2.1. Benefits, Quantified But Not Monetized
Benefits measurement, when fully executed, results in a dollar valuation for pollution avoided or abated. That could be called monetization. However, monetization can be viewed as the final link in a chain. Earlier links might include identification of damages, collection of data on same by experiment or observation and statistical analysis of that data to determine a dose-response or similar function. That technical relationship might be in terms of, say, excess mortality related to a given level of some pollutant. It is helpful to place a dollar figure on such relationships so that a comparison can then be made with the corresponding dollar figure for the cost of avoidance or abatement.
The chain described above is long and each link may be complex. A link in a particular chain may require the specialized skills of investigators such as epidemiologists or air quality modelers who may not have the skills to build other links. For that and perhaps other reasons, some studies do not include the final link of monetization and may not have other links, either. The chain may also be modular in the sense that links that already exist can be reused obviating the need to develop them repeatedly.
Presently, we mention studies that lack the final link and perhaps other links, as well. In addition, this section includes “framework” studies. These works develop and describe a technique for building one or more links in a benefits chain. They may also include an empirical exercise to demonstrate the method. You can see the full list of reports corresponding to this section in the benefits analysis - quantification without monetization subview of the subject view of the Environmental Economics Report Manager (EERM) database and in the benefits analysis - quantification without monetization subview of the subject view of the EPA/NSF Research Funding database.
Air Pollution and Health in Washington, DC: Some Acute Health Effects of Air Pollution in the Washington Metropolitan Area (EE0022) a 1977 report by Seskin, sought to quantify the relationship between changes in ambient air pollutants and unscheduled physician visits. Very little appeared in the data. The large, departmentalized practice he studied did tend to see slightly more unscheduled patients in its urgent care and ophthalmology departments on highly polluted days but the effect was minor. “The statistical results indicated that air pollution levels had a very limited effect on the health-care utilization of the group practice.” The study used data from 1973 and 1974.
The January 1978 study, Impact of Air Pollution on the Consumption of Medical Services – Cost of Hospitalization in the Portland Metropolitan Area (EE0018) by Bhagia and Stoevener, is a sequel to an earlier work by Jaksch and Stoevener, which measured the impact of suspended particulates on the use of outpatient medical services in the Portland, Oregon - Vancouver, Washington area. That earlier study found a statistically significant effect of air pollution on respiratory diseases. The present study failed to find a statistically significant relationship between air pollution and the use of inpatient medical services in the Portland metropolitan area.
St. Louis Pollution-Morbidity Study Procedures for Exposure Estimation (EE-0268A) and The Human Morbidity Costs of Air Pollution in St. Louis (EE-0268B) by Koontz (1978) are parts of a report on a large study of morbidity effects of St. Louis air pollution. Anticipating that collection of health data in St. Louis will be a part of the study, the investigator conducted a pretest of planned collection procedures in Montgomery County, MD. The earlier report describes that effort and finds it to be successful. Air pollution exposure data by subject is another component required in the study. The later document describes how such data were collected and collated.
Socio-Economic and Institutional Factors in Irrigation Return Flow Quality Control (EE0370) by Vlachos et al. (1978) investigates the problem of non-point water pollution resulting from the run off of irrigation water. The study consists of four volumes. The first deals with “methodology.” The other three apply its findings to case studies of the Yakima Valley, the Middle Rio Grande Valley and the Grand Valley. The report emphasizes prevention of pollution over treatment of it. More importantly, it also is concerned with questions of implementation and acceptability of the regulatory procedures. That is, consent of the regulated and all concerned is high on the authors’ list of program characteristics.
Morbidity, Air Pollution and Health Statistics (EE0396) by Ostro and Anderson (1981) contains a substantial review of statistical techniques to measure the morbidity-air pollution relationship and a discussion of issues related to those techniques. The original work of the paper is an epidemiological study. The investigators apply three statistical techniques – ordinary least squares, Tobit and Logit-OLS – to a data set. The independent variables include measures of particulates and sulfates. The dependent variable is work loss days (WLD), a proxy for morbidity. In each case, the particulate coefficient is statistically significant at the 5% level and has a positive sign, i.e., particulates increase morbidity. Also, in each case, sulfates have the “wrong” (negative) sign but are not statistically significant even at the 10% level. The regression results were combined with several assumptions including a 20% reduction in air pollution. The OLS model predicted 52 million fewer WLDs, the Tobit model predicted 45 million fewer WLDs and the logit model 50 million fewer.
Pollution Dose-Response Curves at the Micro Level (EE0038) by Hay and Gareis (undated) discusses an analysis of insurance company data on 455,000 policyholders, of whom 1,191 died during 1978-1982. They combined the vital statistics in the insurance records (e.g., age, sex, smoking habits, income, education) with environmental data from other sources by residential location. Regressions were run to determine the contribution of the factors to mortality. The pollution factors have the following effects. Fine particulate lead is positively associated with increased death risk. Fine particulate sulfates are positively associated. Fine particulate nitrates do not have significant statistical association with death probabilities. Fine particulate mass is significant and negatively associated with death risk. Fine particulate sulfur is significant and negatively associated with death risk. The vital statistics variables have expected signs.
