Implementing the Precautionary Principle

Mark Geistfeld

The author is a Professor of Law, New York University School of Law. This research was supported by a grant from the Filomen D'Agostino and Max E. Greenberg Research Fund at the New York University School of Law. Copyright 2001 Mark A. Geistfeld.

[31 ELR 11326]

I. Introduction

The precautionary principle has become an increasingly important component of environmental policy, considered by the European Union (EU) to be a "full-fledged and general principle of international law."¹ As the precautionary principle has gained prominence, policy analysts have devoted increasing attention to the issue of implementation.² Nevertheless, the practical implications of the principle remain unclear. The difficulty of formulating environmental policy pursuant to the principle is reflected in the European Commission's recent communication concerning the principle.³ Even European commentators acknowledge it is "doubtful that the communication will go far in clarifying the EU's use of the principle or convincing the [United States] of its validity."⁴

The vagueness of the principle explains why it so hard to implement. Consider the version of principle invoked in § 15 of the Rio Declaration on Environment and Development:

Where there are threats of serious or irreversible damage, lack of scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation.⁵

This formulation of the principle merely states that uncertainty does not justify inaction; it provides no guidance on how regulations should account for uncertainty, nor does it specify the objective of environmental regulation — the "cost-effective measures" that ought not be postponed because of scientific uncertainty.

The vagueness of the precautionary principle suggests that its development and implementation will proceed in stages. Initially, implementation is likely to involve the least controversial aspects of the principle. If these minimal requirements of the principle can be turned into a regulatory decision rule, the rule would at least partially satisfy most proponents of the precautionary principle. Disagreements about other aspects of the principle can then be framed in terms of alterations to the regulatory decision rule. By implication, if the core aspects of the principle cannot yield a well-defined decision rule — if the essential, widely agreed upon aspects of the principle do not translate into a coherent set of performance obligations — the hard question arises whether the precautionary principle is anything more than sentiment or political slogan.⁶

For various reasons, the least controversial version of the principle must include evaluation of costs and benefits. The precautionary principle addresses the relation between scientific uncertainty and risk regulation. In situations of scientific uncertainty, the environmental hazards are not known with the requisite degree of confidence, because the scientific techniques of risk assessment yield a range of possibilities. Regulators therefore must choose among the various plausible risk assessments, and that choice is addressed by the precautionary principle. The choice of risk assessment, in turn, defines the risk that regulators use to measure the safety benefit (reduced risk of injury) of the regulation. That benefit is then compared to the costs of regulation. As an analytic matter, then, the choice of risk assessment, which is governed by the precautionary principle, can differ from the evaluation of costs and benefits.⁷

For this reason, the EU defensibly recognizes that implementation of the precautionary principle includes examination of costs and benefits.⁸ Moreover, regulators in the United States often are required by statute or executive order to scrutinize proposed regulations with [31 ELR 11327] cost-benefit analysis.⁹ In these contexts, implementation of the precautionary principle will involve an evaluation of costs and benefits.

Pragmatic reasons also counsel in favor of relying on cost-benefit methodology to implement the precautionary principle. Some of the sharpest critics of the precautionary principle focus on the seemingly irrational outcomes that would be produced by a regulatory regime that pays insufficient attention to costs and benefits.¹⁰ These critiques could be answered, and the controversy surrounding the precautionary principle reduced, if implementation of the principle included evaluation of costs and benefits.

An implementation strategy for the precautionary principle also should consider the rules of international trade, because regulatory measures based on the precautionary principle are a likely source of trade disputes. For example, the EU invoked the precautionary principle in banning U.S. hormone-fed beef.¹¹ The United States alleged the ban violated the World Trade Organization (WTO) Agreement on Sanitary and Phytosanitary Measures (SPS Agreement), which allows Member countries to enact stricter, nonharmonized regulatory standards for the protection of human, animal, or plant life only if based on risk assessment, scientific principles, and scientific evidence.¹² According to the WTO Appellate Body that decided the case, this requirement is not satisfied by mere invocation of the precautionary principle; the SPS Agreement requires a scientific basis for concluding that the regulated activity poses a threat to human health.¹³

Incorporating the scientific techniques of risk assessment into an implementation strategy for the precautionary principle is consistent with the evaluation of costs and benefits. Science is an essential component of the risk assessment, which is required to calculate the injury costs posed by the potential hazard. These components of an implementation strategy for the precautionary principle therefore coherently fit together. The question is whether they cohere into a regulatory rule that adequately satisfies the normative requirements of the principle. If not, there may be no sufficiently noncontroversial core to the principle capable of guiding regulators.

Part II of this Article begins by arguing that a plausible, noncontroversial norm embodied in the precautionary principle requires that scientific uncertainty should not disadvantage those individuals who might be physically harmed by the potential hazard. Part III shows how this norm can justify environmental regulations based on the risk assessment reasonably preferred by potential victims. That assessment can be combined with standard economic methodology to quantify the injury cost of the potential hazard. That cost can then be evaluated pursuant to the regulatory procedure for known hazards. The resultant decision rule protects potential victims from the adverse consequences of scientific uncertainty as required by the norm apparently embodied in the precautionary principle. The decision rule requires a reasonable basis for the risk assessment adopted by potential victims. Part IV uses the SPS Agreement to illustrate how the rules of international trade can provide adequate guidance to regulators on how to apply the requirement of reasonableness.

