studies. This approach is somewhat misleading because another possible way to determine the value of life would be to examine regulatory decisions themselves. Yet the norm is sought in other areas, of course, precisely to guide these decisions and evaluate existing regulatory policies. Such guidance and evaluation are the sole purposes in determining a social value of human life. This issue is a source of enduring controversy. Pricing life seems necessary for both holding regulatory agencies economically accountable and at the same time finding them morally repugnant.

The moral issue, of course, is treating a ''sacred" good as an economic commodity. Kant (1785:Ak. 434-435) wrote that human beings have intrinsic worth or dignity and that whatever has dignity is "above all price, and therefore admits of no equivalent." Solow, however, describes the issue more clearly: "It may well be socially destructive to admit the routine exchangeability of certain things. We would prefer to maintain that they are beyond price (although this sometimes means only that we would prefer not to know what this price really is)."

Decisions about acceptable risk inevitably involve comparing the value of saving lives and protecting health with the costs of doing so. Many would argue that, to make these decisions in a fully rational way, it is necessary to be explicit about these costs and the amount our nation is willing to spend to save lives. Such disclosure is to adopt a moral position about procedures for making important decisions, a position that Gibbard (1986:99) calls one of "technocratic moral reform." He suggests that

a rationally grounded morality will be reformist—perhaps shockingly so. It will not appeal primarily to our capacities to be aroused, as traditional moral reform movements have done, but to ways of regimenting the considerations involved to produce rational, coherent judgments. . .. We need, it seems, to train people in rational methods of risk assessment and so organize society that those methods really do determine policy with regard to risk.

Why do "many, perhaps most of us," as Gibbard says, find this a chilling prospect?

The answer is not, as some have claimed, that human life has an infinite price. That idea is not what Kant meant in saying that humanity is "above all price." The reason is rather, as Solow observes, that civilized people find it morally repugnant to view life as routinely exchangeable for other benefits. People thus feel uncomfortable with even rationally defensible procedures for making difficult decisions if those procedures make finding an exchange rate for life a prominent feature.

What sense can be made of this reaction? I would argue that the value of life is complex.⁷ Human life has intrinsic value, which makes it worth saving and prolonging. This component of life's value favors efficiency and the saving of more lives rather than fewer. It stands behind support for rational methods of risk assessment. Yet human life is also sacred, and this component of its value can work in a different and conflicting direction.

Durkheim (1915) regarded sacred values as "elementary forms" of religious life, by which he meant that even as societies become secular, there remains a kind of need that traditional religions fulfill in older or more primitive cultures. This need involves finding rituals that strengthen social integration. "There can be no society," Durkheim wrote (p. 417), "which does not feel the need of upholding and reaffirming at regular intervals the collective sentiments and the collective ideas which make its unity and its personality."

As a pluralistic culture, the United States does not have a single unity and personality, but there are clearly some basic moral values that all Americans share and that, as Durkheim would point out, are universal. Rituals carry symbolic meaning that call attention to these values. They are marked by special, perhaps nonrational behavior and actions that draw the attention of the community to objects or relationships that have a special place in the life of the group. Because rituals are symbolic, they rely on conventional forms of behavior, which can differ from group to group. In all societies, however, and especially in those areas in which man finds it necessary to "humanize" parts of his existence, characteristic activities marked by rituals can be found (Hampshire, 1983). These rituals surround birth, sex, and marriage, for example, and they also surround death and the taking of life.

Precisely because health and safety decisions have obvious economic consequences, it is necessary to guard against treating human life as exchangeable in these contexts. Some policies and procedures that are inefficient but highly symbolic can be effective guards. Startling examples of inefficient, ritualized behavior are common in our dealings with hazards and risks. For example, Americans are generally willing to engage in rescue missions when identified individuals are involved and to act as if—or certainly to give the appearance that—costs are not a consideration.

To argue that the sacred value of human life in these situations must be respected is not to deny in any way the value of efficiency in life-saving and the importance of saving more lives rather than fewer whenever possible. The point is rather a more subtle one. It is to suggest that there may be irresolvable tensions between our rationalistic, revisionist sentiments, on the one hand, and our conservative, ritualistic sentiments on the other. A rationalistic decision procedure may unavoidably threaten some of these sentiments, which may suggest that it is better not to make that procedure too absolute, too open, or too openly identified with public agencies like EPA that were created to pursue moral as well as other goals. It may perhaps be necessary to live with some controversies rather than to resolve them technocratically, and to tolerate "pockets" or modest levels of inefficiency for this purpose.

I mean also to suggest that the symbolic role of public figures like the EPA administrator should be recognized when he or she appears at a press conference to announce a regulatory decision, often about some hazard that to be reassured that things they value deeply—health, the environment, posterity—are being guarded and protected by the agency that has been established as the trustee of these values. Like it or not, the actions of EPA have important symbolic and expressive significance. Everyone may not approve of such taboos as refusing even to look at benefit-cost analyses, but it is necessary at least to be sensitive to the kinds of symbolic importance they might have.

Conclusion

The issues I have discussed involve very different kinds of arguments that lead to different paths of moral inquiry. Nevertheless, I think they all support the general conclusion that it is important to look critically and perhaps even suspiciously at suggestions that some analytic method should be universally applied to environmental decision making. The comparisons and trade-offs that must be made are often context dependent or may involve symbolic elements. These comparisons make the justification of decisions specific to a particular situation.

