Chapter 3
Risk Management and Regulatory Decision-Making
This report brings a risk management perspective to issues surrounding risk assessment, risk communication, and risk reduction. Risk assessment can provide valuable information to those who set environmental, health, and safety regulatory priorities, allocate resources within regulatory agencies, and make regulatory decisions. Technical risk assessments seldom set the regulatory agenda, however, because of the different regulatory goals specified in the various environmental statutes and the different ways in which the public perceives risks.
This section presents five conclusions that have emerged from our examination of risk management and regulatory decision-making:
• The complex and often confusing process of communicating information about risks to diverse affected parties must be improved.
• Decisions about how to allocate resources to reduce risks can be made and explained partly on the basis of risk comparisons.
• The use of "bright lines" which distinguish between contaminant emissions and exposures associated with negligible risk levels and those associated with unacceptable risk levels, needs to be clarified.
• Moving from command-and-control regulation to nonregulatory approaches to risk reduction can increase both efficiency and effectiveness.
• Criteria for judicial review, a common element in major regulatory actions, should be reaffirmed.
This section offers recommendations on each of those topics in the hope of contributing to the evolution and improvement of risk-based decision-making.
Communicating and Comparing Risks
Effective risk communication is critical to successful implementation of the Risk Management Framework. Risk communication engages both the communicator and the audience in listening and in explaining information and opinions about the nature of risk and other topics that express concerns, opinions, or reactions to risk messages (NRC 1989). Various proposals to increase the transparency of risk assessments, which entails revealing and characterizing the assumptions, uncertainties, default factors, and methods used to estimate risks and to require the use of risk comparisons have been considered by Congress. As Congressman Thomas Bliley (R-VA) noted in his comments to the Commission, risk managers and the public need to understand, and have the right to know, what the weight of scientific evidence says about a health or environmental risk. Some risk comparison proposals would compel agencies to compare a risk to be regulated with other risks also regulated by the agency and other health risks experienced by the public.
This section discusses the need for better risk communication with the public, including the use of risk comparisons and the use of a common metric for describing exposures associated with different types of adverse health effects. A separate section, Comparative Risk Analysis for Risk Management Priority Setting, follows on page 46 and discusses the process of comparing and ranking risks to identify priorities and make resource allocations.
Identifying Risk Communication Needs
Finding
Stories abound of misunderstandings caused by poor communication about risks and risk reduction proposals. After a decade of research at leading universities and experience at all levels of government, much has been learned about how to enhance effective risk communication to gain the confidence of stakeholders, incorporate their views and knowledge, and influence favorably the acceptability of risk assessments and risk management decisions. That knowledge is not reflected commonly in practice, however.
Recommendation
Regulatory agencies should adopt comprehensive risk communication programs that emphasize both the learning and explaining activities of communication. These programs should provide research on risk communication messages, train risk managers and others engaged in communicating risk, and include risk communication funding, objectives, and evaluation in risk management plans.
The Commission’s Risk Management Framework (Section 2) is built on continuous involvement of stakeholders and respectful learning from them. Effective risk communication is an essential ingredient in the success of the Framework, especially in the problem identification and options stages in the process.
Risk assessors now recognize that a community’s response to learning that a local industry has put them at risk through release of pollutants tends to include a sense of outrage that inevitably magnifies their perception of risk. Studies of the differences between expert and public perceptions of risk have identified many of the factors that contribute to outrage (Sandman 1992). Those factors include involuntary exposures, lack of previous knowledge of the risk, and dread of effects and severe consequences (Slovic 1987). People factor in their perceived personal potential benefit and harm. For example, in comments to the Commission, the Argonne National Laboratory’s Environmental Assessment Division pointed out that the dreaded and unfamiliar nature of environmental (non-medical) exposure to ionizing radiation has evoked even greater fear of radiation than of chemicals. A growing body of research provides some guidance on communicating risk information effectively, as detailed in a report prepared for the Commission by David McCallum (see Appendix A7). Our discussion here is not comprehensive; rather, it is intended to indicate the importance of effective risk communication to overcome the potential for mistakes and misunderstandings.
People interpret and use new information in the context of their existing beliefs. We need a basic understanding of the exposure, effects, and mitigation processes relevant to making decisions about a hazardous process, product, or site. Meeting those needs through risk communication should involve well-tested methods; an untested communication should no more be released than an untested product (Morgan et al. 1992). Risk communication means both listening and speaking, and risk communicators should learn about the concerns and values of their audience, their relevant knowledge, and their experience with risk issues. Stakeholders often have important knowledge of sources and patterns of exposure that analysts will need to integrate into a risk assessment. The degree to which information provided by stakeholders is incorporated into risk assessment and risk management decisions may enhance the prospects for trust, a key to effective communication and cooperative action. By listening, risk communicators can craft risk messages that better reflect the perspectives, technical knowledge, and concerns of the audience. Risk communicators must be prepared to explain and answer questions about any specific, relevant tests or surveys done in the community regarding incidences of illness or uptake of pollutants; they cannot rely on general models.
Effective communication must begin before important decisions have been made, as emphasized in the Commission’s Framework for Risk Management. It can be facilitated in communities by citizen advisory panels, such as those supported by the EPA Superfund program and by the Department of Energy. Many corporations work continuously with citizen advisory panels in their communities. For example, Phillip Lewis, Vice President for Health and Safety at Rohm & Haas Company, noted in testimony to the Commission that the citizen advisory panels associated with each of his company’s plants generate a better understanding of the questions and concerns of the community and an opportunity for the company to test risk communication messages before using them with the general public. Of course, not all citizen advisory panels develop a trusting relationship with a company they are advising or are trusted by their communities.
With the growing use of risk assessments and risk estimates by regulatory agencies, there is a need to increase the public understanding and credibility of such information. The media play an important role because they can heavily influence public perceptions about risks and they can instigate concern or draw attention to neglected or underappreciated risks. Of course, broadcast media and print media have different agendas, and must be used in different ways to most effectively transmit information. Material with visual impact will appeal to television reporters. Print journalists are most concerned with who and how many people are affected, the severity of the potential impacts, any possible non-compliance with regulations, and the cost of damage, repairs, or remedies. They are particularly interested in conflicting views, especially among qualified scientists, about the nature and severity of the risks and about the costs and benefits of the remedies. Communicating with the media should be a part of a good risk communication plan. As with the public, honesty and accuracy are essential.
Agencies and Congress have emphasized the importance of improving the quality of risk assessments but have given less attention to the need for training and educating risk assessors and risk managers in communicating information about risk. Comprehensive risk communication programs that stress listening as well as explaining need to be established in regulatory agencies. Training risk assessors and risk managers in risk communication and testing risk communication messages should have as high priority as every other part of the risk management process. Specific communication objectives related to awareness and involvement of stakeholders should be identified in risk management plans, with appropriate methods for evaluating the effectiveness of communication. The National Research Council made the case in Improving Risk Communication that "risk managers need to consider communication as an important and integral aspect of risk management" (NRC 1989). The recent National Research Council report Understanding Risk strengthens and supports our recommendation to place public stakeholders in prominent roles (NRC 1996a).
The practice of risk communication is moving from trying to explain risk information to citizens toward building partnerships between plant managers and nearby residents, between companies and consumers, and between agency risk managers and the public. Although our air, water, and food are measurably cleaner and therefore less risky than they were 30 years ago, the fact that many citizens believe that they are at greater risk indicates that risk communication has a long way to go. Investments of time and resources are clearly needed.
Communicating About Risk by Comparing Different Kinds of Risk
Finding
People make informal judgments about risks every day. Some risks are familiar, even comfortable; others are unfamiliar and can be sources of considerable fear. Different people have different perceptions of the same risks. It is logical and reasonable for people to request comparisons or for Congress to incorporate mandates for risk comparisons in legislation. But some comparisons trigger resentment, as though a substantial risk were being dismissed or belittled.