Of course, this decision rule is unlikely to satisfy all proponents of the precautionary principle, which is to be expected in light of the vagueness of the principle and the associated room for disagreement. The decision rule does show, however, that the principle has a plausible core meaning that yields a well-defined regulatory approach. Translating the principle into a decision rule also has the potential for advancing the debate from vague generalities to the particularized guidance required for regulation.

II. Normative Content of the Precautionary Principle

The precautionary principle may have no content independent of the norms already adopted by the regulatory regime. Consider once again the principle as invoked in § 15 of the Rio Declaration on Environment and Development:

Where there are threats of serious or irreversible damage, lack of scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation.¹⁴

Under this formulation of the principle, anything of practical import could turn solely on the objective of environmental regulation rather than on the particular impact of uncertainty. The reference to "cost-effective measures" requires regulations that attain the given environmental objective at lowest cost. The reference does not specify the objective, however, and the objective itself may already provide a method for dealing with scientific uncertainty.

If the objective is to efficiently regulate environmental harms, for example, then the pursuit of efficiency often will [31 ELR 11328] require regulations in the face of uncertainty. The uncertainty necessarily creates error costs for any regulatory decision — be it one of action or inaction — and an allocatively efficient decision rule will seek to minimize those costs rather than ignoring them. Hence the regulatory objective of allocative efficiency, without more, implies that "the lack of scientific uncertainty shall not be used as a reason for postponing cost-effective [efficient] measures to prevent environmental degradation." The precautionary principle therefore may impose no new requirements on a regulatory regime committed to allocative efficiency.

For these same reasons, the precautionary principle need not impose any additional requirements on a regulatory regime that emphasizes safety and environmental protection over cost considerations. Situations of uncertainty involve the possibility that costly precautions will be required for activities that turn out to be benign. But the mere possibility that an activity is benign may be fundamentally similar to the mere possibility that a known hazardous activity will harm no one. The risk of cancer, after all, does not necessarily translate into actual cases of cancer. Probabilistic outcomes involve an element of uncertainty, so a regime committed to the reduction of cancer risk need not be paralyzed by the possibility that a regulated substance may not in fact be carcinogenic. For such a regime, the precautionary principle may mean nothing.

Because the objectives of an environmental regulatory regime are likely to involve some method of incorporating scientific uncertainty into the decisionmaking process, there is a hard question of whether the precautionary principle imposes any additional requirements on regulators. Assuming the precautionary principle has some independent normative content — that it is a "full-fledged and general principle of international law" as claimed by the EU¹⁵ — the emphasis clearly is on precaution or reducing the risks of environmental harm. The precautionary principle, in other words, involves some sort of linkage between precautionary behavior and scientific uncertainty.¹⁶

Precaution or risk reduction could serve at least two purposes. Most broadly, the purpose could be preservation of the environment, with considerations of human welfare relevant only insofar as humans are part of the overall environment. Environmental preservation is worthy, in short, for reasons other than its impact on people.

Protection of the environment and of nonhuman interests may turn out to be defensible, but surely such an objective is controversial at present. A more limited, less controversial objective focuses on how environmental degradation adversely affects people, now and in the future. Precaution is worthy only insofar as it protects people from harm. On this view, the precautionary principle emphasizes precaution due to a concern for protecting those individuals threatened by environmental degradation. By eliminating uncertainty as a reason for foregoing otherwise desirable precautions, the precautionary principle implies that the uncertainty should not disadvantage potential victims.

In situations of scientific uncertainty, then, the precautionary principle would seem to require that scientific uncertainty should not disadvantage those individuals who might be harmed by the activity in question, particularly if the harm is serious or irreversible as in the case of serious physical injury or death. This focus on the welfare of potential victims makes it possible to consider the precautionary principle in light of cost-benefit methodology, which is derived from the branch of economics concerned with the assessment of governmental policies on individual well-being (welfare economics).

III. Cost-Benefit Methodology and the Precautionary Principle

Cost-benefit methodology tends to be associated with the policy objective of allocative efficiency, because cost-benefit analysis is designed to identify efficient outcomes. The connection between cost-benefit analysis and allocative efficiency would seem to imply that cost-benefit methodology is inconsistent with the precautionary principle. Although the precautionary principle may be the allocatively efficient decision rule in some circumstances, the principle is unlikely to be efficient across the range of cases.¹⁷ But even if the precautionary principle is inefficient, it does not follow that the principle requires rejection of cost-benefit methodology.

Cost-benefit methodology, particularly when applied to risks that threaten human health, provides a useful way to quantify or measure the distributive effects of environmental regulations. Insofar as a distributive concern about the welfare of potential victims motivates the precautionary principle, then the welfare considerations relevant to the principle can be analyzed with cost-benefit methodology. (The point holds for the welfare considerations of future generations.)¹⁸ Once the relevant welfare considerations are identified, the regulatory decision rule can be evaluated in light of the distributive rationale for the precautionary principle.

This approach yields a well-defined decision rule capable of assessing costs and benefits in a manner sensitive to the distributive rationale for the precautionary principle.¹⁹ Whether this decision rule defensibly implements the precautionary principle cannot be determined due to the vagueness of the principle. In important contexts, however, the [31 ELR 11329] rule corresponds to the type of decision rule envisioned by proponents of the precautionary principle. To develop the rule, we must begin with the standard cost-benefit methodology for known risks, and then consider how the methodology might account for scientific uncertainty in light of the distributive rationale for the precautionary principle.

