Breaking the Vicious Circle: Toward Effective Risk Regulation.

• Author: Williams, Stephen F.

A seasoned Breyer-watcher once said, "Steve thinks the solution to most problems is to turn them over to a room full of people as smart as he is."(1) The remark captures both the strength and weakness of Justice Breyer's 1992 Holmes Lectures, now published as Breaking the Vicious Circle. Breyer discusses key anomalies in our current system for regulating environmental risks, seeks out their causes, and proposes a remedy. The first phase, the discussion of the inconsistencies in our current approach to environmental regulation, is a tour de force, confidently integrating science and policy in terms easily accessible to the intelligent layman. The second, the explanation of why the incongruities prevail, is also very skilled, though impaired -- I will argue -- by its almost complete neglect of the role of interest group politics. The third, Breyer's suggested solution, raises more serious problems. He proposes establishing not just a room full of intelligent people, but a whole staff, located in the Office of Management and Budget, charged with nudging the relevant agencies toward policies that are consistent with each other and with economic rationality. A president who pursued the proposed remedy would doubtless seek people with the qualities so well reflected in this book -- versatility, evenhandedness, balance, and intelligence. Assembling such a staff would not be easy; people with these skills are in high demand and short supply. Even assuming their assembly, however, Breyer does not completely convince the reader that they could do the job. Can a SWAT team of technocrats, however brilliant, neutralize the interest groups at play in this field?

Breyer sees current policy as bringing about a wasteful misallocation of resources -- massive overinvestment in reducing some risks and comparative neglect of others. "Tunnel vision," he argues, leads agencies to pursue their missions without regard to competing values, a quest resulting in the attempt to eliminate "the last ten per cent" of a problem even where the benefits are minimal in relation to the costs (p. 11). In a Superfund case in Breyer's former court,(2) the Environmental Protection Agency tried to compel a private party to spend $9.3 million on additional cleanup of a waste site, so that children could eat its dirt without ill effect not merely for 70 days a year -- the level the party, amazingly, agreed to achieve -- but for 245 days a year (pp. 11-12). As there were no children in the vicinity, nor even residences, nor even any likelihood of residential development, and as at least half the offending chemicals would dissipate by the year 2000, the health return on this $9.3 million clean up cost would have been meager.(3)

Nor does the dirt-eating-children episode appear atypical. Breyer points to "mid-range" consensus estimates that the removal of asbestos from schools would cost approximately $250 million per statistical life saved (pp. 12-13). He cites the Fifth Circuit's observation(4) that the EPA's bans on a wide range of asbestos uses, had they been allowed to take effect, would have cost approximately $200 million by the EPA's own estimate while saving only about seven or eight lives over the next thirteen years -- less than half the expected deaths from ingested toothpicks (p. 14). Further, some of the Occupational Safety and Health Administration's benzene rules will evidently cost about $180 million per statistical life saved (p. 15). Breyer includes a table listing regulations by cost per life purportedly saved, including twenty-one regulations with costs over $10 million and one with costs reaching $5.7 trillion.(5) The pursuit of completely pure air, land, and water clearly takes one to a point at which the marginal cost of extra health may daunt all but the most zealous.

Breyer argues that the related problems of random agenda selection and inconsistency compound tunnel vision (pp. 19-28). Not only do experts' assessments of the most serious risks differ radically from the popular understanding, but the government makes no effort to establish overall priorities or to coordinate agency efforts (pp. 19-20). The agencies use different methods to calculate and evaluate health effects, and they ignore the ways in which one set of regulations affects the fulfillment of other environmental goals. For example, Breyer notes that when the EPA proposed rules to reduce sewage sludge it claimed they would save one life every five years, but it overlooked the rules' impact on waste incineration: the resulting increases in this form of disposal would have cost two statistical lives every year -- ten times the expected benefit of the sludge reduction (p. 22). Agencies similarly overlook the offsetting behavior of individuals. A farmer deprived of a relatively low-risk artificial pesticide, for example, may substitute plants with natural pesticides of equal or greater carcinogenicity.(6) At the very least, one might hope for reform that would force the government to coordinate its interventions. With some regulations costing billions but yielding almost no health gain, and with neglected opportunities at hand for saving lives at modest cost,(7) the potential gains from rationalization -- shifting social investments to the most promising health and safety opportunities -- seem huge.(8)

Breyer attributes the persistence of wasteful regulation to three causes: inaccurate public perceptions, congressional action and reaction, and uncertainties in the regulatory process. Though believing that people think rationally in assessing risks and possible remedies, Breyer argues that most people are unlikely to acquire a full grasp of the relevant facts. If people "think dramatically, not quantitatively," for instance, as Justice Holmes observed,(9) they are likely to reach inaccurate conclusions from a few stories associating toxic waste dumps with cancer. With 26,000 waste dumps throughout the country and a large number of locales with above-average cancer rates (by definition roughly one half), the probability that these two facts will coincide is overwhelming. A few dramatic coincidences, though in fact meaningless, may create an impression of a causal link.

Congress, meanwhile, adopts a range of mandates, from extremely detailed linguistic formulae, to absolute bans, to "hammers" -- statutes that allow the EPA some time to devise reulatory control for some activity (for example, underground disposal of a chemical), but ban the activity outright if the agency misses the deadline. However these approaches might be reconciled in theory, in operation they yield widely varying degrees of stringency. The variations are hardly surprising. Congressional bills originate at different times, in different committees, with each subcommittee "competing for political time and attention" and thus likely to "consider the particular problems that it has studied as the most important, deserving a place at the head of the regulatory queue, whether or not dispassionate observers would reach the same conclusion" (p. 42). Although statutory language may offer apparent flexibility, agency fear of congressional oversight may limit its use: EPA officials may know that, "given an individual legislator's political incentive to appear in interesting, positive news stories, hearings are far more likely to mean criticism for leniency than for strictness" (pp. 40-41).

Finally, the facts about environmental risks are most uncertain. Agencies typically infer risk to humans by extrapolating from the response of rats and mice exposed to very high doses. In doing so, they assume a linear dose-response curve; that is, they assume that if a dose of X causes cancer in fifty out of a hundred rodents, a dose of X/1000 -- adjusting for the difference in weights or body surface area, to be sure -- will cause cancer one thousandth as frequently, or in fifty humans out of every one hundred thousand exposed. In other words, the method infers human responses at low doses from rodent responses at very high doses. But rats and mice differ from humans in ways that appear relevant to the...