A branch of economics called "environmental economics" has been trying to establish monetary values for social externalities, natural resources, and public goods such as water quality. It uses the following methods: the "averting behavior" method, the "weak complementarity" method, the "hedonic market" method, and the "direct questioning/contingent valuation" method.

I won't discuss these in detail here, but I will note ther premises. The "averting behavior" method assumes that the value of environmental quality can be measured by examining expenditures on market goods that compensate for reductions in the quality of environment (e.g. water filters purchased to clean up the effects of water pollution). The "weak complementarity" method, used primarily to assess the value of outdoor recreation, assumes that variation in the intensity of use of outdoor sites for recreation can be explained by their relative environmental quality, and that absolute and relative monetary values can be assigned to environmental quality by comparing the costs of visiting different sites. The "hedonic market" method begins from the assumption that environmental quality is part of a bundle of linked characteristics that account for difference in price assigned, say, to residential property. Finally, the "direct questioning/contingent valuation" method surveys respondents and asks them to state how much they are willing to pay for an environmental amenity.

As should be apparent from this quick summary of methodological premises, all these economic approaches to environmental qualityan immeasurable variable-attempt to find its economic value by measuring the measurable and assigning to that measurement an added meaning: the cost or benefit of environmental quality.

Professor Shrader-Frechette has identified three methodological and ethical problems in economic cost benefit assessment:(1) Can social gains and losses be expressed numerically in market prices? (2) Can distributive effects be ignored, provided that overall benefits exceed overall costs? and (3) Can benefits and risks be defined correctly in terms of egoistic hedonism? I believe with her that these are genuine problems affecting the rigor and utility of these methods. I disagree with her only when she asserts that market prices fail to account for social costs, externalities, natural resources, or public goods. Market prices account for these variables, but in a non-obvious way. Let me explain.

About 45 years ago, Friedrich Hayek argued that (1) non-centrally planned markets are the most efficient means of conducting markets and (2) a price established for a good or service is the most efficient way to inform transactions in that market because the price represents the aggregate quantity and quality of information known relevant to a good ormental quality is part of a bundle of linked characteristics that account for difference in price assigned, say, to residential property. Finally, the "direct questioning/contingent valuation" method surveys respondents and asks them to state how much they are willing to pay for an environmental amenity. As should be apparent from this quick summary of methodological premises, all these economic approaches to environmental qualityan immeasurable variable-attempt to find its economic value by measuring the measurable and assigning to that measurement an added meaning: the cost or benefit of environmental quality.

Professor Shrader-Frechette has identified three methodological and ethical problems in economic cost benefit assessment: (1) Can social gains and losses be expressed numerically in market prices? (2) Can distributive effects be ignored, provided that overall benefits exceed overall costs? and (3) Can benefits and risks be defined correctly in terms of egoistic hedonism? I believe with her that these are genuine problems affecting the rigor and utility of these methods. I disagree with her only when she asserts that market prices fail to account for social costs, externalities, natural resources, or public goods. Market prices account for these variables, but in a non-obvious way. Let me explain.

About 45 years ago, Friedrich Hayek argued that (1) non-centrally planned markets are the most efficient means of conducting markets and (2) a price established for a good or service is the most efficient way to inform transactions in that market because the price represents the aggregate quantity and quality of information known relevant to a good or service, in a short-hand available to all buyers and sellers in that market. Hayek was criticizing the efficiency claimed for centrally planned economies. In his view, no central planner can have all the information about all the attributes of the market at any time. Information is dispersed among the buyers and sellers. Pricing, in his view, summarizes efficiently all the demand and supply information available about the market in the only signal available to all participants.

I agree with Hayek: sellers price their products and services in light of all they know or believe about the supply and demand characteristics of their market; buyers pay, or refuse to pay, sellers' prices on the basis of all they know or believe about their market. By this I do not mean that decentralized markets are "equitable." I mean only that prices convey all the available information about the condition of the market in a concise signal. "Information;' as W. Schramm said, "is anything that reduces uncertainty.''

How do I merge these observations about price as an informational signal and information as anything that reduces uncertainty? Consider this example:

A consumer with a limited supply of money is considering whether to purchase a new car. She might have several options: several dealers and several prices of a specific car in which she is interested.

Many things will influence her decision. Certainly she will be influenced by what she knows about the quality and features of the cars and the dealers. She will also be influenced by her knowledge of how else she could use her money and of her own needs and desires. Her information about these things will reduce her uncertainty and she will positively value that information.

Yet, it doesn't matter whether she "rationally" costs out her options and chooses one or instead chooses impulsively. Either way her choice is influenced (consciously or unconsciously) by what she does not and cannot know: which choice will satisfy her. She cannot know which choice will satisfy her when she is deciding because that information about her ultimate satisfaction which would minimize or eliminate her uncertainty-is in the future, when she will have made (and must live with) her decision.

Thus, in assessing her choices she necessarily will evaluate the probable benefits and costs of her choice in terms of her information and her uncertainty. And she inevitably will attach her net assessment of those benefits and costs to the price she decides to pay (or to forego paying) for a car. The price she decides to pay represents in shorthand how she monetarily values her information and her residual uncertainty about that specific decision.

The consumer is engaged in risk benefit assessment. She is juggling what she knows and doesn't know about herself in order to decide what to do. And she will translate her assessment into market conduct.

These observations are relevant here. We have developed our risk assessment tools to aid or justify our public and private decision-making. We use risk assessment to value what we know and don't know about situations in which we can or must make choices. Risk assessment, when it is rigorous, enables us to identify and value our information and our uncertainty. Inherent, therefore, in any kind of risk assessment, including risk-benefit assessment, is a large amount of irreducible uncertainty-owing to our incomplete knowledge of the past, present, and future.

This is the most significant problem inhering in risk assessment. Prices represent buyers' and sellers' conclusions about what they know about goods and services and what they don't know. Risk assessment, no matter how methodologically rigorous or ethically scrupulous, no matter whether it accounts for the differences between "price" and "value;" must assign a current value to uncertainty, in order to arrive at a comprehensive price for a risk or its associated benefit. Risk assessment assigns to uncertainty the value today of some product or service, some risk or benefit, whose value we cannot yet know.

When risk assessors attempt to apply their techniques to the value of environmental protection or resource conservation, they face uncertainty of huge dimensions. Environmental risk-benefit assessment manipulates poorly understood phenomena comprised of innumerable variables with relatively new and unsophisticated tools. Uncertainty dominates environmental cost-benefit analysis. Analysis of environmental quality's "price" and "value" thus may understate, overstate, or accurately state its "true" value.

For these reasons, environmental risk assessment is itself a risky business. We should recognize its limitations. We should examine it whenever we use it.

Yet, we need risk-benefit assessment in public decision-making about environmental quality. Risk benefit assessment permits us to identify, through careful criticism and refinement of assessment methods, what we know and don't know about environmental risks. It thus obliges us to confront uncertainty, and it obliges us to put a price on uncertainty-about the future and about the methods we use to anticipate the future. Risk assessment enables us to recognize that risk assessment itself is an activity to which costs and benefits attach the cost of the labor of risk assessment and the benefits that might accrue from its contribution to policy-making.

Because it deals in known and unknown attributes of the past and present and the unknowable attributes of the future, risk assessment can be deductive in only one sense: taken together, the context and process of any assessment will force a conclusion. Otherwise, risk assessment is far from deductive: the definitions and premises an assessor chooses for an assessment, the tools she employs to collect data, and the analysis that ensues are all shaped by the assessor's judgments. If that aspect of risk assessment is ignored, risk assessment will seduce us ...and betray us in the bargain.