A. Cost-Benefit Methodology for Known Risks

Suppose a safety investment would reduce by 10% the risk of physical injury faced by potential victims. The risk reduction is the benefit of the safety investment. To compare this benefit to the cost of the safety investment, the benefit (reduced risk of physical injury) must be translated into money: the risk must be monetized. The procedure for monetizing the risk of physical injury is the aspect of cost-benefit methodology of particular relevance for the precautionary principle.

The notion of equating risk with money may seem incomprehensible. Often the risk of physical injury includes the possibility of death. How could a fatal risk be translated into a finite amount of money, if no amount of money can substitute for life?

To answer this question, the context must be specified in greater detail. Suppose the potential victims face a given probability of suffering a fatal injury. Suppose further that the potential victims do not directly benefit from the polluting activity and have no preexisting contractual relationship with the potential injurer (polluter). Finally, assume each potential victim is not concerned about leaving money for others in the event of her death. For these individuals, no amount of money could compensate them for the loss of life. Nevertheless, in most contexts each would be willing to accept some positive payment in order to face or assume the risk. More precisely, the minimum amount of money an individual would be willing to accept (WTA) in order to face the risk makes her indifferent between (1) the state of the world in which she does not face the risk and receives no money for facing the risk, and (2) the state in which she faces the risk and has a higher level of wealth due to receipt of the WTA risk proceeds. In effect, the WTA measure is the monetary equivalent of the benefit a potential victim must receive before she would assume the risk, illustrated by the higher wages workers demand to accept riskier jobs. The monetary benefit enhances the welfare of the potential victim (worker) to offset exactly the reduced welfare caused by the risk exposure.

The WTA measure depends on the magnitude of the risk (the probability and severity of injury) and the individual's wealth. If the risk were certain to kill the individual, for example, then she would not be willing to accept any amount of money in exchange for facing the risk: the WTA measure equals infinity. Note, then, how the WTA measure captures the notion that no one would exchange her life for money. Facing a small risk of death, however, is different than giving up one's life. Individuals routinely choose to face such risks, as in the employment context, because they receive some benefit from the risky activity that adequately offsets the chance of harm. This benefit is monetarily expressed by the WTA measure. Since the individual must receive that amount of money to offset the chance of harm, the WTA measure yields the individual's monetary assessment of the risk.

B. Cost-Benefit Methodology and Distributive Considerations

We can now consider the WTA measure in light of the concern for potential victims embodied in the precautionary principle. The WTA measure defines the baseline or status quo in terms of the potential victim's welfare level without the risk: either the individual does not face the risk, or she chooses to face it and receives the WTA risk proceeds. In both instances, the individual is not worse off than she would be in a world without the risk, an outcome that plausibly satisfies the distributive demands of the precautionary principle.

For this reason, the precautionary principle can be satisfied by cost-benefit regulations coupled with appropriate transfer payments. Cost-benefit outcomes correspond to the hypothetical contract that would be reached between the potential injurer and victim. The hypothetical contract, unlike actual contracts, does not involve any monetary transfer between the parties. But if such a transfer can be accomplished by some distributive mechanism, like the tax or tort systems, then cost-benefit regulations need not treat potential victims in a manner violative of the precautionary principle.

To see how the requirements of cost-benefit analysis can be replicated by a hypothetical contract between potential injurers and victims, suppose that for each precaution, there is a cost or burden B that would be incurred by the potential injurer. If the potential injurer does not take the precaution and instead creates the risk, she owes the associated WTA amount to the potential victim. The potential injurer would not incur the burden B if it would be less expensive to pay the WTA amount to the potential victim. By definition, the potential victim would agree to face the risk in exchange for the WTA risk proceeds. Hence the two parties would agree that the potential injurer can impose risks whenever the cost of eliminating the risk B exceeds the safety benefit expressed by the WTA measure: B > WTA. The parties also would agree that the potential injurer must take precautions satisfying the cost-benefit test (precautions for which B < WTA), as it would be cheaper for the potential injurer to take the precaution than pay the potential victim to face the risk. The hypothetical contract between the parties therefore corresponds to regulations that require potential injurers to take precautions satisfying a cost-benefit test.

Hypothetical contracts differ from actual contracts for various reasons, including the absence of any monetary transfer between the parties. Potential victims would agree to face a risk in exchange for receiving the WTA risk proceeds, but that transfer is not accomplished by a hypothetical contract. Without the transfer, why would a potential victim consent to the risk?

Any normative justification for cost-benefit analysis based exclusively on hypothetical compensation is troubling. Consequently, cost-benefit regulations become distributively defensible if coupled with some other transfer mechanism, ideally a lump-sum governmental transfer between households.²⁰ A lump-sum transfer does not involve administrative or other costs and does not affect the behavior of anyone who pays or receives benefits. By relying on such transfers, the government can convert hypothetical [31 ELR 11330] compensation into real compensation. No one loses under a cost-benefit regulation (as the potential victims are monetarily compensated for facing the risk), and some people gain. Everyone presumably would consent to the regulation (as reflected in the hypothetical contract), so cost-benefit regulations coupled with the appropriate compensatory payments can treat potential victims fairly.

