Jump to main content.

2.2. Risk Communication and Other

Quick Links

Environmental Economics Research at EPA


A number of studies deal with benefits questions. Several of them deal with method. Mostly, they concentrate on avoidance as a benefit and more specifically with risk or risk assessment techniques. Others deal with toxicology, epidemiology and pesticide technology. You can see the full list of reports corresponding to this section in the benefits analysis - risk communication and other subview of the subject view of the Environmental Economics Report Inventory (EERI) database and the corresponding full list of NSF/EPA projects in the benefits analysis - other subview of the subject view of the EPA/NSF Research Funding database.

Risk Communication

A Multivariate Characterization of Risk (EE0236) by Carson and Horowitz (1990) deals with the psychology of risk. The authors attempt to determine what aspects of risk are of particular concern. Dread, whether exposure is voluntary, newness, familiarity, technological complexity and catastrophic potential are all risk aspects. The paper relies on three major surveys and two minor ones plus a focus group. The investigators find “that subjects’ reactions to environmental health risks are based primarily on how large they believe the risks are and how effective they believe they or the government would be in reducing them.” “The most important risk characteristics for explaining subjects’ attitudes toward risk are Personal Risk, Future Deaths, and Current Deaths.” “The next most important risk characteristics are Government Effectiveness and Personal Effectiveness. Subjects prefer government action toward substances for which government effectiveness in reducing the risks is high ...”

The 1991 study Incorporating Uncertainty in Risk Communication (EE-0223) by W. Kip Viscusi and Wesley A. Magat comprises three papers. They all focus on the reaction of respondents to hypothetical, risky situations. The investigators note anomalies in the way subjects responds. For example, they place very different values on different kinds of deaths. They place more weight on more recently presented risk data even when there is no temporal order to the data. Nevertheless, individuals show learning behavior and do not respond in an alarmist manner. The findings may help sharpen contingent value studies.

The third of the papers, entitled "Ambiguity and Health Risks" (with coauthors Huber and Payne), is a study of reaction to uncertainty as it affects health. Uncertainty may mean the absence of a probability distribution. The authors conduct and cite psychological experiments. They reach several conclusions. First, “... when several studies of a risk yield widely divergent estimates of that risk, people are likely to respond to those estimates as if their best estimate of the risk exceeds the average of the reported risk estimates, and the greater the range of those estimates, the larger the degree of ambiguous belief aversion.” The individual’s assessment of risk will be sensitive to the order in which risk estimate studies are presented. If they are presented in ascending order of risk estimated, the individual will have a higher risk assessment than if the order is reversed. Third, “... when people understand that several studies have produced identical estimates of the health risk, or if they are presented with only one risk estimate ... from an unknown population, they are more averse to the risk than if the multiple studies showed divergent estimates (with the same mean risk).” There are some suggestions as to how regulatory agencies might adapt their public communications in light of these findings.

The 1988 paper Valuing Risk in Experimental Markets: Self-Protection, Self-Insurance, and Collective Action (EE-0249) by Jason F. Shogren is an exercise in game theory and experimental economics. Its purpose is to improve the design and accuracy of contingent value field studies. The goal of the project is to determine if and how the value of risk differs under various reduction mechanisms. Risk is a characteristic of many environmental policy decisions. Risk and other factors of interest can be portrayed to respondents more accurately in a laboratory, hypothetical setting than in a field test. The investigator found that feedback after an intial bid helps produce a more stable second bid. For example, if the subject is told the average bid, that would constitute feedback. Even without providing feedback, the initial bid is an accurate predictor of the second bid.

In the 1993 paper Community Preferences and Superfund Responsibilities (EE0042) Cantor explores an aspect of the Superfund program which EPA by law must deal with: negotiations and communications with affected parties including communities, users of the site in question and local governments. The different parties have different interests and different incentives under varying remediation (cost sharing) arrangements. The study includes a pilot effort to determine the preferences of the constituent groups at one site. Clearly, the preferences of local government leaders differ from the preferences of neighbors of the site on many of the issues. The purpose of the effort is to shed some light on underlying differences to facilitate dispute resolution.

The NSF/EPA-funded study from 1991, “Communicating Environmental Risks” by Viscusi explores the degree of uncertainty and ambiguity that exists in the minds of people about the risks they may face varies according to the methods used in communicating those risks and is of particular importance to their response. This project explores how people respond to ambiguous information about environmental risks, how EPA should best communicate those risks, and how individuals respond to different forms of information about the environmental risks they may face. A survey of different ways of communicating risks will be used and analyzed.