It is extremely important to use analysis to organize complex data and make decisions more consistent and efficient. It is also important, however, to realize that different values may have to be treated differently. It is this fact, and not measurement problems, that makes value comparisons in environmental policy making so difficult.

Acknowledgment

Many people commented on earlier versions of this paper. Claudia Mills, Talbot Page, Amartya Sen, and Susan Wolf provided some particularly useful criticisms. A. Myrick Freeman, III, and Roger Noll sent me detailed written criticisms. I have benefited from, though I am aware that I have not fully responded to, the questions they raised.

References

Durkheim, E. 1915. The Elementary Forms of Religious Life. Translated by J.W. Swain. London: George Allen and Unwin.
Gibbard, A. 1986. Risk and value. Pp. 94-112 in D. MacLean, editor. , ed., Values at Risk. Totowa, N.J.: Rowman and Allenheld.
Hampshire, S. 1983. Morality and Conflict. Cambridge, Mass.: Harvard University Press.
Kahneman, D., J. Knetsch, and R. Thaler 1986. Perceptions of unfairness: Constraints on wealth seeking. American Economic Review 76:728-741.
Kant, I. 1785. Grundlegung zur Metaphysik der Sitten. Page references to Prussian Academy edition of Kant's works (1911). Vol. III. Berlin. Translated by J.W. Ellington as Grounding for the Metaphysics of Morals (1981). Indianapolis: Hackett.
Keeney, R.L. 1984. Ethics, decision analysis, and public risk. Risk Analysis 4:117-129.
Kneese, A.V., S. Ben-David, and W.D. Schulze 1983. Ethical basis of benefit-cost analysis. Pp. 59-76 in D. MacLean, editor. , ed., Energy and the Future. Totowa, N.J.: Rowman and Littlefield.
Lind, R.C., editor. , ed. 1982. Discounting for Time and Risk in Energy Policy. Washington, D.C.: Resources for the Future.
MacLean, D. 1983. Valuing human life. Pp. 89-107 in D. Zinberg, editor. , ed., Uncertain Power. New York: Pergamon.
1986. Social values and the distribution of risk. Pp. 75-93 in D. MacLean, editor. , ed., Values at Risk. Totowa, N.J.: Rowman and Allenheld.
Marglin, S.A. 1963. The social rate of discount and the optimal rate of investment. Quarterly Journal of Economics 77:95-111.
Morrall, J.F. 1986. A review of the record. Regulation 10(Nov/Dec):25-34.
Nozick, R. 1974. Anarchy, State, and Utopia. New York: Basic Books.
Parfit, D. 1983. Energy policy and the further future: The social discount rate. Pp. 31-37 in D. MacLean, editor. , ed., Energy and the Future. Totowa, N.J.: Rowman and Littlefield.
1984. Reasons and Persons. Oxford: Clarendon Press.
Pigou, A.C. 1932. The Economics of Welfare. 4th ed. London: Macmillan. Ramsey, F.P.
1931. A mathematical note on saving. In F.P. Ramsey, editor. , Foundations of Mathematics and Other Logical Essays. London: Kegan Paul.
Rawls, J. 1971. A Theory of Justice. Cambridge, Mass.: Harvard University Press.
Sagoff, M. 1986 Values and preferences. Ethics 96:301-316.
Scanlon, T.M. 1982. Contractualism and utilitarianism. Pp. 103-128 in A. Sen, editor; and B. Williams, editor. , eds., Utilitarianism and Beyond. New York: Cambridge University Press.
Scheffler, S. 1982. The Rejection of Consequentialism. Oxford: Clarendon Press.
Schelling, T.C. 1984. The life you save may be your own. Pp. 113-146 in Choice and Consequence. Cambridge, Mass.: Harvard University Press.
Sen, A.K. 1967. Isolation, assurance and the social rate of discount. Quarterly Journal of Economics 81:112-124.
1982. Approaches for the choice of discount rates for social benefit-cost analysis. Pp. 325-353 in R.C. Lind, editor. , ed., Discounting for Time and Risk in Energy Policy. Washington, D.C.: Resources for the Future.
Solow, R. 1981. Defending cost-benefit analysis: Replies to Steven Kelman. Regulation 5(Jan/Feb): 39-41.
U.S. Office of Management and Budget 1972. Circular No. A-94. (Revised) To the heads of executive department and establishments, subject: discount rates to be used in evaluating time distributed costs and benefits. Washington, D.C.: OMB, March 27.

Footnotes

: ^1. This quotation comes from comments on an earlier draft of this paper by Roger Noll. Both the suggestion about how philosophical issues can be most usefully addressed and the list of philosophical perspectives or theories are common in the policy analysis literature (see, e.g., Kneese et al., 1983, and Keeney, 1984).
: ^2. For a detailed discussion of different approaches to moral inquiry, see Parfit (1984).
: ^3. I am indebted to John Broome for discussion of these points.
: ^4. For a good discussion of moral theories and moral doctrines, see Scanlon (1982).
: ^5. I owe this example to Robert Dorfman; see note 6 below.
: ^6. I first discussed this paradox at a workshop on discount rates at Resources for the Future in 1985. Robert Dorfman reformulated it more precisely and elegantly than I had. I rely here on his reformulation.
: ^7. I have argued this point more fully (MacLean, 1983,1986).
: Douglas E. MacLean is professor of philosophy at the University of Maryland.

Bookshelf ID: NBK235534

Contents