Assessing the Risk Assessors
Can We Rely on the Findings of Government Experts?
MARCH 01, 1995 by DANIEL BENJAMIN
Dr. Benjamin is Professor of Economics at Montana State University and Fellowship Project Director at PERC in Bozeman, Montana.
Life is risky business. As we travel its uncertain voyage, hazards must be assessed and choices made among them. For most of recorded history, both assessment and choice have been the prerogative of the individual. Society and its institutions—most notably the law—have exhibited great faith in the ability of individuals to regulate the risks of their personal environments.
Over the past 30 years, however, the right (and obligation) of the individual to assess and choose among risks has been displaced by decisions of courts and government regulatory agencies. There are, I believe, two reasons for this development. One of these is the erroneous notion that individuals make systematic errors in their assessment of risks, an idea I have challenged elsewhere. Here I address the second factor, the widespread view that experts at government agencies and in the judicial system can perform the risk assessment process more accurately because they are, in the words of Webster’s, “very skillful or highly trained.”
There is mounting evidence that this view of expertise is misguided. Within regulatory agencies such as the Occupational Safety and Health Administration (OSHA) and the Environmental Protection Agency (EPA), experts have adopted policies of extreme conservatism. At every point in the process, the worst-case assumptions are made, so that the compound effect may be to overstate the true risks by a factor of ten or even 1,000—without saying so. Simultaneously, the judicial system has increasingly dismissed the capacity of market participants to determine the appropriate level of risk. Consumers are viewed as unable to assess risk and unwilling to bear it, while producers are judged unwilling to sell the level of safety and health that consumers demand.
The result of this regulatory and judicial intrusion into the risk assessment process is twofold. First, there have been a rapidly growing number of costly attempts to reduce health and safety hazards that are simply not very dangerous. Just as importantly, the relative hazards of different threats have been severely distorted; as a result, we attack the less serious risks, leaving the more deadly to wreak their havoc.
Expertise and Evidence
Two recent policy issues illustrate why experts must be tools rather than arbiters in the process of assessing risk. First, consider cadmium, a toxic industrial metal used to coat metals and to make batteries and pigments. There is general agreement that at high doses cadmium can damage kidneys and possibly cause lung cancer. During the early 1980s, OSHA investigated the possibility of imposing stringent controls on workplace exposure to cadmium. OSHA’s review of the evidence led it to propose rules that would have reduced exposure levels 100-fold, saving the lives of 14 workers each year.
The Office of Management and Budget (OMB) subsequently reviewed the same body of evidence. It concluded that the proposed regulations, far from saving 14 lives each year, would actually kill at least 25 additional individuals! OMB reasoned that the high cost of the proposed OSHA regulations would reduce America’s wealth. As people become wealthier, they demand more safety; as they become poorer, they choose less safety. Thus, the decline in wealth caused by the costly new regulations would cause a decline in the demand for safety and lead to a rise in total fatalities.
So one group of experts claim that the proposed regulations are an improvement; the other group of experts say they are not. The relevant issue is that the experts disagree. Which expert should we choose?
The second question is which evidence should be relied on. Consider the “greenhouse effect,” the apparent tendency of carbon dioxide (CO2) and other gases to accumulate in the atmosphere, acting like a blanket that traps radiated heat and increases the earth’s temperature. Humankind is producing greenhouse gases at a record rate, and they are accumulating in the atmosphere. Over the past 50 years, the amount of CO2 in the atmosphere has risen about 25 percent.
In principle, higher CO2 levels should lead to higher global temperatures; the National Academy of Sciences has suggested that by the middle of the twenty-first century, greenhouse gases could be double their levels of 1860 and global temperatures could rise by as much as 2° to 9° Fahrenheit. Many of today’s temperate climes could become arid dust bowls.
At first glance, such a view seems consistent with the fact that, on average over the past century, greenhouse gases have been rising and so has the average global temperature. Yet almost all of the temperature rise occurred before 1940, while most of the increase in greenhouse gases has occurred after 1940. In fact, global average temperatures fell about 0.5° between 1940 and 1970; this cooling actually led some prominent scientists during the 1970s to forecast a coming ice age. Thus, less than 20 years after experts assured us the evidence pointed directly at falling temperatures, experts are now saying that the evidence points toward rising temperatures. The key is that here, as in many cases, the evidence is mixed. We must still decide which evidence to rely upon.
The Role of Incentives
Once we admit the possibility that experts err (or at least disagree) and that evidence is sometimes ambiguous (or at least arguable), we are forced to recognize that experts’ judgments often reflect the incentives they face. Indeed, the cadmium episode reveals quite clearly how those differences can depend on the incentives at hand.
As reflected in its name, OSHA is charged with protecting worker safety, and one must presume that the rewards facing its employees are structured accordingly. In contrast, OMB emphasizes managerial efficiency and ferreting out the cost implications of government actions. The positive safety implications of the proposed cadmium regulations stem from the physical protections offered workers who handle cadmium; it was the evidence on these effects that OSHA’s experts found most compelling. The adverse implications of the regulations, so compelling to the OMB experts, arise from their negative impact on our standard of living and the resulting decline in the demand for safety.