Risk and Uncertainty

Problems in the Use of Toxicological Data for Human Risk Assessment (EE-0102) by Martyn T. Smith delivers the author’s thoughts and reservations on the use of animal testing models to estimate toxicity in humans. He notes with dismay that high doses are substituted for long-term exposure. He also notes that the test subjects are typically a genetically inbred, homogeneous strain of small rodents whereas the concern is for the outbred, extremely heterogeneous human population. Further, symptoms which might take years or decades to appear in human are sought almost immediately after exposure in the test animals. “In conclusion then, acute toxicity testing is usually cruel and of little relevance to man.”

Dose-Response Assessment in Regulatory Risk Analysis (EE-0103) by Kenneth T. Bogen discusses various issues in the field of toxicology as they pertain to risk assessment. The document is a review of established methods in the field and their evolution. The author believes the risk analysis process suffers from a failure to incorporate any explicit consideration of uncertainty in the dose-response relationship that is selected for use in the final risk analysis.

Raucher’s 1987 report Evaluating the Risk Tradeoffs of Drinking Water Disinfection (EE0037) provides a discussion of the risk-risk tradeoff involved in chlorinating drinking water. Chlorine is an effective and inexpensive disinfectant but presents carcinogenic and other health hazards. The study surveys some of the economics and psychology literature dealing with risk. The paper is more descriptive than analytical.

The 1984 report by Harrison, et al. entitled Research and Demonstration of Improved Methods for Carrying Out Benefit-Cost Analyses of Individual Regulations Vol IV Strategies for Dealing with Uncertainty in Individual Regulations (EE0269D) contains three papers. The third paper, "The Value of Acquiring Information Under Section 8(a) of the Toxic Substances Control Act: A Decision-Analytic Approach," by Nichols, Boden, Harrison and Terrell is another paper about uncertainty and its antidote, information. Section 8(a) of the Toxic Substances Control Act permits the Administrator of the Environmental Protection Agency to promulgate rules requiring chemical manufacturers to submit detailed information about their specified products. The subject paper considers the use of the rule and alternatives including gathering no further data. An 8(a) rulemaking, is a costly and time-consuming process, which may explain its infrequent use. The investigators conclude that a decision-analytic framework, which they outline, is the preferred approach. They also provide an 8(a) case study of ethylene dichloride.

Carrying Out Benefit-Cost Analyses of Individual Regulations Vol IV Strategies for Dealing with Uncertainty in Individual Regulations (EE0269D) contains three papers. In "An Overview of Scientific Uncertainties in Benefit Estimation," John S. Evans and Katherine Walker discuss uncertainty in benefits estimates that result from our ignorance of the various links in the chain of causation between emissions and damages. The chain can be divided into two parts, exposure assessment and hazard assessment. The former involves determining the spatial, concentration and temporal pattern of human exposure to a pollutant. “Hazard assessment is the determination of the health risks posed by exposure to the pollutant ... obtained from the exposure” assessment. In turn, both of these assessments can be further subdivided into other tasks or chain links. Each link has its own uncertainties or outcomes distribution. For example, dose-response studies done on small animals can be extrapolated to humans but the extrapolation is subject to uncertainty. The final, resulting uncertainty between abatement and benefit may be very great.

A Program of Economics Research on Improving Estimation of Benefits from Reduced Pollution (EE-0033), a 1981 report, discusses EPA information needs with respect to the measurement of economic benefits of pollution control, grouping them into three categories: (1) development of methods and further benefit assessments; (2) data needs; and (3) reduction of overall system uncertainties in national benefit estimates.

Suggestions for further research included:
  • further work using statistical methods to estimate the dose response relationship of humans to environmental pollution.
  • using wage premiums, property price differentials, and other market prices to value morbidity and mortality risks.
  • comparing different survey techniques for assessing willingness to pay.
  • applying survey techniques to measure visibility, existence values, option values, and anxiety concerning threats to health and well being.
  • examining contributions to environmental public interest groups as a basis for inferring amenity, option, existence and anxiety-reduction benefits.
  • estimating agricultural damages from environmental pollution using farm cost functions.
  • pilot studies on health effects from pollution in the workplace; damage to materials from pollution; and methods for valuing reductions in anxiety.

    The report also identified several data needs related to human health and pollution; participation in water-related recreation activities; and damage to materials due to pollution. Finally, the report suggested a pilot project to determine the sources and magnitude of uncertainties in national benefit estimates from air and water pollution control.

