The AEI-Brookings Joint Center for Regulatory Studies held a conference on October 4, 1999, to discuss the new Joint Center working paper by Cass Sunstein on the constitutionality of the Clean Air Act. The paper argues that the Environmental Protection Agency (EPA) should specify the range of benefits that it believes will result from a proposed rule. The agency should also consider at least two alternatives to the proposed rule, and should explain why the proposed rule is preferable to either alternative. Provided EPA reasonably performs these tasks, the courts should uphold the regulations.

The opening presentation by Cass Sunstein was followed by critiques from three discussants: Christopher DeMuth of AEI, Lisa Heinzerling of Georgetown University, and Jonathan Wiener of Duke University.

Cass Sunstein, University of Chicago

Mr. Sunstein argues that the May 1999 decision by the D.C. Court of Appeals, American Trucking Association v. EPA, is not the right way to ensure better and more transparent policy analysis. The American Trucking court held that the EPA's new air quality standards represent an unconstitutional delegation of legislative power, and based its decision on a new interpretation of a legal principle known as the nondelegation doctrine. Mr. Sunstein questions this ruling for three reasons. First, the ruling did not apply the nondelegation doctrine in the traditional sense because the court did not order Congress to define the limits of the EPA's legal authority to set standards. Instead, the court ordered the EPA to define its own limits. Second, it is unlikely that federal judges can successfully enforce the nondelegation doctrine because the lines between legislation and execution are not clear. Third, agencies operating under clear congressional guidance are not necessarily more competent and reliable than those that are governed by comparatively open-ended congressional guidance. Mr. Sunstein concludes that the nondelegation doctrine is not the right way to refine the existing approach to solving environmental problems.

Mr. Sunstein explains that the American Trucking decision does, however, address an important question that the EPA has not yet satisfactorily answered: Why did the EPA choose that specific standard rather than one a bit higher or lower? He suggests that the EPA should meet two requirements when it issues national ambient air quality regulations. First, EPA should specify the range of benefits it expects from each proposed rule, and from at least two alternative approaches, one stricter and one more lenient than the chosen regulation. In the process, the agency should describe the level of "residual risk" under all three options. Second, it should convincingly justify its choice of the regulation over the other alternatives. If the EPA performs those tasks adequately, the court should uphold the regulation.

Mr. Sunstein makes two final recommendations to the EPA and the Congress related to establishing regulations on particulates and ozone. First, he recommends that the EPA make "health-health comparisons," also known as "risk-risk tradeoffs," as part of the benefit analysis. Such comparisons may reveal that efforts to decrease a target risk may actually increase other health risks. The American Trucking court, for example, remanded EPA's ozone standard for EPA's failure to consider the health benefits of tropospheric ozone. Tropospheric ozone screens out harmful ultraviolet radiation that increases the risk of cataracts and skin cancer. Second, Mr. Sunstein recommends that Congress should permit the EPA to consider costs when setting national air quality standards, rather than to set the standards on a health-only basis. He asserts that "it is impossible to assess safety in a cost-vacuum."

Christopher DeMuth, American Enterprise Institute

Mr. DeMuth faults Mr. Sunstein's solution of "agency minimalism," referring to the minimum hurdles an agency must surmount to survive judicial review under either constitutional or Administrative Procedure Act standards. He does not think the EPA would propose defensible air quality standards if it had more information about affected populations or a better understanding of the relative magnitude of various health problems. In addition, Mr. Sunstein's recommendation that the EPA should specify a floor and a ceiling level of standards is too similar to the analysis many agencies already conduct to survive judicial review. Mr. DeMuth believes that "the essential problem of regulation, which amounts to a constitutional problem, is that the agencies act without a spending constraint."

Mr. DeMuth asserts that a statute instructing an agency in general terms to protect the public health should be considered constitutional if it requires agencies to pursue such protection "up to a point where going further would not be worth the costs." He prefers that the agencies adopt this principle rather than enforce the principle through the courts, and is therefore comfortable with the court's decision to remand the regulations to the EPA. He concludes that Mr. Sunstein's recommended use of benefit analysis will not solve the problem of agency abuse of discretion. Benefit-cost analysis, on the other hand, could reduce such abuse.