Methods Development for Assessing Air Pollution Control Benefits Vol. V, Executive Summary (EE-0271E) written in 1979 by d’Arge, Kneese and Schulze summarizes original efforts that construct consistent and verifiable methods for assessing environmental quality improvement benefits.
Methods Development in Measuring Benefits of Environmental Improvements, Vol. I - Executive Summary (EE-0272A) from 1979 summarizes a large body of underlying research some of it dealing with experimental or contingent valuation approaches to valuing air and water quality improvements and the rest dealing with questions related to air pollution impacts on human health.
Experiments in the Economics of Air Pollution Epidemiology (EE0271A) by Crocker et al. tests air pollution data, demographic data on race and age, doctors per capita, dietary data, and smoking data as explanatory variables for mortality rates in 60 cities. Because of the simultaneous nature of the relationship between doctors and illness (doctors tend to gravitate to where there will be customers causing a correlation suggesting that doctors “cause” illness), the authors employed a two-stage estimation process. The authors found statistically significant relationships between particulates and pneumonia and between SO2 and early infant disease. The coefficients were an order of magnitude smaller that what was reported in an earlier study by Lave and Seskin that did not control for dietary variables.
The volume also explores the relationship between air pollution and various measures of acute illness such as work-days ill using a cross-sectional study of 5,000 randomly selected individuals nationwide over the nine-year period from 1967 through 1975. The study, sometimes referred to as the “Michigan Survey” as the data were collected by the Survey Research Center of the University of Michigan, is preliminary in nature.
Estimated Human Health Benefits From Chemical Plant Effluent Control In Two Regions by Industrial Economics, January 25, 1983 (EE0066), estimates the human health benefits that would result from the implementation of proposed BPT and BAT effluent standards for plants in the organic chemicals and plastics industries. The method identifies two regions as case studies (Kanawha River in West Virginia until it joins the Ohio River and the Houston Ship Channel as it flows from Houston to Galveston Bay).
Methods Development for Environmental Control Benefits Assessment Vol. VII Methods Development for Assessing Acid Deposition Control Benefits (EE-0278G) by Crocker, Tschirhart, Adams and Forster from 1985 provides “a basis for the ecological and the economic disciplines to ask better-defined questions of each other.”
Methods Development for Environmental Control Benefits Assessment Vol. II Six Studies of Health Benefits From Air Pollution Control (EE-0278B) is a 1985 collection of papers, mostly epidemiological, by numerous authors. Two deal with mortality and three with morbidity effects of air pollution.
Methods Development for Environmental Control Benefits Assessment Vol. VI The Value of Air Pollution Damages to Agricultural Activities in Southern California (EE-0278F), 1985 papers by Adams, Crocker, Thanavibulchai and Horst, estimate the losses in producer and consumer surplus and in farm workers’ incomes resulting from air pollution damages to agriculture.
Air Pollution Exposure and Lung Function in Children: A Micro-Edipemiological Study (EE0158) by Krupnick (1985) derives concentration-response functions for the effects of daily average TSP, respirable particulates, sulfate and nitrate concentrations on FEV scores of up to 4,800 elementary school children living in one of six neighborhoods of Birmingham, Alabama. Before testing for FEV began, a questionnaire was sent home with each child asking the mother to answer demographic, socioeconomic, and health-related questions, as well as whether the home contained a gas stove.
Outdoor pollution data came from six EPA monitors in the area. Linear models explained about 74% of the variance in FEV scores. Age, height and sex explained most of this variance. Educational attainment had a smaller but statistically significant effect. TSP and sulfate had the expected negative signs. Nitrates, smoking habits of the parents, and type of stove fuel used did not have a significant effect.
Effects of Air Pollution on Health Outcomes (EE0205A), by Coulson, Duan, Keeler, Korn and Manning (1985) is an attempt to identify statistically the adverse health effects of air pollution. The Rand Corporation had collected health data in Seattle and Dayton for another purpose and pressed it into service for this application as well saving time and money. The investigators have no quarrel with the data but it offers scant support of a link between air pollution and ill health. A few quotes follow. “Ozone was responsible for many of the significant effects -- largely positive (beneficial) in the case of outpatient health expenditures and time lost to illness, and both positive and negative (adverse) in the case of the health status measures… It is so commonly assumed that air pollution is bad for health under all circumstances that our finding that ozone is frequently associated with significant beneficial short-run effects seems puzzling. ... the aggregated day-to-day approach yields many significant findings, most of which are of the wrong sign, i.e., they indicate a beneficial association of air pollution with health outcomes.”