Of course, the government cannot effectuate the type of lump-sum compensatory transfers that would ensure the fair treatment of potential victims.²¹ Tax transfers, for example, can address distributive issues based on income or wealth, but the monetary transfer contemplated by the WTA measure depends on important factors unrelated to income or wealth. An ex post damages award for the injury, like that provided by the tort system, also is inadequate, because such damages cannot fully compensate individuals who suffer fatal injuries. Moreover, the diffuse nature of environmental risks often makes it impossible to assign responsibility for the injury.

Absent tax transfers or other forms of compensation, cost-benefit regulations can yield the distributive inequity in which potential victims are not adequately compensated for facing the nonconsensual risks sanctioned by the regulations. With this distributive problem in mind, we can now consider its relation to scientific uncertainty and the precautionary principle.

C. The Precautionary Principle as Partial Solution to a Distributive Problem

The WTA measure is the payment a potential victim requires to be compensated for facing a specified threat to her bodily security. In situations of scientific uncertainty, the nature of that threat is unknown. What is the appropriate compensatory payment (WTA measure) under these conditions? An answer is provided by the precautionary principle, interpreted as a partial solution to the distributive problem that arises whenever potential victims are not adequately compensated for facing the risk in question.

Consider a situation of uncertainty in which scientific evaluation yields the following defensible risk assessments: the activity may pose a significant risk of serious physical injury (substantial risk), a very low risk of serious physical injury (moderate risk), or an insignificant risk (no risk). Suppose further that the weight of the scientific evidence supports the conclusion of moderate risk. The possibilities the risk is substantial or nonexistent, however, have adequate scientific foundation. How should regulators proceed?

The potential victim prefers to assume that the risk is substantial, because this assumption ensures that the ultimate resolution of the scientific uncertainty will not be disadvantageous. If the risk were assumed to be moderate or nonexistent, and turned out to be substantial, the potential victim would be undercompensated. For these same reasons, the potential injurer prefers to assume the risk is nonexistent, for if the risk is assumed to be moderate or substantial and turns out to be nonexistent, then the potential injurer will have expended money unnecessarily. Each party has a reasonable basis for his or her preferred assessment of the risk. How should the dispute be resolved?

The dispute could be resolved by an objective third-party assessment of the risk, suggesting that regulators should assume the risk is moderate (the conclusion supported by the weight of scientific evidence). The implications of this approach are troubling, however, in light of the precautionary principle's concern for potential victims. This approach does not give any special emphasis to the security interests of potential victims, because the objective assessment of the risk focuses only on scientific evidence. The potential victim, however, is the party forced to face the risk. The potential victim can determine the cost of facing a risk of known probability and severity of injury. That cost depends on a variety of factors determined by the potential victim (distaste for risk, the degree to which money substitutes for physical injury, and so on). Presumably, the cost can be determined only on the basis of reasonable factors, so that irrational fears of risk, for example, would not count. The cost still depends on various reasonable factors that could be altered to yield lower WTA measures (such as increased substitutability between money and health). Yet these aspects of the potential victim's assessment of the risk are not second-guessed, presumably because they are reasonable. Is it unreasonable for the potential victim to assume the worst-case scenario, given a defensible scientific basis for the conclusion of substantial risk? And if it is not unreasonable for the potential victim to monetize the risk in this way, what justifies rejection of that measure in the regulatory assessment of the costs and benefits?

These considerations suggest the desirability of a regulatory approach that adopts the potential victim's assumption that the risk is substantial, an approach that gives priority to the potential victim and her interest in physical security. This approach conforms to the precautionary principle, because the potential victims will assume the worst-case scenario, thereby ensuring that the uncertainty is not disadvantageous to them. In light of this assumption, the benefit of risk reduction is determined by aggregating each potential victim's WTA measure for the substantial risk. That benefit can then be compared to the cost of safety precautions in a cost-benefit evaluation of the regulation.

To be sure, this regulatory approach can be criticized on the ground that the potential victim's preferred assessment of the risk — the worst-case scenario in the context under consideration — often will mischaracterize the risk. At times, the best-case scenario will correctly characterize the risk. Most of the time, the scenario supported by the weight of scientific evidence presumably will be correct. But no approach will provide the correct characterization always. Risk assessments under conditions of scientific uncertainty will not be correct each and every time. Mistakes will be made. Mistakes create error costs, which will be borne by potential injurers, potential victims, or both parties. Different regulatory approaches under conditions of uncertainty therefore yield different distributive outcomes, and so a choice among approaches can be made on distributive grounds.

Consider further the distributive implications of a regulatory approach based on the precautionary principle. Suppose a regulation is based on the potential victim's reasonable assessment that the activity poses a substantial risk. Suppose the risk in fact is moderate. Regulations based on the assumption of substantial risk will produce error costs; potential injurers will be forced to take more costly precautions [31 ELR 11331] than would be required in a world of complete scientific knowledge. But what is unfair about this outcome? Potential victims face a lower level of nonconsensual risk than they otherwise would face in a world of no scientific uncertainty, as regulations for substantial risks impose more demanding safety requirements than do regulations for moderate risks. Any amount of nonconsensual risk, however, makes potential victims worse off than they would be in a world without the risk, unless they are adequately compensated by some other means (like tax transfers or tort damages). Absent such compensation, partial compensation is afforded by the precautionary principle. The approach compensates potential victims or is advantageous to them only insofar as it makes potential injurers bear the error costs of uncertainty. However, potential injurers directly benefit from the potentially hazardous activity and do not face the prospect of physical injury. Moreover, potential injurers can avoid the more burdensome regulatory requirements by financing the research needed to reduce the uncertainty.