Recommendation
Risk comparisons should help to convey the nature and magnitude of a particular risk estimate. Such comparisons should systematically address risks associated with different decision options, with chemically related agents, with the same agent from different exposure sources, with different kinds of agents with the same exposure pathway, or with different agents that produce similar effects.
Many kinds of risk comparisons exist. At the simple end of the spectrum are arithmetical comparisons of magnitude, such as a one-in-a-million cancer risk compared with a length of one inch in 16 miles; comparisons of chemically related agents, such as one organophosphate pesticide with another; comparisons of the same agent with different exposure sources, such as polycyclic aromatic hydrocarbons from motor vehicle exhaust and broiled meat; comparisons of different agents with the same exposure pathway, such as foods containing naturally occuring carcinogenic components as well as synthetic additives; and comparisons of different agents that produce similar effects, such as the risk of lung cancer from inhaling radioactive radon particles versus smoking a particular number of cigarettes. Toward the complex end, multiple risks are compared across a variety of dimensions, such as the hazards to the public, workers, and ecosystems of different energy-producing or Superfund cleanup technologies.
In general, risk comparisons can help people to comprehend probabilities or magnitudes. Most people, including physicians, do not easily relate low-risk probabilities or ratios, such as "one-in-a-million," to their everyday experience. One solution is to make quantitative comparisons between familiar and less familiar risks. A better solution might be to use analogies—one-in-a-million is equivalent to 30 seconds in a year, 1 inch in 16 miles, or 1 drop in 16 gallons. Another solution might be to express risk in terms of the number of persons who might be affected per year or per hypothetical 70-year lifetime. Even more difficult to communicate is the fact that a one-in-a-million risk estimate currently is not an estimate of actual risk, but a statistical upper bound on the likelihood that a risk could exist; that is, the actual risk is likely to be much lower, and it could be zero, but it is quite unlikely to be higher.
Many people perceive the reduction of risk by two or more orders of magnitude as though each "power of ten" reduction were an equivalent reduction in risk. A better way to illustrate orders of magnitude of risk reduction for linear dose-response effect is shown in Figure 3.1: the bar graph depicts better than words that a reduction in risk from one in 1,000 (10-3) to one in 10,000 (10-4) is a reduction of 90% and that a further reduction to one in 100,000 (10-5) is a reduction 10-fold less than the first reduction of 90%, i.e., a 9% reduction in the original risk. The amount of the risk reduction associated with lowered emissions and exposures is a much more meaningful concept to communicate verbally and graphically than estimates of absolute risk levels, such as 10-5.
A different proposal for communicating risk magnitude is to use time intervals, which might be better understood than numerical probability estimates. Commissioner Goldstein indicates that converting probabilities per unit of population to periods per event, such as one death expected in 3,500 years, substantially altered the perception of threat (Weinstein et al. 1996). The city of Columbus, Ohio, did an analysis estimating that one death would occur in Columbus in 204 years from an additional cancer risk at the theoretical one-in-a-million level, compared with frequencies of several deaths per day or every few days for measurable risks, such as ordinary rates of heart disease, cancer, homicide, and automobile collisions. The mayor of Columbus, Gregory Lashutka, in testimony before the Commission, stated that that analogy helps citizens to understand the magnitude of the effects that any federal or state regulation concerning the environment, transportation, labor, or education might have on the community. We recommend expressing risks both as numbers of events in an actual exposed community or on an annual basis and per million hypothetical people over a lifetime.
Using comparisons to explain the magnitude of risks will be increasingly important as advances in analytic chemistry improve our ability to detect smaller and smaller amounts of chemicals in air, water, and other media. This phenomenon of a plummeting "nondetectable" level or a "vanishing zero" poses a problem, particularly in the assessment of risks associated with human carcinogens, when no level of exposure is assumed to be without risk.
Risk comparisons can be helpful, but they should be used cautiously and tested if possible. There are proven dangers in comparing familiar and unfamiliar risks, natural and manufactured risks, and voluntary and involuntary risks, such comparisons can be perceived as minimizing a risk (NRC 1989). It is sometimes difficult to find risks that are sufficiently similar to make a comparison meaningful. In general, comparisons of unlike risks should be avoided; they are often perceived as manipulative and confusing. Comparisons of risks associated with chemically related agents, risks associated with the same agent with different exposure sources, risks related to different kinds of agents with the same exposure pathway, or comparisons of different agents that produce similar effects can improve communication. Those comparisons are better because the qualitative characteristics are similar.
Risk comparisons can either improve or hinder risk communication. Testing messages that use risk comparisons, even informally, can help to avoid miscommunication and misunderstanding.
Finding
Chemicals suspected of causing cancer are regulated by assuming that every exposure has some risk. In contrast, chemicals suspected of causing other effects, such as developmental or reproductive toxicity, are regulated by assuming that there is a safe level of exposure. That simple dichotomy is not fully supportable by current scientific evidence. Furthermore, it results in expressions of risk for cancer and for other kinds of toxicity that cannot be compared and in striking discrepancies among maximal exposures considered to have negligible risk. These discrepancies are particularly salient when the same chemical has both carcinogenic and noncarcinogenic effects.
Recommendation
To assist in comparative risk assessment and risk communication, a common metric for comparing health risks should be sought by environmental protection and public health agencies. The Commission recommends that two complementary approaches be evaluated: (1) EPA’s margin-of-exposure approach, which compares exposure levels of a chemical associated with specific carcinogenic or noncarcinogenic health effects to actual exposure levels; and (2) the margin-of-protection or safety factor approach, which is currently used everywhere to set "safe" exposure levels for chemicals causing noncarcinogenic effects, and could be applied as well to carcinogens here in the U.S. as it is in many other countries.
Having a common metric with which to compare the health risk implications of actual exposure concentrations with exposure concentrations thought to be associated with negligible risk, and exposure concentrations shown to be associated with toxicity of various kinds would have important advantages over the present situation:
• Harmonizing risk assessment methods for carcinogens and noncarcinogens might permit noncarcinogens greater emphasis than they now receive.
• Using a common metric for both carcinogens and noncarcinogens could improve risk communication. The differences between toxic exposure levels, actual exposure levels, and exposure levels considered to pose negligible risk could be compared more easily for all types of exposures and health effects. Discrepancies in the size of acceptable margins of exposure or margins of protection for different chemicals and different effects would be highlighted. There is often no margin between the clinically manifest effect levels of section 109 criteria air pollutants and actual exposures, especially in areas where levels exceed the national ambient air quality standards. In contrast, there is a 1,000- to 100,000-fold difference between the concentration that would affect 10 percent of people (ED10) and the virtually safe doses of hazardous air pollutants calculated to pose an incremental lifetime cancer risk of no more than one in 10,000 to one in one million. That discrepancy is illustrated by a comment from the Health Effects Institute: the number of excess annual deaths in Philadelphia attributed to airborne particulate matter at an exposure level equal to its national ambient air quality standard is greater than the cancer incidence attributable to any individual section 112 hazardous air pollutant ("air toxic") for the entire nation, based on an EPA study (EPA 1996a).
• The distinct but complementary roles of risk assessment and risk management would be transparent. Identifying a chemical’s relevant effects and the doses at which they occur would be a science-based activity (as it is now), and drawing conclusions about levels of exposure that might be associated with negligible risk would clearly be a risk management responsibility, requiring consensus as to the level of protection that is desired and feasible for different effects and for different situations and population groups. For example, FDA uses a larger margin of protection for a substance in food that is consumed by most of the U.S. population than OSHA does for protection of workers exposed to a solvent used in industrial processes. OSHA has to consider feasibility when it sets workplace standards; FDA considers only health impacts and deals with much larger populations with greater variation in susceptibility. It is reasonable for an agency to choose different allowable exposures for different effects, such as, specifying a more stringent standard to protect against lung cancer than reversible breathing problems.
• It would be easier to compare cancer risks to noncancer risks for making risk-management decisions. For example, it might be easier to decide whether a hazardous-waste site classified as posing an upper-bound incremental lifetime cancer risk of 1 in 10,000 should receive a higher or lower priority for cleanup than a site classified as having a noncancer hazard index of 10. The same problem arises when residual risks from hazardous air pollutant emissions at various facilities are characterized and compared (see the section on EPA and residual risks on page 109).
