Approximately two million carloads of hazardous materials (e.g., liquefied petroleum gas, sulfuric acid, anhydrous ammonia, liquid caustic soda) are shipped annually in the U.S. According to a recent study of railroad technology done by the U.S. Office of Technology Assessment (OTA), roughly 65 percent of these tank cars are annually involved in the release of hazardous material. The OTA also confirmed that U.S. railroads have double the number of serious accidents, per train mile, as Canadian railroads. Despite these facts, the OTA study concluded that no new laws or regulations were needed to reduce U.S. rail accidents.

How could the U.S. railroad?accident rate not be seen as a problem? And why did the OTA recommend no new regulations? At least one rea son is that the OTA analysis followed classical risk?benefit, market methodology in assessing railroad risks. In so doing, the OTA study ignored the non-market costs or externalities associated with railroads. Externalities in this case include things like the social costs of evacuations and the pollution from derailments. By excluding non-market costs, the OTA study overestimated the benefits of current railroad policy and underestimated the risks associated with it.

A variant of cost-benefit analysis, risk-benefit analysis (RBA) consists of evaluating a proposed action in terms of the monetary values assigned to each of the various risks and benefits associated with the action. Each monetary assignment is made by calculating the "compensation variation" (CV) associated with a particular risk or benefit. The CV is the amount of money whose loss or gain, respectively, would perfectly compensate a person for the benefits or risks associated with a hazardous activity. (For example, a person living near a proposed airport might claim that a C V of $1000 per year was necessary to compensate him for the risks associated with airplane noise and crashes.) If the CV's of the gainers (e.g., those who benefit from the airport) are added to those of the losers, and if the resulting sum is positive, then the benefits outweigh the costs and risks, and the economic action (e.g., building a new airport) is usually judged acceptable. If the CV's of the gainers are added to those of the losers, and if the resulting sum is negative, then the costs and risks exceed the benefits, and the action is typically judged questionable.

Many of the important ethical problems associated with risk?benefit analysis concern how to assign a numerical or monetary value to each of the risks, costs, and benefits associated with a proposed action, e.g., transporting hazardous materials by railroad cars. Some of the more questionable assumptions built into the assignment of numerical values, by means of CV's, are: (1) that the gains and losses can be expressed numerically, on the basis of market prices; (2) that distributive effects can be ignored, provided that overall benefits exceed risksicosts; and (3) that benefits and risks can be defined in terms of individual preferences.

Expressing risks and benefits in terms of market values, assumption (1) is particularly problematic because it ignores Aristotle's distinction between price and value. Price is a measure merely of the intensity of human wants, but value is a measure of the intensity of ethically desirable human wants. Hence, it is arguable that market price does not measure value. Moreover, as Kenneth Boulding warned, the concept of value always includes an integrative system or "grants economy" involving things ike love, duty, honor and community. None of these has market value?yet, without them, there could be no market economy.

Apart from ethical limitations, there are several economic reasons why market prices diverge from authentic values. Market values ignore social costs or externalities (as the railroad example showed), as well as the distorting effects of monopolies and speculative instabilities. Market prices also assign a zero value to natural resources, public goods (e.g., public education), and free goods (e.g., clean air). Hence, by definition, RBA undervalues risks to natural resources and public goods.

The economists' rationale for failure to include externalities and natural resources in their analyses is that it is impossible to assign an objective numerical value to such parameters. They claim that because social costs (like the risk of evacuation) and natural resources (like air) are not priced or traded on a market, they are external to eco nomic evaluations.

Such a response, however, seems ethically questionable. It ignores the interests of persons in breathing clean air and drinking clean water, merely because common resources have no market value. It also ignores the rights of potential risk victims (e.g., those evacuated) to equal protection and to compensation for the risks they bear. Indeed, if Oscar Wilde is correct that a cynic is someone who knows the price of everything and the value of nothing, then RBA might dictate policy that is cynical as well as unethical.

RBA has also been attacked on ethical grounds because of its presupposition (2) that the distributive effects of an action can be ignored. Built on a utilitarian ethics that attempts merely to provide the greatest amount of good for the greatest number of persons, RBA ignores the fact that one subset of persons (e.g., those living along toxic waste transport routes) might receive a disproportionately greater share of overall risks or benefits. Questionable on grounds of equity, RBA implicitly condones a "politics of sacrifice" in which certain minorities bear much higher, uncompensated risks that benefit society as a whole.

Another fundamental ethical controversy is whether RBA errs in presupposing (3) that benefits and risks are defined in terms of individual preferences and egoistic hedonism. Many ethicists would argue that such a definition ignores the distinction between fulfilling wants and securing justice, or between utility and morality. Moreover, they argue, societal welfare is not merely the aggregate of individual, egoistic preferences. It obviously would not serve societal welfare, for example, if everyone behaved egoistically during a gasoline shortage and attempted to maximize his personal stock of fuel. A recent OTA assessment of automobile technology illustrated this point quite effectively. Using the norm of egoistic hedonism, the study concluded that, because "personal mobility" was important to each individual, future group welfare was best served by private automobiles, rather than by expanded use of mass transit. Following classical economic (RBA) methods, the OTA study defined social welfare as the aggregate of individual preferences, even though it admitted that, by the year 2000, its sanctioned automobile policy would cause more than half the U.S. population to be exposed to extremely hazardous levels of automobile?generated pollution. One solution to some of these ethical problems with risk?benefit analysis is to introduce ethical weights into economic methodology. For example, if a particular risk is inequitably distributed, then this inequity could be factored into the analysis and counted as a cost. Or, if a particular good (e.g., clean air), has no market value, then it could be "shadow priced" and air pollution (an externality) counted as a cost. Unless we devise some way for RBA to take account of the ethical principles that ought to constrain risk policy, our economics and our technology policy will be divorced from our ethics. To paraphrase Browning, our ethical reach will continue to exceed our risk?benefit grasp.