Cases of scientific uncertainty therefore pose a particular type of distributive problem. In such cases, regulatory errors are inevitable, and someone must bear the associated costs. A regulatory approach based on the precautionary principle places the cost on those who directly benefit from the potentially hazardous activity, while seeking to minimize the impacts of uncertainty on those parties who might be physically injured. This solution to the distributive problem seems defensible, and once the precautionary principle is conceptualized in this way, its method of implementation becomes apparent.

D. The Precautionary Principle and Context

Any distributive rationale for the precautionary principle must account for the way in which distributive considerations change across a broad range of contexts. So far we have assumed the potential injurer derives all of the benefit from the activity, whereas the potential victim faces an unknown risk of suffering severe physical injury. The distributive problem arises because the potential victim is entitled to compensation for facing the risk (captured by the WTA measure), but does not receive such compensation from the potential injurer nor from other transfer programs like the tax or tort systems. In such a setting, the distributive problem is partially redressed by the precautionary principle, which ensures that the uncertainty does not disadvantage potential victims.

In actuality, the distributive problem is less severe than the foregoing analysis suggests. The analysis so far has assumed the potential injurer unilaterally imposes a risk on a potential victim who poses no risk to the potential injurer. Not all risk situations are of this type. For situations in which two parties impose risks on one another, the nature of the distributive problem can fundamentally change.

Consider the extreme case of perfect reciprocity in which the two individuals involved in the risky interaction (such as automobile drivers) are identical in all relevant respects, including the degree of risk that each imposes on the other, the severity of injury threatened by the risk, the reasons for the risky behavior, and so on. As there are no relevant individual differences, the two individuals can be conceptualized as one entity. Whatever safety precautions required of one individual will be required of the other. Whatever safety benefits accrue to one person will accrue to the other. Each individual effectively pays for her own safety, just as a single individual pays for her own self-protection. Consequently, each individual monetizes risk in terms of the willingness-to-pay measure rather than the WTA measure.²² Rather than receiving monetary compensation for facing the risk (the WTA measure), these individuals must make the expenditures required for risk reduction.

In these contexts, the individuals threatened by the risk no longer prefer to adopt the worst-case scenario that reasonably can be maintained. Why pay for safety investments that may be unnecessary? These individuals also would not assess the risk in terms of the best-case scenario, for if the risk materializes it will harm them. Individuals who are simultaneously a potential injurer and victim bear the error costs created by scientific uncertainty, which take the form of overly costly precautions or an excessive risk of injury. Because they bear the error costs, these individuals prefer risk assessments that minimize the cost of error. Risk assessments of this type are likely to be set on the basis of the weight of scientific evidence, which presumably yields the right answer more frequently than other approaches.

Considerations of reciprocity explain why an outright ban of any potentially hazardous activity is not in the best interests of potential victims. Anyone threatened by a potentially hazardous activity also participates in other activities that may pose threats to others. The losses someone might suffer as a potential victim must be considered in relation to the gains she receives as a potential injurer. Banning any potentially hazardous activity therefore is likely to make everyone worse off compared to a regulatory regime in which each potential victim is sometimes undercompensated for facing potential hazards. For this conclusion to hold, though, the legal regime must regulate potentially hazardous activities in a manner adequately protective of potential victims. Such protection is provided by the precautionary principle.

Considerations of reciprocity also show why the distributive rationale for the precautionary principle yields different regulatory requirements for different contexts. In situations of perfect reciprocity, potential victims prefer risk assessments that minimize the cost of error. This type of risk assessment is based on all the available scientific evidence. For contexts involving risks monetized with the WTA measure, potential victims prefer the risk assessment corresponding to the worst-case scenario, yielding the type of decision rule apparently contemplated by proponents of the precautionary principle.

Once the uncertainty has been resolved in this manner, the cost of the risk is defined and regulators can apply the decision rule for known risks. Thus, the activity can be regulated with cost-benefit analysis or any other method of comparing costs and benefits. Moreover, these methods of risk regulation need not be inconsistent with the victim-oriented [31 ELR 11332] rationale for the precautionary principle, because the stringency of the regulatory standard can be altered to accommodate distributive concerns.

Much like the distributive rationale for the precautionary principle stems from the fact that individuals who suffer serious physical injury or death are not adequately compensated by an ex post damages award or other forms of compensation, that fact also provides a distributive rationale for altering the regulatory standard for known risks. For contexts in which potential victims are not adequately compensated for facing the risk, some compensation or protection can be provided by a regulatory standard that reduces risk below the level that would be obtained when potential victims are fully compensated for facing the risk. Situations of full compensation correspond to cost-benefit risk levels as illustrated by the hypothetical contract described earlier. Consequently, if the regulatory standard reduces risk below the cost-benefit level, the added risk reduction serves as a form of compensation to potential victims that somewhat offsets the inadequacy of other forms of compensation. Just as the precautionary principle can be a partial solution to a distributive problem posed by potentially hazardous activities, the same purpose can be served by the stringency of a regulatory standard based on a comparison of costs and benefits. This point further shows why there need not be any fundamental incompatibility between the precautionary principle and cost-benefit methodology.²³

IV. Science, "Reasonable" Risk Assessments, and the Rules of International Trade

The precautionary principle, understood in distributive terms, translates into a decision rule that monetizes or quantifies the cost of potential hazards in any reasonable manner preferred by potential victims. This rule critically depends on the concept of reasonableness. Such a concept is likely to be developed by international trade treaties.