• It is misleading to express cancer risk in a manner that implies great precision, when cancer risk often is based on little or no more information than is available on noncancer effects. Risks from carcinogens are generally expressed in terms of upper-bound or worst-case predictions of incidence or numbers of deaths per unit of the population over 70 years. Although those predictions are not intended to be interpreted as actual or measurable cancer risks, they often are, even when the information base is restricted to observable dose-response data from rodent bioassays. In only a limited number of cases have additional mechanistic data aided in extrapolating between species and from high to low exposures.
Two potentially useful common metrics exist that we believe should be evaluated: the margin of exposure and the margin of protection.
Margin of exposure (MOE). A margin of exposure is a ratio defined in EPA’s Proposed Guidelines for Carcinogen Risk Assessment as a dose derived from a tumor bioassay, epidemiologic study, or biologic marker study, such as the exposure associated with a 10% response rate, divided by an actual or projected human exposure (EPA 1996b). Lower margins of exposure pose greater concern. For example, EPA determined that a margin of exposure of only 10 exists for neurotoxicity in workers exposed to acrylamide while engaged in sewer grouting applications. Thus, the typical sewer grouting worker may be exposed to levels of acrylamide close to the level estimated to produce neurotoxicity in humans (October 2, 1991 at 56FR49865).
Margins of exposure are specific to individual situations and do not by themselves communicate the likelihood of risk in an exposed population. Margins of exposure simply reflect the ratio between a level associated with observed toxicity in humans or animals and the actual level of exposure in a particular situation. Risk managers and stakeholders can evaluate a particular margin of exposure and decide whether it reflects an appropriate level of protection given the relevant risk management criteria. Those interpreting a margin of exposure in risk characterization and risk management can consider factors such as the slope of the dose-response relationship in the observable range, mode of action, nature and extent of the uncertainties, human variation in susceptibility to the response of concern, human sensitivity as compared with laboratory animals, and comparisons with margins of exposure of alternatives or other relevant hazards.
A margin of exposure cannot be equated with safety. Further discussion with stakeholders is necessary, as part of the risk communication and risk management processes, to consider likely levels of risk or safety and influence priorities for action. Typically, margins of exposure are much smaller in the workplace than in community settings.
Margin of protection (MOP). The margin of protection or safety factor method is used to derive estimates of acceptable daily intakes (ADI), reference doses (RfD), or reference concentrations (RfC) for noncancer effects. Those values represent chemical exposure concentrations that would be associated with negligible risk. A no-observed-adverse-effect level (NOAEL),1 a lowest-observed-adverse-effect level (LOAEL), a benchmark dose, or some other level derived experimentally is divided by factors thought to account for variability and uncertainty to obtain the ADI, RfD, or RfC. Typically, a margin of protection is 1,000 when three safety factors of 10 each are used and multiplied together. If the resulting ADI, RfD, or RfC in food, water, soil, or air is not exceeded, adverse health effects are so unlikely that exposures are considered "safe." These negligible risk levels are inversely related to a chemical’s toxic potency. This method is used in Canada, Europe, and many other countries for carcinogens as well as for chemicals causing other types of adverse effects. It is used in the U.S. primarily for noncarcinogens.
Several factors affect the interpretation of margins of exposure and margins of protection:
• Severity of the health effects being compared. Margins based on nasal irritation, for example, are not easily compared to those based on lung cancer or reproductive toxicity.
• The exposure concentration associated with observable health effects that is chosen as the basis for deriving the margins (the "point of departure"). It may be difficult to compare a margin of protection derived from a NOAEL to a margin of protection based on a benchmark dose. Not only are NOAELs and benchmark doeses derived differently, NOAELs are associated with some level of safety because no effects are being observed, while benchmark doeses and ED10s relate to observed effects. An additional safety factor of 10 is routinely used if the LOAEL is used instead of the NOAEL.
• Dose-response relationships. Dose-response relationships influence the interpretation of margins of exposure and margins of protection; as mechanistic knowledge increases, more sophisticated interpretation is facilitated. Margins of exposure and margins of protection are the same only when dose-response relationships are linear. If a contaminant has a nonlinear dose-response relationship, a margin of exposure of 10 might confer a margin of protection of, perhaps, 100 or more. It is important to try to evaluate the nature of the dose-response relationship when discussing the acceptability of a particular margin of exposure; doing so is difficult when extrapolating way below the range of exposures that produce observable effects in rodent bioassays or epidemiology studies.
• Exposure duration and latency. Differences in duration of exposure and in latency until effects become manifest complicate comparisons of dose-response relationships.
In the end, both margins of protection and margins of exposure can provide useful starting points for a broad range of stakeholders to query experts and regulators about risks of various adverse effects from individual agents or multiple agents and help build a consensus about risk management options and priorities.
Comparative Risk Analysis for Risk Management Priority Setting
Priority setting is necessary when money, time, and staff are limited. The Carnegie Commission on Science, Technology, and Government, the National Academy of Public Administration, many members of Congress, and Supreme Court Justice Stephen Breyer’s book, Breaking the Vicious Circle, have recommended comparative risk assessment approaches for priority setting.2 The process of comparing various types of risks includes problem identification, data collection and analysis, and risk ranking of environmental problems to developing an action plan and implementing new strategies for risk management and reducing risk. This process is a generic version of our Risk Management Framework for specific problems (Section 2).
Most comparative risk projects for priority setting have been conducted by state, local and tribal governments, typically led by one or more environmental protection, natural resource, or health agency. EPA’s Regional and State Planning Division has provided grants, training, and problem-solving assistance. Our recommendation here is directed primarily at federal agencies.
Risk-Based Priorities and Resource Allocation
Finding
Federal regulatory agencies are confronted with many problems and issues related to health and environmental protection, but have limited time and resources for action. The risks associated with the problems and the resources available to act on them are often misaligned. EPA, state, local, and tribal comparative risk projects have been useful in identifying such mismatches and in refining the comparative risk process to better manage risks.
Recommendation
Federal regulatory agencies should try a comparative risk analysis approach on an experimental or demonstration basis to seek consensus on setting priorities for risk management of environment, health, and safety hazards. The priorities, reflecting diverse stakeholder values and opinions, should influence agency resource allocation decisions and be fully discussed in Congressional oversight hearings. The Commission’s Risk Management Framework—with its emphasis on context, total exposure, and early stakeholder engagement—should be useful.
EPA undertook some of the earliest efforts to use comparative risk assessment to rank environmental risks and set priorities for agency efforts. In 1987, EPA staff prepared a report, Unfinished Business: A Comparative Assessment of Environmental Problems (U.S. EPA 1987b), that identified risks receiving in their view inadequate attention from the agency. An important conclusion of the report was that the EPA’s program priorities tended to reflect the public’s perception of risks, rather than the most serious risks as judged by EPA scientists and staff. The Science Advisory Board reviewed that report and issued Reducing Risk: Setting Priorities and Strategies for Environmental Protection (EPA 1990). The Science Advisory Board emphasized the subjective nature of rankings and called for broad public participation in ranking environmental risks so that risk reduction policies based on imperfect and evolving scientific understanding and subjective public opinion might be more widely supported.
In 1995, EPA and Congress asked the Science Advisory Board to undertake an integrated ranking project as a follow-up to the risk rankings in Reducing Risk. The difference in those efforts and the EPA-funded state, local, and tribal comparative risk projects is the explicit incorporation of public values and perceptions of risk, a process of diverse stakeholder involvement, and inclusion of elected officials’ representatives in the state, local, and tribal activities. As a result, it appears that the state, local, and tribal comparative risk assessment projects may have been more successful in influencing agency priorities and resource allocations. Unfortunately, legislative proposals requiring federal agencies to perform comparative risk analysis for priority setting and make appropriate adjustments in budget requests have had the same weakness. These proposals have not included stakeholders in the priority setting process and thus miss the opportunity to build public support for changes in funding allocations.