    Other Studies

    Research Needs and Priorities: Water Pollution Control Benefits and Costs - Vol II (EE-0210B) by Jordening and Allwood (1973) is an early action plan to research the issue of water pollution. Much has happened in a quarter century. The report correctly notes that recreation and aesthetics are the main remediation benefit categories. It favors partial equilibrium approaches over general equilibrium for reasons of expediency. It sets forth data requirements and methods development tasks.

    The 1979 report entitled Studies on Partial Equilibrium Approaches to Valuation of Environmental Amenities, Volume IV of Methods Development for Assessing Air Pollution Control Benefits (EE-0271D) is a collection of four papers. The first paper, "Public Goods Decisions Within the Context of a General Competitive Economy," by William R. Porter is an attempt to reconcile partial equilibrium analysis of a public good such as pollution abatement with general equilibrium. The investigator concludes that “partial equilibrium methods of cost-benefit analysis will not lead to allocations that are Pareto superior if the project is of discrete size,” i.e., large enough to affect prices. He also concludes that his “approach emphasizes the logical impossibility of separating costs from benefits and valuation from taxation and trade.” Several proposals are offered to rectify the identified problems with partial equilibrium analysis.

    The 1983 study Economic Benefits From Control of Major Environmental Episodes (EE-0098) by Mark F. Sharefkin contains three case studies of high-profile, large-scale environmental incidents: Kepone contamination of the James River, contamination of the Cohansey aquifer by Price’s landfill and the Chemical Control incident. The case studies illustrate methods for estimating avoidance benefits. For each study, the principal damage categories are identified, and data are assembled from which those damage categories are estimated. The author notes that each facility was operated lawlessly and questions whether marginal changes in incentives would have avoided incidents where avoidance costs were clearly less than damages. He conjectures on the efficacy of performance bonds.

    Gregory’s 1985 paper Measures of Consumer’s Surplus: Interpreting the Disparity in Views (EE-0058) explores the relation between willingness to pay and compensation demanded as measures of economic losses for environmental goods. As many previous studies have demonstrated, though economic theory suggests the two approaches should yield similar estimates, willingness to pay is typically much less than compensation demanded. Several experiments with subjects revealed that asking and interpreting valuation questions is quite complex. Though it has promise in the view of the author, much more additional research is urgently needed, especially to explain the persistent gaps between WTP and WTA.

    The Econometrics of Pesticide Use: Why Specification Matters (EE-0104) by Erik Lichtenberg and David Zilberman discusses the technology and econometrics of pesticide use. The authors argue that pesticides are different from other inputs to the agricultural production process. Pesticides cause pests to evolve to become more resistant to the pesticide. The perverse effect is to increase the marginal productivity and demand for the pesticide. Appropriate production functions and econometric methods are discussed. Understanding better the productivity and value of pesticides may help in developing regulations for them.

    Framework -- Case Study Design for a Risk Benefit Analysis of Pesticides in the Special Review Process (EE0105), by Zilberman and Lichtenberg (1986), offers several closely related sections on the welfare effects of pesticide regulation. The first augments the conventional welfare analysis of regulation by extending the competitive model to account for pre-existing policies that subsidize the agricultural sector. Consumer surplus surprisingly can, in some such situations, be enhanced by regulation.

    The second section examines econometric modeling of pesticide productivity. The authors note that previous studies “... found average values of marginal pesticide productivity several times greater than marginal cost, suggesting that pesticides are underutilized in U.S. agriculture.” They explain that this anomaly results from a specification of the regression equation that ignores the development of pesticide resistance over time. They also develop econometric methods that better capture this phenomenon and apply them to cotton production.

    The third section considers re-entry regulation. That is, farm workers are excluded from fields where certain toxic pesticides have been applied until specified safety intervals have elapsed. Corresponding harvest delay regulations exist to protect consumers. The authors model the productivity and health effects of these regulations to capture the inherent trade-off.

    Methods Development in Measuring Benefits of Environmental Improvements Vol IV Valuing Ecosystem Functions: The Effects of Air Pollution (EE0272D) by Crocker, Tschirhart, Adams, and Katz, a report from the early 1980s, attempts to develop a link between ecosystems and economies that will allow an economic evaluation of ecosystem structure and diversity.” The authors develop a model of nature in economic terms and then introduce humans and economic activity. Energy and entropy play central roles in their effort. The second part of the study is a contingent valuation exercise to determine willingness to pay to avoid air pollution damage to a forest. The final section of the study attempts to develop a method for deriving a dose-response relationship for agricultural production to air pollution and applies that method to the effect of ozone on four crops: corn, cotton, soybeans and wheat.