A relation between the precautionary principle and the rules of international trade is to be expected in light of the trade disputes that have arisen and are likely to arise over regulations based on the precautionary principle. When the EU blocked importation of U.S. hormone-treated beef, did it properly invoke the precautionary principle or was it trying to protect European agricultural interests from foreign competition? The European concern over genetically modified organisms raises a similar question. The difficulty of divining the true intentions of regulators, and the stakes involved, makes it easy to understand why regulations based on the precautionary principle are likely to be at the center of trade disputes.

Not surprisingly, the rules of international trade already address the reasonableness component of the precautionary principle. The SPS Agreement, which governs sanitary and phytosanitary measures necessary for the protection of human, animal, or plant life or health that directly or indirectly affect international trade, lets WTO Members adopt a higher level of protection than that which would be achieved by international standards if there is "scientific justification."²⁴ This provision means that measures must be "based on" a risk assessment that "shall take into account available scientific evidence" of risk.²⁵ This requirement, as interpreted by the WTO Appellate Body in the hormones case, can serve as the basis for evaluating the reasonableness of the risk assessment implied by the distributive rationale for the precautionary principle:

We do not believe that a risk assessment has to come to a monolithic conclusion that coincides with the scientific conclusion or view implicit in the SPS measure. The risk assessment could set out both the prevailing view representing the "mainstream" of scientific opinion, as well as the opinions of scientists taking a divergent view. [The SPS Agreement] does not require that the risk assessment must necessarily embody only the view of a majority of the scientific community. In some cases, the very existence of divergent views presented by qualified scientists who have investigated the particular issue at hand may indicate a state of scientific uncertainty. Sometimes the divergence may indicate a roughly equal balance of scientific opinion, which may itself be a form of scientific uncertainty. In most cases, responsible and representative governments tend to base their legislative and administrative measures on "mainstream" scientific opinion. In other cases, equally responsible and representative governments may act in good faith on the basis of what, at a given time, may be a divergent opinion coming from qualified and respected sources. By itself, this does not necessarily signal the absence of a reasonable relationship between the SPS measure and the risk assessment, especially where the risk involved is life-threatening in character and is perceived to constitute a clear and imminent threat to public health and safety. Determination of the presence or absence of that relationship can only be done on a case-to-case basis, after account is taken of all considerations rationally bearing upon the issue of potential health effects.²⁶

This interpretation of the SPS Agreement appears to sanction the type of risk assessments that regulators would adopt when implementing the distributive version of the precautionary principle. Under this approach, regulators adopt the risk assessment reasonably preferred by potential victims. The reasonableness of any given risk assessment presumably is a scientific matter, and regulators could evaluate the adequacy of scientific support in terms of the type of scientific evidence required by the SPS Agreement. Thus, if potential victims prefer to assess risk in terms of the worst-case scenario, then regulators would base the risk assessment on the worst-case scenario adequately supported by science. The scientific support can be adequate even though the majority of the scientific community would adopt a different risk assessment. In other contexts, potential victims prefer to assess risk in terms of all the available evidence, and so regulators could adopt the risk assessment based on mainstream scientific opinion. Such variable use of scientific evidence on a case-by-case basis appears to be consistent with the SPS Agreement.

Thus, the SPS Agreement illustrates how the issue of reasonableness and the associated role of scientific input are likely to be addressed by international trade treaties. In applying the reasonableness component of the precautionary [31 ELR 11333] principle, then, regulators should be able to derive sufficient guidance from the rules of international trade.

V. Conclusion

The vagueness of the precautionary principle makes it hard to implement, indicating the desirability of proceeding in steps. The first step involves identification of the core aspects of the precautionary principle likely to be included in most plausible versions, followed by an attempt to turn those features into a well-defined decision rule. Subsequent steps of implementation can focus on more controversial aspects of the principle.

Whatever the full scope of the principle may be, it would seem to minimally express the idea that scientific uncertainty is not a sufficient reason for foregoing precautionary measures. Such measures can serve various objectives, the least controversial of which involves the protection of potential victims. If uncertainty cannot bar the protection of potential victims, then one plausible interpretation of the precautionary principle is that scientific uncertainty should not disadvantage potential victims.

This norm can be turned into a decision rule by applying conventional cost-benefit methodology and the scientific techniques of risk assessment. The decision rule monetizes risk with a nonconventional procedure, defining the cost of risk in terms of the risk assessment reasonably preferred by potential victims. This procedure will not be allocatively efficient, because it is not designed to minimize the error costs necessarily created by the uncertainty. Instead, the procedure implements the precautionary principle by ensuring that potential victims are not disadvantaged by scientific uncertainty.

Although the decision rule is not allocatively efficient in important settings, such inefficiency should not be controversial as a matter of welfare economics, the branch of economics that includes cost-benefit methodology. Welfare economics provides no basis for challenging distributive objectives. Welfare economists recognize that distributive concerns (tax transfers from the rich to poor, for example) can justify departures from allocatively efficient outcomes. Typically, the tax system distorts economic behavior, so the pursuit of distributive objectives via the tax system often requires the sacrifice of efficiency to attain equity (the "efficiency-equity trade off"). Rather than challenging the distributive objective, welfare economics merely requires that the distributive objective be attained in the least costly fashion. This requirement is satisfied by a decision rule that monetizes risk in terms of the risk assessment reasonably preferred by potential victims.