There is wide disagreement about the efficacy of using comparative risk analysis for setting priorities. In comments to the Commission, the Consumer Product Safety Commission and the Natural Resources Defense Council emphasized the limitation of comparative risk analysis and the need to provide agencies with some discretion in setting priorities. In contrast, the Chemical Manufacturers Association concluded that there has been sufficient experience with the comparative risk process to start using it to guide agency priorities. Of course, EPA priorities are often dominated by statutory and court-ordered deadlines, and Department of Energy priorities are influenced by Congressional appropriations and triparty agreements with state and federal regulators.
Comparative risk analysis for priority setting brings together science and public values by making clear what is known and what is not known about the environmental challenges we face. The comparative risk process includes organizing teams of agency and nonagency stakeholders, such as representatives of business and environmental groups; making a comprehensive list of environmental problems; assembling the available good information about the sources of the problems and the risks that they pose to human health, ecosystems, and quality of life; ranking the problems in order of the group’s view of the risks posed; and using the rankings to guide strategic planning and budgeting. Methods for ranking the risks of identified problems have included voting by participants; formulas that rely more heavily on quantitative data; matrix-based discussions categorizing risks as high, medium, or low; decision-seeking consensus; and bargaining or tradeoffs among stakeholders. Comparative risk analysis for risk management tracks the six steps of the Commission’s Risk Management Framework (see Section 2) and can mobilize and energize stakeholder participation.
Each federal agency will need to adapt the fundamental elements of the comparative risk ranking approach to its mission, statutory mandates, and current and emerging responsibilities. At the federal level, agencies can substitute staff of authorizing committees of Congress for state and local representatives and can identify agency staff and affected stakeholders, including representatives of national and local environmental organizations, as participants on the basis of programs and projects of specific agencies. Depending on the agency, it will be important to include representatives from state, local, and other federal agencies with relevant programmatic responsibilities or interests. State and local participation will be especially important as roles and obligations change under the Unfunded Mandates Act of 1995, which places limits on the capacity of the federal government to implement new programs that will cost state and local governments over $50 million in any year, beginning with 1996.
Benefits other than priority setting often justify putting time and effort into the comparative risk analysis process (Minard and Jones 1993). Most comparative risk projects produce a catalog of the major environmental problems facing a state or locality, which can be a valuable resource for the public and for risk managers. Participants in a comparative risk project learn about a range of problems that might not be part of their daily interests or responsibilities. The comparative risk process improves understanding of competing priorities, provides an appreciation of the complexity of decision-making, and can stimulate new insights into solutions. As a result of increased communication among institutions and interest groups, new avenues of cooperation might be established. Adversarial relationships among interest groups and jurisdictional conflicts among agencies might not disappear, and could even be intensified, but comparative risk projects have revealed unexpected agreement among parties and enhanced understanding of differences in perspectives and values in some cases. Most important, experience has shown that the process itself can help to build coalitions that favor priority setting and shifting resources to the identified priorities. Broader public support for a common agenda might allow agencies, state legislatures, and Congress to move money and staff into priority problems with less litigation and less controversy. For example, Charles Kleeburg, director of the Seattle Drainage and Wastewater Utility, explained to the Commission that the city’s success in forging consensus on ten priority problems that were acted on by the city government was a direct result of the influence and effectiveness of the comparative risk analysis process in the Mayor’s Environmental Priorities Project. In contrast, testimony from EPA indicated that a great deal of controversy is generated when it tries to address problems that it knows are real but has not been directed by Congress to address.
Numerous challenges and limitations impede the usefulness of the process, as pointed out by Patricia Buffler and Carl Craner in testimony before the Commission about the California Comparative Risk Project. There is no guarantee that the process will produce consensus among stakeholders, agencies, and funding authorities. Resolving inconsistent data across problems, forcing all risks to conform to a common measurement, and integrating problems into a single list are important methodologic and political challenges. The degree of uncertainty varies across problems, making comparisons even more difficult. The process might not adequately account for environmental equity, emerging issues, and effects across jurisdictional boundaries. Those problems can result in some groups’ objecting strongly to the rankings, in loss of opportunities for preventing future risks, and in the neglect of risks imported from or exported to other geographic areas. Lack of sufficient resources and time constraints can limit data collection, diminish the quality of data analysis, and hinder development of risk management strategies and recommendations. For example, changing the scope and criteria after a comparative risk project was well underway created conflict in California. For federal agencies, there may be additional problems of proposing changes to statutory mandates when priorities for resources change and in taking action in the absence of clear or explicit statutory direction.
The comparative risk process emerging from the state, local, and tribal projects supported by EPA’s Regional and State Planning Division constitutes a worthy starting point for federal agencies to use in ranking priorities and making resource allocation decisions. For example, the risk-based process being introduced by the Department of Energy’s Environmental Management Program at the nation’s nuclear waste sites is intended to test how well identification, analysis, and comparison of risks and remedies can be translated into budget decisions for those extremely complex sites. The Commission encourages federal regulatory agencies to continue to develop comparative risk approaches for priority setting.
Strategies for managing risk have centered traditionally on command-and-control regulatory programs and specific bright lines delineating between acceptable and unacceptable levels of cancer risk. Command-and-control regulatory programs have led to improved air quality, water quality, and better handling of solid and hazardous waste; however, additional ways of improving human health and environmental protection are becoming available and credible. Futhermore, use of risk estimates with bright lines, such as one-in-a-million, and single point estimates in general, provide a misleading implication of knowledge and certainty. As a result, reliance on command-and-control regulatory programs and use of strict bright lines in risk estimates to distingush between safe and unsafe are inconsistent with the Commission’s Risk Management Framework and with the inclusion of cost, stakeholder values, and other considerations in decision-making.
Risk Management Options: Alternatives to Command and Control
Finding
Many risks to human health and the environment have been reduced over the last 25 years, primarily through command-and-control regulations of existing and new sources of emissions and through testing requirements for newly developed chemical products. In some cases, those practices have led to very high compliance costs and increased litigation, causing delays in human health and environmental protection. Performance goals for environmental protection can increase the flexibility risk managers and stakeholders have to pursue the most effective and efficient solutions. Having a range of solutions can facilitate decision-making when options to reduce or eliminate risk are identified. Implementation of the Government Performance and Review Act may provide a means of judging whether alternatives achieve environmental goals.
Recommendation
Risk managers and stakeholders should aggressively seek alternatives to command-and-control regulation to improve the efficiency and effectiveness of health and environmental protection and to reduce compliance and litigation costs. A sense of experimentation and a commitment to evaluation should be key elements of identifying and implementing alternatives. A safety net of command-and-control regulations should be maintained, however, to avoid reducing current levels of protection.
In the last quarter century, the United States has made extraordinary progress in human health and environmental protection as a result of substantial investments by governments and by industry and through effective public and political advocacy. We now have a system of regulatory controls, enforcement, and sanctions that has established a floor for environmental protection.
Command-and-control regulations set environmental standards that are enforced through penalties for violating permits. While this system has resulted in significant reductions in pollution, we appear to have reached a point of diminishing returns in many situations, in that each incremental improvement in community health and environmental risk reduction comes only with a large increase in control costs. In those cases, the benefits of additional regulation may be slight because so much risk reduction has already been achieved. In other cases, the cost of risk reduction is aggravated by the rigidity of the underlying command-and-control regulatory system. Rule-makings and permitting processes become de facto design standards requiring the use of specific technologies for pollution control. There may not be adequate flexibility for tailoring remedies to reflect the circumstances of individual sources and locations, including the relative advantages that different companies might have in choosing risk reduction options. Ironically, some companies, especially small businesses, may prefer design or technology standards because resources for research and innovation are limited.