    The 1994 memorandum Inventory of Ecosystem Effects (EE0364) by Bell, Deck, Michaels and Paciorek provides an inventory of ecosystem benefits that could be attributed to reductions in nitrogen oxide emissions. The purpose is to more fully describe the benefits from reducing nitrogen oxide emissions at the coal-fired electric power plants subject to Phase I acid rain rules.

    The inventory summarizes previous and ongoing work to understand the dynamics of ecosystem impacts of nitrogen oxide emissions. Several types of ecosystem services are considered: direct (hiking/camping, birdwatching, scenic beauty, boating, commercial fishing, recreational fishing, and waste assimilation); and indirect services (storm protection and wave buffering, flood control, nutrient removal, pollutant uptake, sediment control, biodiversity preservation, and agricultural pest control). The authors are not optimistic regarding the prospects for future economic valuation work concerning the damage to ecosystems caused by nitrogen oxide emissions. In their view, the likely magnitude of economic loss is too small to justify the major effort that would be required.

    The 1984 report entitled Research and Demonstration of Improved Methods for Carrying Out Benefit-Cost Analyses of Individual Regulations Vol. II Benefit Methodologies Applied to Ecological Hazards From Toxic Substances (EE0269B) contains two papers. The first, "Methods for Estimating the Benefits from Mitigating Ecological Damages from Toxic Chemicals" by Robert Repetto is an abstract investigation of how pollution damage to various species might be valued. This problem entails some measure of philosophy because not all damage may be to marketable commodities. The paper includes models linking species population to economic activity and the environment.

    The second paper, "The Assessment of Ecological Hazards from Pesticides: The Use of Qualitative Modelling in Decision Analysis" by Robert Repetto and Anthony C. Janetos develops a framework for testing pesticides prior to possible registration. The investigators note that prevailing methods seek the effect of a pesticide on each species individually. However, in an ecosystem, there may be ripple effects even if the first exposed species is uninjured. They propose that information be gathered sequentially with a decision made at each stage as to how to proceed. Some of the information may be qualitative.

    Methodology Report For Benefits Associated With PM Reductions Resulting From Reduced Diesel Fuel Sulfur and Aromatics (EE0371) by Duff and Brennan (1988) describes how the authors would value some of the benefits of reducing particulates and sulfur dioxide that are produced when diesel fuel is burned. These benefits are obtained by reducing sulfur and aromatics in the fuel. The model is divided into five sequential stages. The first is to identify air quality changes. A base case of no controls and four control cases are planned. Second is to develop estimates of the exposed population. There is a spatial component to this effort to reflect differential exposures as well as population projections into the estimated years. The third stage is estimation of the health or welfare changes, ie, risk assessment. Here, three benefits are identified: reduced mortality, reduced morbidity and reduced soiling. Models produced earlier by others will constitute this stage.

    The fourth stage is valuation of health or welfare changes. Here, too, the effort will rely on the works of others. As an example, the value of a statistical life would be pegged at $1.87 million ($1987) with a range of $0.47 -$3.31 million. Comparable values would be retrieved for the other benefits. The final stage, aggregation across benefit categories, involves projecting benefits over the entire spatially-distributed population.

    If the goals of the Federal Water Pollution Control Act are to be met, control measures will be required for nonpoint source pollution. The 1979 report Costs and Water Quality Impacts of Reducing Agricultural Nonpoint Source Pollution (EE0024) by Meta Systems states that it should be feasible to model various control measures for agricultural nonpoint runoff with respect to both cost and and effectiveness.

    Comprehensive Management of Phosphorus Water Pollution, a 1974 report by Porcella et al, (EE0035) begins the development of methods for analyzing and assessing impacts of pollutants on surface waters and for estimating the cost-effectiveness of alternative control strategies. In the analysis, costs are limited to those for treatment, though the authors recognize the desirability of including other economic costs. The model contains no information on benefits, though the authors suggest that they, too, would be useful.

    The NSF/EPA-funded study from 1996, “Aggregative and Deliberative Contexts for Valuation: A Philosophical Contribution to Experimental Research in Environmental Decision Making” by Sagoff will consider conventional aggregative procedures versus unconventional deliberative ones for reaching environmental decisions. In an aggregative procedure, an expert decision maker aggregates the preferences of individuals to develop a policy. In a deliberative procedure, the individuals with diverse viewpoints and interests are convened by the expert and they reach a decision democratically perhaps with technical input from the expert.


  • Local Navigation


    Jump to main content.