Short-Term Health Effects of Air Pollution A Case Study (EE0205B) by Duan, Hayashi, Carlson, Keeler, Kom and Manning (1987) is the second phase report of a Rand Corporation project to assess pollution’s health effects. The first phase, EE-0205A reviewed above, found little relation or a perverse one. The present study concentrates on one of the more promising statistical methods utilized in the first phase, Whittemore-Korn (WK). The earlier study used data for Dayton and Seattle collected for another purpose and applied WK to Seattle data. The present study uses the Dayton data and applies the WK method along with a Monte Carlo study to correct for possible biases in WK. The results for Dayton parallel those found earlier for Seattle using WK. “Two of the criteria pollutants, SO2 and NO2, were found to have significantly adverse health effects in both sites. Total suspended particulates were found to have significantly adverse health effects in Dayton, but not in Seattle. The rather puzzling finding in the Seattle results that ozone had a significant perverse (beneficial) health effect is not replicated in Dayton.”
In the 1986 paper Economic Valuation of Aquatic Ecosystems (EE-0237) by Fisher et al., the authors indicate the ways aquatic ecosystems are valuable and suggest how these values might be assessed.
The report Methods Development in Measuring Benefits of Environmental Improvements Vol VII Economic Benefits of Controlling the Effects of Environmental Pollution on Children’s Health by Atkinson, Crocker, Murdock and Needleman (EE0272G), which dates from the mid 1980s, contains two papers. The first paper, "The Economic Consequences of Elevated Body Lead Burdens in Children: A Proposed Study Framework" by Atkinson, Crocker and Needleman uses a household production function model to develop a framework for valuing the child-health effect of lead. As used by the authors, “child-health” refers to the child’s expected adult consumption efficiency and capital stock. The framework captures the “lead-induced changes in health status on length of schooling and schooling success” and shows “how these health status changes can influence subsequent occupational choices and life-cycle incomes.” Regression functions are derived and data sources are outlined but estimates are not made.
The second paper, "Have Priors in Aggregate Air Pollution Epidemiology Dictated Posteriors?" by Atkinson, Crocker and Murdock is an investigation of the robustness of the statistical results obtained by Lave and various associates in their studies of air pollution’s effects on human mortality. The investigators note that Lave’s critics have found radically different results by altering his functional forms or using somewhat different data. Hence, the general or methodological problem of a researcher’s beliefs (priors) dictating the result (posteriors) arises. The authors address this problem by applying techniques developed by Leamer. Although sympathetic to Lave’s conclusions, they find his conclusions to be “fragile” in the sense that they are sensitive to the econometric choices he and his associates made.
Health Econometric Methods for Multimedia Pollutants (EE0204) by Needleman and Grossman (1985) examines the effects of the six criteria air pollutants and one multimedia pollutant (lead) on blood pressure, hematocrit levels, and self- or parental- reported chronic health outcomes for subjects in the second National Health and Nutrition Examination Survey; the effects of the same pollutants on neonatal mortality; and the effects of five occupational pollutants (benzene, carbon monoxide, cadmium, lead, and solvents) on three general health outcomes (restricted-activity days in the past two weeks due to illness, work-loss days in the past two weeks due to illness, and self-rated health). Of the pollutants SO2 is identified as having the largest effect on neonatal mortality. Increased levels of CO and SO2 in ambient air and lead concdentrations in blood are associated with increased blood pressure. The relationship between occupational exposures and health outcomes is not statistically sgnificant for the data in this study.
Oxidants and Asthmatics in Los Angeles: A Benefits Analysis Environmental Benefits Analysis Series (EE-0218C) and Oxidants and Asthmatics in Los Angeles: A Benefits Analysis - Executive Summary (EE-0218A) by Rowe and Chestnut from 1986 examines behavior, expenditures and willingness to pay as related to changes in air pollution and to changes in asthma symptoms. In addition, a panel of 82 asthmatics in Glendora, California provided survey data.
The model recognized that the subjects might engage in adaptive behavior. The survey data addressed this expectation. “... the modeling efforts indicated that if respondents who are concerned about air pollution expect a bad asthma day with air pollution as a contributing aggravating factor, they will substitute behavior in a manner that may be expected to reduce adverse asthma symptoms. These changes in behavior are potentially substantial, on the order of a 20 percent reduction in time spent in active outdoor activities.” “Of the 47 respondents employed full or part-time, 20 felt their choice of job was affected by their asthma.” “Nineteen percent ... hired help on a regular basis to perform chores that they would perform themselves if their asthma were less severe. These individuals spent an average $1,478 per year for these services ....”
The 1988 study A Health Econometric Study of Diary Studies of Air Pollution and Health (EE-0213) by Wypij, Ware and Schwartz has as its purpose to refine one link in the chain between air pollution and abatement value, i.e., the statistical analysis of diary data participants maintained daily on respiratory symptoms.
A Framework for Measuring the Economic Benefits of Ground Water (EE-0259) developed in 1995 by an inter-office group within EPA sets forth a framework for all offices preparing Regulatory Impact Assessments to use in determining the direct and side effects of proposed regulations on the value of groundwater.