Risk–Benefit Analysis and Richard Wilson , Edmund A. C. Crouch Harvard U. Press, Cambridge, Mass., 2001 [1982]. $25.00 paper (370 pp.). ISBN 0-674-00529-5
A recent US Geological Survey report on “pollutants” in US waterways catalogs widespread presence of all manner of chemicals, none of which are attractive to those of us who drink water. The mere detection of contaminants gives cause for concern, but to know whether or not to take action, we must understand the risks (costs) of consumption and the costs of preventing or removing the contaminants. What does 3.6 ppb of antibiotics in water mean? What about 2.0 ppb of reproductive hormones? Our present exquisitely sensitive ability to detect things can render decision-making more complicated.
“Risk” is arguably the most complex single-word topic I can think of. Life in all its forms constantly faces risk (assumed and imposed) and responds in manifold ways. In simpler forms of life, experiences with risk result over time in genetic (Darwinian) change. In more complex forms, learning from risk experiences leads to behavior modification. In still higher forms of life, cumulative experience (folk wisdom, for example) leads to the observation that “a wise person learns from experience but a wiser person learns from the experience of others.” Risk processes for humans also include all the simpler responses but extend to quantitative analysis and integration of relevant information so that vastly more informed risk decisions can be made in advance of undertaking the risk. As a consequence we find human procedures, equipment, and services (medicine, for example) designed to minimize adverse risk to the user while providing maximum benefits.
Richard Wilson and Edmund Crouch, the highly qualified authors of Risk–Benefit Analysis , provide copious examples of both the evolution of the ways analysts approach (model) risk estimation and the translation of analysis into meaningful terms. Wilson and Crouch make clear the pitfalls of incomplete analysis and insufficiency of good data. Their examples of actual cases are instructive in illustrating the seeming inability of some people, however well-meaning, to interpret the significance of numbers. One such example is the risk presented by the Cassini spacecraft’s flyby of Earth on its way to Saturn a few years ago, with plutonium-238 on board for thermoelectric energy generation. Another is that of the Tennessee Valley Authority nuclear reactor disaster in 1975 at Brown’s Ferry, Alabama, due to a “common mode” failure (bad construction design). Bertrand Russell once observed that “man would rather commit suicide than learn arithmetic.”
Risk analysis can be useful to the informed individual in making personal decisions, but its greatest value lies in forming the basis for corporate and public policy decisions about acceptable risk. We must have “rules of the road” that are technically as well informed as possible but at the same time are acceptable to society. These constraints can lead to great consternation, amply illustrated in the book.
The book would be more useful if it contained more discussion of ways to accommodate better to the natural variability of reaction to exposure. It is not fair to assume an “average” or “randomly selected” person when setting permissible doses. Neither does it seem fair to set limits based on the most sensitive population independent of the costs. The authors should have gone more deeply into such questions.
More discussion would also have been helpful in the area of accepting assumed versus imposed risk, because our lives are increasingly confounded by both types. (By “assumed” I mean risks that could be avoided by voluntary action, smoking or charbroiling meat for example; “imposed” refers to natural and human-caused actions such as earthquakes and the production of environmental carcinogens from industrial sources.)
Chapter 4, “Perception of Risk,” puts human nature squarely into the discussion: The perception of risk is deeply influenced by the past experience of the observer, which includes events that have affected trust. It appears that trust-decreasing events have considerably more impact than trust-increasing events. Risk managers should give special attention to this chapter.
All in all, Wilson and Crouch have given us an enlightening and entertaining tome on risk, risk perception, and public policy. I do hope that future editions will be graced by a more careful edit (“millions” on page 73 is not an acceptable substitute for “microns,” for example). I believe that material on methods of risk calculation and estimation, in Chapter 2, could be less distracting if it were organized by types of models that are used with different types of data rather than according to old risks and new risks. The authors would also serve their readers better with less one-sided critiques of regulatory discussions.
Finally, while Wilson and Crouch focus on the risk side of things, it is disappointing (although not surprising) that they devote little energy to benefits. This is particularly the case for such noneconomic benefits as reduced risk to future generations, aesthetic considerations, and such nonhuman factors as ecosystem health that still lie outside the universe of traditional national income accounts. Such nonmarket values are emerging as subjects of public attention as we wrestle with governance and stewardship in the 21st century.
Read the book, enjoy the cartoons and quotes, and ponder the lessons learned.
John H. Gibbons, after 20 years practicing physics, migrated to energy efficiency, environment, technology, and public policy; he directed the US Congressional Office of Technology Assessment for 14 years before becoming science adviser to President Clinton.