Lisa Heinzerling, Georgetown Univesity

Ms. Heinzerling criticizes Mr. Sunstein's proposal that the EPA should consider a less strict and a more strict alternative to its preferred option. She explains that the proposal does not limit the discretion of the agency since there are no specific criteria for setting the level of the more lenient and the stricter alternative standards. Furthermore, neither the court nor Mr. Sunstein pays adequate attention to the EPA's statement about its own constraints on discretion. The EPA chose the standards that were within the ranges suggested by its scientific advisory committee. The court and Mr. Sunstein also pay insufficient attention to the details and complexities of rule making. Both discuss the EPA's chosen ambient standards for ozone and particulate matter as if choosing the standard was the EPA's only decision. The EPA made decisions about many other components of the standards, including indicators of the size of particulate matter and averaging times. If examination of the benefits of alternatives requires consideration of the various components of the standards, the complexity of the examination increases -- especially if the components also involve uncertainties. She also states that waiting for the EPA to present this complex analysis will take a long time. During that time, many people may die.

Ms. Heinzerling also comments on the court's suggestion that the EPA set standards by establishing a "generic unit of harm." She finds suggestion puzzling. The idea that individuals make risk-based decisions according to their own determination of generic units of harm seems unlikely. Moreover, the court's suggestion that generic units of harm be established based in part on quality-adjusted life-years is problematic insofar as it would ration regulatory interventions based on age, health, and disability status.

Jonathan Wiener, Duke University

Mr. Wiener agrees with Mr. Sunstein that the modern version of the nondelegation doctrine does not resemble the two traditional versions of the doctrine. According to one traditional version, illustrated by the Supreme Court's decision in Schechter Poultry Corp. v. United States, the court requires the Congress to remedy a statute that delegates excessive agency authority. According to the second version, the court remedies the excessive delegation itself by supplying specific limits on the agency's legal authority, as the Supreme Court did in Industrial Union Department, AFL-CIO v. American Petroleum Institute (1980), also known as the "Benzene" case. According to the modern version of the nondelegation doctrine, illustrated by International Union, UAW v. OSHA (1991), the court asks the agency to provide the limiting principles to curb excessive delegation. According to Mr. Wiener, the decision by the American Trucking court is consistent with the modern version of the nondelegation doctrine.

Mr. Wiener suggests that the real purpose of the modern version is to ensure that the agency provide a "reasonable or permissible interpretation" of the statute, not to require Congress to make every decision. He further suggests that the modern version converges to the doctrine in the Supreme Court's landmark 1984 case Chevron v. Natural Resources Defense Council. According to the Chevron ruling, the agency gets the first crack at interpreting ambiguous statutory language that authorizes agency action. The courts review the agency's interpretation to ensure that it is not arbitrary. Mr. Wiener believes that the American Trucking court's decision to remand EPA's air quality standards manifests such an act of judicial deference-at least as compared with the courts' much less deferential rulings in Schechter Poultry and in Benzene-and is not, as Mr. Sunstein seems to suggest, an act of judicial hubris.

Mr. Wiener agrees with Mr. Sunstein that agencies must take "risk-risk tradeoffs" into account. He points out that the American Trucking court ruled unanimously in Part III.B of the decision that the EPA acted arbitrarily and violated the Clean Air Act by ignoring the countervailing health risks of reducing ozone levels. He suggests that the court could have written a much simpler and potentially more enduring opinion if it had remanded both the ozone and the particulate matter standards on this risk-risk basis, without applying the nondelegation doctrine. If so, the court would not have had to question the constitutionality of the Clean Air Act or, inevitably, involve the Supreme Court. Moreover, attention to such risk-risk tradeoffs could provide the limiting principle needed to avoid violating the nondelegation doctrine: the agency would set its standard to minimize the overall risk to health and the environment, including both the risks of not regulating and the countervailing risks of regulating.