It is unlikely to be sheer coincidence that the respective experts’ conclusions dovetailed closely with the mandates of their respective agencies. It is far more probable that these experts—like all experts—responded to the incentives they faced. They chose to evaluate the evidence not in a vacuum, but in accordance with the costs and benefits they perceive likely to accrue to them as a result of their decisions.
It is arguable that only the people who bear the consequences of decisions can fully know the advantages and disadvantages of each expert decision, including decisions about what evidence to accept. In the case of global warming (or cooling), one body of evidence suggests that we act now to avoid further atmospheric warming; other evidence suggests that we should adopt a wait-and-see attitude. The physics and chemistry of the issue are insufficient to determine which body of evidence should be acted upon, just as the principles of aerodynamics and internal combustion are insufficient to determine whether you should fly to Los Angeles next Tuesday. We must know the benefits and costs for human beings. The ultimate “experts” on these are the individuals who will gain or lose as a result of the decisions that are made.
It is tempting to argue that we can instruct the experts to behave in ways that are consistent with the wishes of the people affected by their decisions. But this argument assumes away key aspects of the risk assessment problem. First it supposes that the experts are already in place, without answering the question, Which expert should we rely on? Second, and perhaps more importantly, it supposes that these experts know all relevant facts about the individuals affected by their decisions. As F. A. Hayek demonstrated a half-century ago, the body of knowledge that enables the economic system to function coherently is not in the hands of any individual or subset of individuals within the economy—no matter how great their expertise. This knowledge begins as disparate bits and pieces of seemingly unrelated facts that are communicated and augmented by the operation of the market. To suppose that experts can bypass the process is to foolishly suppose that they are somehow endowed with the information that is produced by the price system.
I am driven to the conclusion that both the experts and the rules under which they operate must be chosen by the people who will bear the effects of the experts’ assessments. The ideal way to ensure this is to maximize the extent of risk assessment that takes place in the marketplace rather than in regulatory agencies or the courts. But if for other reasons it is necessary that government experts play a role, it is important that none be anointed “czar” of his or her respective risk arena, and that all be subject to the broadest possible review.
The participation of OMB in actively crafting OSHA and EPA regulations is likely to reduce the damages caused by regulations, because OMB participation adds to the spectrum of individuals and interests represented in the regulatory process. Similarly, requiring regulatory agencies to take the costs and benefits of their decisions explicitly into account forces them to broaden the range of people whose interests are considered. The point is not that the economists computing these numbers have more expertise, but that more interests are brought into the process.
The flip side of this is that those who are choosing the experts (or choosing to do without them) must bear the full range of costs and benefits associated with their choices. For example, if the firms that use cadmium play a role in crafting the rules limiting workplace exposures, these same firms should not then be able to hide behind the limitation of liability implied by workers’ compensation laws. In other words, the firm should not enjoy all the benefits of a production process but then be able to avoid some of its costs in the event that something goes wrong.
The consequences of risk assessment decisions will be borne by someone, whether that someone is the decision-maker or not. Those consequences will be the most advantageous possible only if those individuals that have the greatest incentive to decide among the unknowns—and the unknowable—are making those decisions. The answer to Juvenal’s query is this: Only the owners of what is guarded are competent to select and monitor those who would guard it. And if the owners fail in these duties, their status soon will be that of former owners. 
1. See Daniel K. Benjamin, “Risky Business Rational Ignorance in Assessing Environmental Hazards,” in Roger E. Meiners and Bruce Yandle (eds.), Taking the Environment Seriously (Lanham, Md.: Rowman & Littlefield, 1993), pp. 209-31.
2. See Risk Assessment in the Federal Government: Managing the Process (Washington, D.C.: National Academy of Sciences, 1983); Albert L. Nichols and Richard J. Zeckhauser, “The Perils of Prudence: How Conservative Risk Assessments Distort Regulation,” Regulation, November/December 1986, pp. 13-24,; and Peter W. Huber Galileo’s Revenge: Junk Science in the Courtroom (New York: Basic Books, 1991).
5. Laboratory analysis of glacial ice dating back 160,000 years indicates that global temperatures and CO2 levels in the atmosphere do tend to move together, which would seem to suggest that the impact of today’s rising CO2 levels may be higher global temperatures in the future. But the historical CO2 changes do not precede the temperature changes (which is the essence of the theory); instead they are either contemporaneous with the temperature changes, or they actually follow the climate changes. See Policy Implications of Greenhouse Warming, ibid., and Sherwood B. Idso, Carbon Dioxide and Global Change: Earth in Transition (Tempe, Ariz.: IBR Press, 1989), pp. 51-53.
6. Robert J. Charlson et al., “Climate Forcing by Anthropogenic Aerosols,” Science, 252 (January, 1992), pp. 423-30, argue that the warming effects of greenhouse gases are currently being offset by the cooling effect of additional particulate matter produced by the same forces (chiefly hydrocarbon combustion) that generate the greenhouse gases. The net impact on global temperatures—for the moment—thus appears to be roughly nil.
7. Today’s decisions may affect the (potential) well-being of persons not yet born. This does not mean that these future generations will be ignored in present decisions, nor that experts must be appointed to act on behalf of our descendants. Although future generations cannot directly participate in today’s decision-making, they are (except for unforeseeable mutations) fully represented genetically by individuals comprising the current generation, and thus best represented by members of the present generation. See Richard Dawkins, The Selfish Gene (New York: Oxford University Press, 1976).