While government must set environmental and worker protection standards, there are important economic and environmental benefits in allowing risk managers and stakeholders greater flexibility in determining how to meet those standards. Greater flexibility must be coupled with agency monitoring and enforcement, however, to ensure that the expected level of environmental protection is being achieved. In addition, the fairness of who benefits and who pays the cost under alternative environmental protection approaches should be compared with the equity of who benefits and who pays the cost under the status quo. Environmental accounting, industrial ecology and life-cycle analysis, and environmental audits are emerging analytic tools that can assist in understanding the interaction between economic activity and environmental protection efforts.
For progress to continue in protecting human health and the environment, we must look beyond command-and-control regulatory programs for managing risk. The call for new tools to manage risk was particularly strong in presentations to the Commission outside Washington, DC. Walter Buckholz from Exxon Chemical Company in Houston, Texas, testified that command-and-control regulations were not controlling some contaminants well and called for the use of performance standards as more cost-effective. Jonathan Howes, Secretary of the North Carolina Department of Environment, Health, and Natural Resources, reported that a National Academy of Public Administration committee concluded that many businesses have chosen to exceed environmental standards if they can use their own strategies to achieve established pollution reduction targets. In other testimony, representatives of federal agencies emphasized their commitment and cited their projects aimed at finding additional options for achieving environmental and worker protection.
Education and information, incentives, monitoring, research, and surveillance are methods that may be helpful elements in risk management options. Right-to-know requirements are measures that rely on information and education for achieving risk reduction and environmental protection. Market-based incentives, subsidies, alternative compliance, and consensus, mediation, and dialogue projects are incentives that can be used when and where they make sense in responding to additional risk reduction opportunities. Research, monitoring, and surveillance are important ways of increasing knowledge about the problem, tracking the change that may be occurring, and observing health effects. When alternative tools for risk management are used, it is important to evaluate them for reliability in meeting or exceeding environmental goals, feasibility of implementation, and general effectiveness and efficiency.
The National Environmental Partnership Program System, started by EPA and the states in 1995, may provide a way to measure the success of the increasing use of alternatives to command-and-control regulations. The system is designed to give states greater flexibility to focus resources on the most serious environmental problems while enhancing accountability to the public and taxpayers for the improvement of environmental conditions and trends. Six states signed pilot Environmental Performance Agreements in 1996 and approximately 30 states are negotiating agreements in 1997. The development of core program performance measures will be an important tool for judging the functioning of state programs. Similarly, the Government Performance and Results Act of 1993 requires EPA to establish performance goals and accountability in carrying out the environmental statutes adopted by Congress.
Tools for Understanding Consequences of Economic Activity and Environmental Protection
Environmental Accounting. There is a movement from traditional accounting systems toward "environmental accounting" for both national and business accounts. In June 1995, EPA published An Introduction to Environmental Accounting as a Business Management Tool: Key Concepts and Terms (EPA 1995c). Many private-sector and private-public partnership forums are addressing this topic.
In traditional accounting of revenue, expenses, and net income of businesses, energy costs are lumped into overhead. Effects on and uses of natural resources—such as air, rivers, soils, and other environmental components—are neglected altogether. The challenge is to incorporate all costs involved in design, production, use, disposal, and reuse so as to arrive at a life-cycle analysis of a product or process. Assigning values to various environmental assets used and to real or potential environmental effects that have varied probabilities is problematic, however. Those assigned values may well drive the results of the analysis. Nevertheless, the process of environmental accounting can link environmental costs with activities and products and provide information that results in win-win opportunities to increase operational efficiency, improve worker safety, enhance product quality, and meet environmental protection goals. Unfortunately, bankers and investment advisers have been slow to encourage investments in these cost-saving initiatives. The President’s Council on Sustainable Development (1996) recommended that national business associations provide technical assistance to companies interested in identifying environmental management costs and innovative ways to increase profits by reducing energy and materials use while better protecting public health and the environment. We agree.
Industrial Ecology and Life-Cycle Analysis. Proponents of industrial ecology envision a closed-loop system in which no resources are depleted; that is, all materials are perpetually reused, and no waste is produced or discarded. The loops might be closed within a factory, among industries in a region, or within national or global economies. Industrial ecology would integrate the producing and consuming segments of an economy to optimize the use and recycling of industrial materials and products. "Benign by design" chemistry, in which synthetic chemistry is designed to use and generate fewer hazardous substances, is a step toward achieving a closed-loop system. Quad Graphics, a Wisconsin based printing business, and Stonyfield Farm, a yogurt producer located in New Hampshire, are trying to establish eco-industrial parks where companies with compatible production processes can use resources more efficiently and reduce waste. Life-cycle analysis is important to the implementation of industrial ecology, because it provides information that can be used to understand the consequences of choices among materials, product designs, and process designs and to understand the fate of products when they are finally discarded by consumers. Life-cycle analyses have been mandated in the European Community. Nevertheless, industry representatives emphasize that life-cycle analysis relies on many assumptions and needs further research and development before it can be a reliable and cost-effecive tool.
Environmental Audits. Audits by industry and by third parties are another tool for influencing corporate compliance with command-and-control regulations, especially when penalties are eased for self-disclosed violations. Audits also allow emitters to highlight voluntary reduction of pollutant emissions to the air, water, and land. Environmental audits have become controversial with the passage of recent state legislation providing blanket protection from penalties for self-disclosed violations.
Using Education To Manage Risk
Right-To-Know Requirements. EPA’s Toxic Release Inventory (TRI), mandated by the Superfund Amendments of 1986, and California Proposition 65 require the disclosure of information about chemical releases to the environment and labeling of chemicals in products, respectively. Those right-to-know laws educate the public and rely on attitudes toward toxicants to encourage industry to reduce or eliminate their use or release.
The TRI is an annual measure of chemicals used, manufactured, transported, or released into the environment by facilities in communities throughout the United States. The 1995 TRI included approximately 600 chemicals. The TRI allows EPA, the states, industry, and the public to gauge industry’s progress in reducing chemical use and waste generation. Reported toxic releases have declined by nearly 44 percent between 1988, the baseline year, and 1994, the last year for which data are available. Several companies have reported success in achieving a voluntary 75% reduction in toxic air emissions since 1988.
In the case of Proposition 65, the requirement to warn people about exposures to chemicals known to cause cancer, birth defects, or other reproductive harm has been an incentive to businesses to eliminate such chemicals or reduce exposures and associated risks below the bright lines for cancer and reproductive risks. Rather than relying on command and control, Proposition 65 uses disclosure of information and labeling requirements as risk management tools. Proposition 65 places the burden of proof of safety on manufacturers rather than on government agencies, requiring businesses to present a risk-based analysis to avoid having to label their products and substances as cancer-causing or reproductive toxicants. David Roe of the Environmental Defense Fund told the Commission that a key decision by the state environmental protection agency was to put the bright line for cancer risk at 10-5, rather than 10-6 as proposed by environmentalists or 10-4 as proposed by business. Proposition 65 has been criticized by some in industry as using questionable science to produce faulty warnings. Others have reported that the responsible agency welcomes good risk-based analyses. The California Environmental Protection Agency’s Risk Assessment Advisory Committee has called for the use of much more scientific information in evaluating cancer and reproductive hazards under Proposition 65 (RAAC 1996).
Using Incentives To Manage Risk
Market-Based Incentives. Market-based incentives rely on economic motivators to encourage environmental protection and cost-effectiveness. A prominent example of market-based incentives to achieve environmental protection is the use of tradable sulfur dioxide emission allowances to reduce acid rain. This program, mandated under the 1990 amendments to the Clean Air Act, permits electric utilities to reduce their emissions of sulfur dioxide, the precursor to acid precipitation, below allowable levels and sell the unused emission allowances to companies whose cost of compliance is substantially greater. The program caps aggregate sulfur dioxide emissions well below historical levels, while allowing emission reductions to be achieved more cost-effectively than by requiring every company to install the most expensive sulfur dioxide control technology. The cost of a ton of sulfur dioxide emission allowances has fallen well below projected costs, presumably reflecting technological advances. Similar programs are being developed to reduce regional nitrogen oxide emissions. The use of caps and tradable pollution allowances may not work well in some cases, such as toxic air pollutants, where sources create highly localized risks.