The distributive objective embodied in the precautionary principle cannot be attained by the standard compensatory or transfer mechanisms — the tax and tort systems — for various reasons, including the diffuse nature of environmental risks and the prospect of serious, irreversible harm, like death. Protecting the welfare of potential victims from the adverse consequences of scientific uncertainty therefore must be provided by some other means. Such protection can be provided by the regulatory rule itself. By accounting for the distributive concerns of the precautionary principle, the regulatory rule requires precautionary measures that ensure potential victims are not disadvantaged by scientific uncertainty. Attaining the distributive objective in this manner is the least costly option available. The cost-effective precautions required by a regulatory rule based on the precautionary principle therefore satisfy the requirements of welfare economics. The regulatory rule also is consistent with President William J. Clinton's executive order requiring federal agencies to consider issues of "equity" and "distributive impacts" when employing cost-benefit analysis.²⁷

In important contexts, though, the regulatory rule selects allocatively efficient precautions. No doubt many proponents of the precautionary principle will be dissatisfied with this aspect of the rule. But to justify more exacting regulations, these proponents must rely on some rationale other than the protection of potential victims. Perhaps there is such a rationale capable of gaining the requisite consensus. Absent such a rationale, proponents of the precautionary principle ought to be satisfied with a regulatory approach that addresses at least one of their core concerns.

So too, many will find this version of the precautionary principle to be overly onerous. But these critics must explain why alternative approaches do not result in potential victims being unfairly disadvantaged by scientific uncertainty, or why the precautionary principle means something else. My objective here is not to defend a particular conception of the precautionary principle. Rather, for the debate over the precautionary principle to be fruitful, it must move from generalities to the particularized guidance required for principled regulation. A concrete proposal for implementation is a step in that direction.

1. See COMMISSION OF THE EUROPEAN COMMUNITIES, COMMUNICATION FROM THE COMMISSION ON THE PRECAUTIONARY PRINCIPLE 11 (2000), available at http://europa.eu.int/comm/dgs/health_consumer/library/pub/pub07_en.pdf (last visited Sept. 11, 2001) [hereinafter COMMUNICATION ON THE PRECAUTIONARY PRINCIPLE]. The European Commission initiates proposals for all new legislation in the EU. The Council of the European Union has endorsed "the broad lines" of the Communication on the Precautionary Principle. See Council Resolution on the Precautionary Principle, Bull. Quotidien Eur. No. 7860, Dec. 10, 2000.

2. See, e.g., INTERPRETING THE PRECAUTIONARY PRINCIPLE (Timothy O'Riordan & James Cameron eds., 1994); PERSPECTIVES ON THE PRECAUTIONARY PRINCIPLE (Ronnie Harding & Elizabeth Fisher eds., 1999); PROTECTING PUBLIC HEALTH AND THE ENVIRONMENT: IMPLEMENTING THE PRECAUTIONARY PRINCIPLE (Carolyn Raffensperger & Joel. A. Tickner eds., 1999); THE PRECAUTIONARY PRINCIPLE AND INTERNATIONAL LAW: THE CHALLENGE OF IMPLEMENTATION (David Freestone & Ellen Hey eds., 1996).

3. See COMMUNICATION ON THE PRECAUTIONARY PRINCIPLE, supra note 1.

4. When in Doubt, BUS. EUR., Feb. 9, 2000, at 6.

5. Rio Declaration on Environment and Development, United Nations Conference on Environment and Development, U.N. Doc. A/CONF. 151/5 (1992), reprinted in 31 I.L.M. 874, 879 (1992).

6. Cf. Christopher D. Stone, Is There a Precautionary Principle?, 31 ELR 10790 (July 2001) (questioning whether there is a precautionary principle, given the range of possible meanings and lack of customary expectation regarding its obligations).

7. Regulators could select the risk assessment that minimizes the cost of error, which would preclude application of the precautionary principle or render it meaningless. See discussion infra. But if the cost-benefit exercise accounts for distributional concerns, then any distributive rationale for the precautionary principle would impose a distributional constraint on the evaluation of costs and benefits, making the precautionary principle analytically distinct from the cost-benefit exercise. Cf. note 27 infra and accompanying text (citing executive order that enables federal regulators to account for distributive concerns when applying cost-benefit analysis).

8. See COMMUNICATION ON THE PRECAUTIONARY PRINCIPLE, supra note 1, at 18 (observing that "reliance on the precautionary principle is no excuse for derogating from the general principles of risk management," including examination of the benefits and costs of action).

9. Numerous examples of federal legislation requiring cost-benefit analysis are provided in American Textile Mfrs. Inst., Inc. v. Donovan, 452 U.S. 490, 510-11 & n.30, 11 ELR 20736, 20741 & n.30 (1981). Prominent recent examples include the Safe Drinking Water Act (SDWA), 42 U.S.C. § 300g-1(b)(3)(C), ELR STAT. SDWA § 1412(b)(3)(C) and the Toxic Substances Control Act (TSCA), 15 U.S.C. § 2605(c)(1), ELR STAT. TSCA § 6(c)(1). Ten states currently require cost-benefit analysis of all proposed agency rules, and seven states require cost-benefit analysis of selected rules. ROBERT W. HAHN, STATE AND FEDERAL REGULATORY REFORM: A COMPARATIVE ANALYSIS 3 (American Enterprise Inst.-Brookings Inst. Joint Ctr. for Regulatory Studies, Working Paper No. 98-3, 1998), available at http://www.aei.brook.edu/publications/working/working_98_03.pdf (last visited Sept. 11, 2001). President Ronald Reagan's executive order requiring agency use of cost-benefit analysis is Exec. Order No. 12291, 3 C.F.R. 127 (1981), reprinted in 5 U.S.C. § 601 (1988), ADMIN. MAT. 45025. President George H.W. Bush did not repeal this order. President William J. Clinton subsequently repealed the order and replaced it with another requiring the agency use of cost-benefit analysis. See Exec. Order No. 12866, 3 C.F.R. 638 (1993), reprinted in 5 U.S.C. § 601 (1994), ADMIN. MAT. 45070. The differences between the two orders are described in Richard H. Pildes & Cass R. Sunstein, Reinventing the Regulatory State, 62 U. CHI. L. REV. 1, 6-7 (1995).

10. E.g., Frank B. Cross, Paradoxical Perils of the Precautionary Principle, 53 WASH. & LEE L. REV. 851 (1996); Editorial, Fear of the Future, WALL ST. J., Feb. 10, 2000, at A18 (arguing that the precautionary principle could result in a ban of clearly desirable risky activities like driving).

11. The history of this trade dispute is described in George H. Rountree, Raging Hormones: A Discussion of the World Trade Organization's Decision in the European Union-United States Beef Dispute, 27 GA. J. INT'L & COMP. L. 607, 611-12, 632-33 (1999).

12. Agreement on the Application of Sanitary and Phytosanitary Measures, Apr. 15, 1994, Marrakesh Agreement Establishing the World Trade Organization, reprinted in 1994 WL 761483, at *90-120 [hereinafter SPS Agreement].

13. See WTO Appellate Body Report on EC Measures Concerning Meat and Meat Products (Hormones), Report from the Appellate Body, WT/DS26/AB/R, WT/DS48/AB/R (Jan. 16, 1998) [hereinafter Hormones].

14. Rio Declaration on Environment and Development, supra note 5.

15. See COMMUNICATION ON THE PRECAUTIONARY PRINCIPLE, supra note 1.

16. Cf. Stonc. supra note 6, at 10799 (concluding that "there are several reasons to doubt that any single principle can be fashioned more trenchant than 'take care'").

17. See Christian Gollier et al., Scientific Progress and Irreversibility: An Economic Interpretation of the Precautionary Principle, 75 J. PUB. ECON. 229 (2000) (identifying a class of restrictive but plausible conditions under which scientific uncertainties justify an immediate reduction of the consumption of a potentially toxic substance); Giovanni Immordino, Looking for a Guide to Protect the Environment: The Development of the Precautionary Principle (unpublished manuscript, 1999) (concluding that the precautionary principle is unlikely to be an efficient rule of thumb), Social Science Research Network Electronic Library, available at http://www.ssrn.com (last visited Sept. 11, 2001).

18. Some proponents of the precautionary principle reject cost-benefit methodology on the ground that it does not adequately respect the welfare of future generations. There is not a persuasive economic rationale for discounting the welfare of future generations relative to the current generation, however, nor is such discounting ethically justified. See Lisa Heinzerling, The Temporal Dimension in Environmental Law, 31 ELR 11055 (Sept. 2001); Richard L. Revesz, Environmental Regulation, Cost-Benefit Analysis, and the Discounting of Human Lives, 99 COLUM. L. REV. 941 (1999).

19. The arguments in this section are developed more fully in Mark Geistfeld, Reconciling Cost-Benefit Analysis With the Principle That Safety Matters More Than Money, 76 N.Y.U. L. REV. 114 (2001).

20. E.g., RICHARD W. TRESCH, PUBLIC FINANCE: A NORMATIVE THEORY 39 (1981); HAL R. VARIAN, MICROECONOMIC ANALYSIS 405 (3d ed. 1992).

21. For more complete discussion of this issue, see Geistfeld, supra note 19, at 153-58.

22. The maximum amount of money the individual would be willing to pay (WTP) to eliminate a risk makes her indifferent between (1) the state of the world in which she does not face the risk and has a lower level of wealth due to payment of the WTP amount, and (2) the state in which she faces the risk and has a higher level of wealth. As this definition implies, the WTP measure monetizes risk for contexts in which potential victims must pay for the safety precautions. Consumers, for example, rationally monetize product risks in this fashion, since the cost of any safety investments (like airbags in automobiles) are included in the product's price.

23. See generally Geistfeld, supra note 19.

24. SPS Agreement, supra note 12, art. 3.3.

25. Id. arts. 5.1, 5.5.

26. Hormones, supra note 13, P194. For a thoughtful discussion of the scientific input required by the SPS Agreement, and its relation to democratic deliberation, see Robert Howse, Democracy, Science, and Free Trade: Risk Regulation on Trial at the World Trade Organization, 98 MICH. L. REV. 2329 (2000).

27. See Exec. Order No. 12866, § 1(a), 3 C.F.R. 638, 639 (1993), reprinted in 5 U.S.C. § 601 (1994), ADMIN. MAT. 45070.