Other Incentives. In addition to the use of direct economic incentive policies, other positive incentives are available to encourage pollution prevention, some of which EPA has implemented. For example, some pesticides that require approval by EPA before they can be distributed, used, or sold can be given priority for approval if they are deemed safer for human health and the environment and thereby reach the marketplace faster than other pesticides. Safer products could receive more favorable treatment if labelling regulations are implemented, such as authority to use a special label, to give them greater prominence in the market. To encourage pollution prevention by manufacturing facilities, businesses could be given tax incentives to replace old facilities with new, cleaner processes that do not generate waste and pollution.
Another example pertaining to Title V permits under the Clean Air Act is EPA’s Pollution Prevention in Permitting Pilot Project (P4 Project) with Intel Corporation, the Oregon Department of Environmental Quality, and the Northwest Pollution Prevention Resource Center as private, public, and nonprofit sector partners. The pilot is now being extended to five other companies in EPA regions 1, 4, 6, 9, and 10. The aim is to reduce production of air emissions, rather than control their release in ways that generate solid waste or waste water.
Subsidies. Subsidy programs can be powerful tools for reducing pollution while selectively encouraging economic activity. For example, agricultural land retirement programs have prevented excessive soil erosion and damage to water bodies and wildlife habitat. Government purchasing practices can also encourage the development of markets for products that are environmentally more sound. Care is needed to avoid excessive acquisition costs for products with small markets and to avoid buying products with one attractive attribute but other unfavorable characteristics, however. Some subsidy programs have had detrimental effects on the environment.
Alternative Compliance. Alternative compliance provides greater flexibility to industry by allowing choices in achieving emission or risk reduction specifications. Designed to achieve higher levels of environmental protection at lower cost and to foster integration of local concerns in environmental risk management decisions, this option can result in substantial savings for industry, communities, or any regulated entity that participates. For example, EPA’s Project XL allows six companies (Intel Corporation, Anheuser Busch Companies, HADCO Corporation, Merck & Co., Inc, AT&T Microelectronics, and 3M Corporation) and two government agencies (California’s South Coast Air Quality Management District and the Minnesota Pollution Control Agency) to experiment with different strategies for improving environmental protection. As of 1996, some projects were making faster progress than others. Intel negotiated the first agreement, but it was criticized by a number of organizations not included in the process as not achieving greater environmental protection. We recommend use of the Commission’s Framework to address multimedia problems and to incorporate stakeholders. As noted in Volume 1, some of the difficulties with Project XL have arisen from the stakeholder involvement process used by EPA. Government also can provide greater compliance flexibility for those attempting to use innovative pollution reduction and control technologies. Use of the concept of a bubble to encompass a facility or geographic area and seek the best way to reduce a pollutant or pollutants within the bubble has also provided flexibility in compliance.
Consensus, Mediation, and Dialogue Projects. Negotiated rule-making and dialogue projects, such as EPA’s Common Sense Initiative, offer opportunities for stakeholders to design new standards and solutions that protect human health and the environment more reliably and with greater cost-effectiveness and public acceptance. The Office of Management and Budget (OMB) ruled in 1996 that the cap on numbers of chartered advisory committees no longer applied to negotiated rule-making processes. With the Common Sense Initiative, begun in 1994, EPA has convened consensus-oriented teams of stakeholders to look for opportunities to turn complicated and inconsistent environmental regulations for six major industries—automobile manufacturing, computers and electronics, iron and steel, metal finishing, petroleum refining, and printing—into comprehensive sector-specific strategies for environmental protection. In 1996, the state of Michigan withdrew from the automobile manufacturing initiative, saying EPA was not providing sufficient flexibility. Several industrial sectors have launched their own initiatives, such as Responsible Care by the Chemical Manufacturers Association.
Using Monitoring, Research, and Surveillance To Manage Risk
Monitoring. Monitoring emissions and ambient conditions has been a long-term component of command-and-control regulatory programs and other efforts to characterize the status of the environment. Monitoring can also be a useful risk management tool when a community is skeptical or suspicious of the effectiveness of risk management actions, making alternatives to traditional command-and-control programs more acceptable.
Research. Research can be an important management option when lack of knowledge about the source of a problem or its impacts make a course of action unclear (see Value of Obtaining Additional Information on page 91). EPA’s cooperative effort with scientists to identify and design appropriate research projects on hormonally active agents, as directed by Congress, is an example of research to inform future risk management decision-making.
Surveillance. Health surveillance is an under used technique for observing effects of pollution on human health. While the incidence of cancer from environmental toxicants may be so small as to be unobservable, surveillance of other health effects such as asthma or heart attack death rates may lead to identification of problems needing additional risk reduction.
Bright Lines for Risk Management
Finding
There is much controversey about bright lines, "cut points," or decision criteria used in setting and evaluating compliance with standards, tolerances, cleanup levels, or other regulatory actions. Risk managers sometimes rely on clearly demarcated bright lines, defining boundaries between unacceptable and negligible upper limits on cancer risk, to guide their decisions. Congress has occasionally sought to include specified bright lines in legislation. A strict bright-line approach to decision-making cannot explicitly reflect uncertainty about risks, population variation in susceptibility, community preferences and values, or economic considerations, however, all of which are required by the Commission’s Risk Management Framework.
Recommendation
Bright lines can be helpful as guideposts in screening risk assessments (see Tiered Scheme for Determining and Managing Residual Risks on page 109). Bright lines or ranges of bright lines tied to specific exposure or contaminant concentrations can be used for compliance. In addition to bright lines intended to protect the general population, bright lines can be used by regulators to protect especially susceptible subpopulations, such as young children, pregnant women, or adults with lung disease. Because of the need for flexibility, Congress should leave the establishment of specific bright lines or ranges of bright lines to the regulatory agencies.
A "bright line" is a single numerical value between unacceptable and negligible magnitudes of exposure or of risk. Bright lines are chosen to provide pragmatic definitions of "safe" and "unsafe" for those making risk management decisions and especially for those implementing, monitoring, and enforcing those decisions. An example of a measurable bright line is a tolerance level of 20 parts per billion for the carcinogen aflatoxin, which is produced by a fungus that grows on peanuts and corn. Peanut and corn crops are tested for aflatoxin contamination and if the level is greater than 20 parts per billion, they cannot be sold or consumed. If the level is less than 20 parts per billion, the crops are considered fit for human consumption.
Bright lines tied to upper-bound estimates of excess lifetime cancer risk which cannot be measured, are limited in their usefulness. Consider, for example, an excess lifetime cancer risk of 10-5: if a risk assessment predicts that more than one case of cancer is likely to occur as a result of exposure to a substance in a population of 100,000 people exposed to it, that risk may be judged unacceptable and protective action be required; a predicted risk of less than 10-5 may be considered negligible and require no protective action. Risk-based bright lines must be converted to regulatory standards expressed as measurable exposure, emission, or contaminant concentrations for implementation and compliance. Regulated parties are expected to demonstrate compliance that estimated risks are below the bright line by showing that measured or estimated exposure concentrations are below the regulatory standard to operate a manufacturing facility, introduce a new product to the market, or sell foods with low concentrations of contaminants.
Risk-based bright lines are generally compared to single point estimates of risk to judge safety; Science and Judgment in Risk Assessment characterizes such bright lines and point estimates of risk as "magic numbers" whose use is inconsistent with knowledge about the variability and uncertainty inherent in estimates of risk (NRC 1994a). Strict use of bright lines is also inconsistent with the Risk Management Framework and with the inclusion of cost and other considerations in decision-making.
Several major problems arise in using of bright lines tied to risk levels. The all-or-nothing nature of use of a bright line could be misunderstood and construed to imply that there is an exact boundary between safety and risk, even though risk-based bright lines are burdened by all the uncertainty, variability, and assumptions inherent in cancer risk estimation. Risk assessments themselves can be manipulated so that their results emerge above or below the bright line according to a risk manager’s particular policy preferences. Bright lines have the potential to be applied inflexibly, leading to decisions that do not reflect the unique characteristics of particular populations. As Roger Pryor, executive director of the Missouri Coalition for the Environment, testified before the Commission, health considerations, cost, and cultural differences all play a role in risk management decisions.
Ranges of bright lines have sometimes been adopted by regulatory policy. For example, under Superfund, a pair of bright lines has been used to define a potentially acceptable risk range for carcinogens. A contaminated site is considered to pose a negligible risk if a risk assessment of the site produces an upper-bound lifetime incremental cancer risk estimate not exceeding 10-6. The site is considered to pose an unacceptable risk, requiring remediation, if the risk estimate is 10-4 or higher. Between 10-6 and 10-4, remedial actions, if any, are determined case by case.
Congress has included bright-line risk provisions in several legislative bills in recent years. In the 1990 Clean Air Act amendments, Congress specified a quantified risk level, when it mandated the development of a strategy for evaluating residual risks after maximum available control technology (MACT) implementation based on an incremental lifetime cancer risk of 10-6 (see Tiered Scheme for Determining and Managing Residual Risks on page 109).
In addition to ranges of bright lines, multiple bright lines should be considered. For example, Accounting for Differences in Susceptibility on page 71 discusses the need to consider sensitive subpopulations in risk assessments. The results of such risk assessments might be expressed in terms of an estimated risk for the general population and a different estimated risk for a sensitive subpopulation. Those risk estimates could be used to establish a bright line for the general population and a different bright line for the sensitive subpopulation. Decisions about appropriate levels of risk reduction could then be made with the benefit of the knowledge of those differences. EPA’s deputy administrator, Fred Hansen, noted in his testimony before the Commission that getting away from single bright lines would be consistent with incorporating environmental justice considerations into risk management.
As noted above, regulatory standards expressed as contaminant concentrations are much easier to implement than risk levels. Risk managers are accustomed to the clear guidance provided by regulatory standards, expressed as emissions or exposure concentrations, for implementing and determining compliance. Measurable contaminant concentrations—such as permissible exposure limits or threshold limit values in the workplace, action levels for food contaminants like aflatoxin on peanuts or mercury in swordfish, and national ambient air quality standards for carbon monoxide or ozone in air—are intended to provide assurance that risks will be negligible so long as contaminant exposure concentrations are below the standards. Small quantitative differences from those standards, whether above or below, can make a big difference in whether protective actions are taken, although some discretion may exist, analagous to failing to arrest a driver traveling at 57 miles per hour in a 55 mile-per-hour zone. Such regulatory standards reflect some judgment about what exposure constitutes negligible risk (or, in other cases, technologic feasibility).
Judicial Review of Regulatory Decisions to Manage Risk
Issues of judicial review that were raised by the 104th Congress—in the context of "regulatory reform" legislation and amendments to the Administrative Procedure Act (APA)—were carefully analyzed, vigorously debated, and are likely to be revisited by Congress. Those issues focused debate on the proper role of judicial review of agency action in the regulatory process.
Conceptually, judicial review is the check by the judicial branch on agency activity at an appropriate stage of the administrative process, and in an appropriate manner and degree. Agencies are authorized to act and promulgate regulations under enabling statutes passed by Congress. The various enabling statutes also grant the right and limit the extent of review of agency action by courts. Both agency action and judicial review of regulatory rulemaking are governed by the provisions of the APA. A party that is affected by agency action can seek judicial review of that action in court when all other administrative remedies and appeals have been exhausted. A preliminary, procedural, or intermediate action by an agency that is not directly reviewable by a court is subject to review under the APA only upon final agency action, however, so that it will not interrupt the regulatory process prematurely.
A reviewing court adjudicates procedural issues, interpretations of constitutional and statutory provisions, and determinations of the meaning or applicability of the terms of agency action. It can compel agency action unlawfully witheld and hold agency action to be unlawful if the court finds it to be arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with the law or in observance of procedure required by law. Moreover, when a reviewing court considers the record developed through formal agency hearings (formal hearings are required under certain enabling statutes), or when "substantial evidence" is otherwise required by statute, the court can hold agency action unlawful if that action is not supported by substantial evidence.
The Commission carefully considered these issues and the effect of each on the regulatory rule making process. In short, as discussed below, we submit that legislative initiatives should not provide for premature interruption of the administrative process nor expand the nature and extent of judicial review in ways that will require courts to devote substantial time and resources to the oversight of agency compliance with detailed procedural requirements or the resolution of complex scientific issues, but should consider the use of alternatives that assure rational and cost-effective regulatory action.
Premature Interruption of the Administrative Process
Finding
Interlocutory, or intermediate, appeals of discrete issues prematurely interrupt the administrative process.
Recommendation
Judicial review should be available only after agency action is complete and all administrative remedies have been exhausted.
Historically, provisions for judicial review under the APA grant review of the rulemaking record for "final agency action." This practice limits parties from interrupting the administrative process by seeking judicial review of discrete issues until all other administrative remedies have been pursued and exhausted. The APA provides a procedural safeguard that not only ensures the establishment of a rulemaking record but preserves that record. Thus, in the administrative context, an agency has the opportunity to apply its expertise, exercise its informed discretion, and create a more complete record, so that there is a full record upon which a court can adjudicate if judicial review is invoked.
Administrative procedure and practice require a party to challenge issues within the internal agency’s deliberative process. Issues raised in an administrative proceeding allow an agency to monitor and correct its mistakes, omissions, or oversights. Without resorting to costly lawsuits and court-imposed remedies, the administrative review process provides agencies with an opportunity to research and develop more fully a record that identifies issues considered as part of the rulemaking process.
Proponents of some legislative initiatives maintained that they preserved the APA’s premise that only final agency action is reviewable, but there were suggestions and debate as to what was considered to be final agency action. In various drafts of proposed legislation in the 104th Congress, a number of initial and intermediate agency determinations in the rule-making process were deemed final agency action. That would have created an opportunity to leap immediately out of the administrative context, where issues could be developed fully, and into the judicial arena, under the guise of final agency action. Interested parties could prematurely, and in piece-meal fashion, seek judicial review of discrete issues during the drafting and implementing of agency regulations, thereby delaying and hamstringing the regulatory process. Allowing premature interruption of the administrative process would impede development of the rulemaking record. As a consequence, judicial review would proceed on an incomplete record and issues would be adjudicated without a full and fair development of the underlying data and benefit of scientific analysis.
The Nature and Extent of Judicial Review
Finding
Legislation was proposed in the 104th Congress that specifies detailed requirements for making risk management decisions and for regulatory rule making. Those provisions would have governed the content of risk assessments and benefit-cost analyses, the procedures for preparing the analyses, and the regulatory decisions based on the analyses. Under accepted administrative law principles, all such requirements would be judicially reviewable, potentially leading to increased and more complex litigation over agency decision-making.
Recommendation
To be consistent with the Commission’s Risk Management Framework, provisions that would make substantive risk assessments and benefit-cost analyses and their underlying factual support subject to expanded judicial review, as well as prescriptive and detailed procedures for conducting those assessments and analyses, should not be legislatively grafted across the board onto existing enabling statutes.
Although issues of scientific method and factual support for agency Findings are currently subject to judicial review, courts typically have confined themselves to broad oversight in deference to agency scientific decision-making. Such deference allows agencies substantial flexibility in drawing upon their specialized expertise while ensuring that they follow accepted procedures and standards. Indeed, one of the primary reasons administrative agencies were created was to bring specialized expertise to bear on complex issues.
Some proposed legislative initiatives would have changed the nature and extent of judicial review of agency decisions. A legislative mandate to agencies to follow intricate, detailed procedures in developing benefit-cost analyses and risk assessments, combined with a change in the standard of judicial review of agency decision-making from the "arbitrary and capricious" standard to the less deferential "substantial evidence" standard (see below), inevitably would involve courts in an investigation of much more than whether a rational basis exists to support an agency rule. In examining agency compliance with detailed substantive and procedural requirements under a broadened standard of review, courts would be likely to delve far more deeply into the many complex scientific issues affecting a rule. That change could create not only increased opportunities for litigation, but more complicated, more expensive litigation.
Some proposals also would have legislatively established criteria ("decisional criteria") that would be used to evaluate the validity of a rule and would supplement all existing enabling statutes.
Consequently, the Findings of cost and risk evaluations, conflicts with regard to scientific data, postulates representing the most reasonable inferences from supporting toxicologic and epidemiologic data, and determinations of whether an agency sufficiently used the appropriate information in its analysis would become part of the agency record subject to judicial scrutiny.
The Commission’s Risk Management Framework (see Section 2) emphasizes the importance of evaluating information about risks, costs, benefits, and stakeholder values in regulatory decision-making. Legislatively specifying additional decisional criteria, however, such as requiring that incremental benefits exceed incremental costs, would limit agency flexibility in the rulemaking process and could expand the scope of judicial review. For example, section 109 of the Clean Air Act has been interpreted by the courts to preclude reliance on benefit-cost analyses. When EPA sets national ambient air quality standards under section 109, those standards must be based on evidence of adverse health effects and must be protective of even the most susceptible population subgroups. In contrast, the recently reauthorized Safe Drinking Water Act is an example of a statute that emphasizes the importance of considering costs and benefits in addition to feasibility and human health risks when setting drinking water standards.
We acknowledge that a compromise position was advanced that called for a program-by-program legislative approach tailoring decisional criteria to fit individual statutory schemes instead of insisting on an overarching mandate. We support a program-by-program legislative approach; however, to be consistent with the Commission’s primary recommendation—the Risk Management Framework—we must emphasize the importance of providing flexible direction to agencies that enables them to consider risks, costs, benefits, and public values in decision-making without rigid decisional criteria.
Finding
Proposed changes in standards for judicial review would reverse years of precedent and expand the historical role of the courts in reviewing agency action.
Recommendation
The standards by which courts review agency regulatory action, exercising great deference to agency interpretations of highly technical and scientific areas, should not be revised to broaden the scope of judicial inquiry.
Historically, the standard by which courts have reviewed most agency regulatory action has been the narrow "arbitrary and capricious" standard. Under the arbitrary and capricious standard, courts consistently have held that agencies are entitled to great deference with regard to factual questions involving scientific matters in their own fields of expertise. Such deference has extended to mixed questions of law and fact, at least to the extent they have been fact-dominated. For example, in Northwest Motorcycle Association v. United States Department of Agriculture,3 an off-road vehicle (ORV) association petitioned for review of the United States Forest Service’s decision to close forest trails to ORVs in designated areas of the Wenatchee National Forest. After exhausting all administrative remedies, the ORV association argued before the United States Court of Appeals for the Ninth Circuit that the Forest Service’s conclusion was arbitrary and capricious.
In holding that the decision to close the trails was not arbitrary and capricious, the circuit court limited its review to the administrative record as required under the provisions of the APA.4 The court recited "evidence in the administrative record" that supported the Forest Service’s Findings, and cautioned that "the court here is reviewing the evidence only to determine whether such evidence existed that justified the [Forest Service’s] decision."5
The ORV association pointed to a number of alleged deficiencies in the administrative record. The court, however, replied that these deficiencies did not "mandate a Finding that the [Forest Service’s] decision was arbitrary and capricious."6 Rather, the court opined that the Forest Service, as fact-finder, was in the best position to determine the credibility of the evidence.7 Acknowledging the long-standing precedents of judicial review under the APA, the court noted that it "is not empowered by [the APA] to substitute its judgment for [the] agency."8 Thus, the basic standard for review of informal regulatory rulemaking is whether the agency action is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." The scope of review under this standard is a narrow one. In Citizens to Preserve Overton Park v. Volpe,9 the United States Supreme Court held that agency action is entitled to a "presumption of regularity" and while that does not "shield [it] from a thorough, probing, in-depth review," the "ultimate standard of review is a narrow one." The reviewing court is to search for a "clear error of judgment," and cannot "substitute judgment for that of the agency."10
A starting point for analysis of the proper standard of review is an explanation of the type of Findings and type of file that are typical to informal rulemaking. The Findings and file reviewed under the arbitrary and capricious standard differ substantially from those required in formal adjudications under the APA.11 The agency is not required to supply specific and detailed Findings and conclusions, but need only "incorporate in the rules a concise general statement of their basis and purpose." The agency need not discuss every item of fact or opinion included in the written comments submitted to it, although it must respond to significant comments and not be arbitrary and capricious. The "basis and purpose" statement must identify "what major issues of policy were ventilated by the informal proceedings and why the agency reacted to them as it did." In addition, the record "ordinarily will contain more generalized than specific information, may not contain information tested by cross-examination and will frequently contain much more conclusory information based on data gathered by interested parties."12
The court’s paramount inquiry is whether a reasoned conclusion from the record as a whole could support and explain the agency’s course of action.13
There have been proposals that would appear to greatly expand use of the broad "substantial evidence" standard now reserved for formal agency adjudications at the expense of the more narrow arbitrary and capricious standard. Proposed amendments to the APA would compel courts to hold agency action unlawful if the agency findings and conclusions are found to be "without substantial support in the rulemaking file, viewed as a whole, for the asserted or necessary factual basis …" [emphasis added]. Thus, the use of a substantial evidence standard apparently would be expanded beyond formal hearings to all rulemakings.14
While the substantial evidence standard is not a new standard of review, it typically (although not exclusively; see, for example, Toxic Substances Control Act) has been reserved for formal rulemaking and hearings. Courts have expressed some question about the application of the substantial evidence standard to informal rulemakings where the evidentiary standards and record development are different than in formal hearings (see Aqua Slide ’n’ Dive v. CPSC15 for an example). Courts that have historically deferred to agency interpretation and action under an arbitrary and capricious standard16 would, instead, have to find substantial support for that action in the agency file. Requiring a court to find substantial evidence lessens its ability to defer to agency decisions.
We believe that years of judicial and administrative precedent are well founded. Agencies, not courts, are better equipped to analyze highly scientific and technical findings. That precedent should not be legislatively overruled by expanding the standard of review.
Consensual Approaches as Alternatives to Increased Judicial Review
Finding
Consensual approaches to decision-making that could help assure rational and cost-effective regulatory actions affecting health, safety, and the environment are not commonly used as alternatives to increased judicial review.
Recommendation
Regulatory agencies should maximize consensual approaches to decision-making—such as negotiated rulemaking, alternative dispute resolution techniques, expert peer review, and informal practices such as meetings with groups of stakeholders (such as regulated parties and community representatives) and workshops—to explore alternative regulatory approaches.
Alternatives to judicial review that promote dialogue, interplay, and negotiation among regulators, the regulated community, and other stakeholders are used infrequently, other than in the context of agency policy initiatives. While variations of alternative dispute resolution procedures are sometimes used in the rulemaking and enforcement arenas, those uses clearly are the exception and not the rule.
For example, members of the regulated community, public-interest groups, and other interested parties engaged in a negotiated rulemaking process work together to analyze and discuss certain proposed regulatory initiatives. Those negotiated rulemaking sessions help the promulgating agency to better understand and develop possible alternatives to usual regulatory actions. EPA has embraced alternatives to regulatory controls with its Common Sense Initiatives, for example; for those stakeholders involved, the process has opened up communications with the regulatory agency, and it is hoped that fewer legal challenges will be filed in the course of the rulemaking process.
Chapter 3 Notes
3. 18 F.3d 1468 (9thCir. 1994).
9. 401 U.S. 402, 91 S.Ct. 814, 28 L.Ed.2d 136 (1971).
10. See Citizens to Preserve Overton Park, 401 U.S. at 415-16, 91 S.Ct. at 823-824.
11. Formal agency adjudications, on appeal, are reviewed under the substantial evidence standard.