Economic Analysis of Environmental Externalities

Home , Economic cost

Pollution Prevention and Abatement Handbook WORLD BANK GROUP Effective July 1998 Economic Analysis of Environmental Externalities

In order to perform economic analysis of pollution prevention and abatement measures, estimates of the potential benefits from controlling pollution, as well as the better-known costs of new equipment or processes, are needed. This chapter discusses the economic analysis of environmental externalities, using a wide range of valuation techniques.

Pollutants produced by industrial activities— turn increasing recreational use and property solid wastes, toxic wastes, and substances that values. Such externalities are real costs and cause air and water pollution—may impose costs benefits attributable to the project and should on society and individuals. The identification and be included in the economic analysis as project quantification of these pollutants and the assess- costs or benefits. ment of their monetary and nonmonetary im- Conceptually, the externalities problem is pacts are important elements in a broader quite simple. Consider Figure 1, where MPC is economic analysis of the benefits and costs of the marginal production cost of a good (e.g., various production alternatives. Information on power produced by a coal-fired boiler), as per- the costs of pollution is also important in help- ceived by the project entity. Suppose that the ing decide what level of pollution control is eco- process produces a negative externality—for nomically justified. example, it emits soot that increases the main- The effects of pollution can generally be clas- tenance costs of adjacent buildings. Because the sified into four major categories: health impacts, production process also produces an external- direct and indirect effects on productivity, effects ity, the marginal social cost is higher and is given on the ecosystem, and aesthetic effects. All these by the line MSC. For any given level of output, are commonly encountered examples of economic q*, the total cost of producing that level of out- externalities of industrial production activity, that put is given by the area under the curve. The is, the externality occurs because the individual difference between the areas under the two or resource affected is not part of the enterprise’s curves gives the difference between the private decisionmaking process. For example, a factory and the social costs. The financial costs of the may emit soot that dirties surrounding buildings, project will not include the costs of the exter- increasing maintenance costs. The higher main- nality, and hence an evaluation of the project tenance costs are a direct result of the factory’s based on MPC will understate the social costs use of a resource—air—that from the plant’s of the project and overstate its net benefits. In point of view is free but that has a cost to society. principle, to account for the externality, one The same analogy applies to health impacts simply works with social rather than private linked to air pollution. Sometimes a project costs. In practice, the measurement difficulties makes certain groups better off, but the nature are tremendous because often the shape of the of the benefits is such that the project entity can- MSC curve, and hence its relationship to the not extract a monetary payment for them. A sew- MPC curve, are unknown. Also, it is not always erage and water supply project, for example, may feasible to trace and measure all external ef- not only improve water quality and yield direct fects. Nevertheless, an attempt should always health benefits but may also produce benefits be made to identify them and, if they appear from decreased pollution of coastal areas, in significant, to measure them. When externalities

54 Economic Analysis of Environmental Externalities 55

Figure 1. Private and Social Costs tion of coastal areas). A monetary value is as-

Price signed to the costs and benefits, and they are en- MSC MPC tered into the cash flow tables just as any other costs and benefits.

Project Boundaries and Time Horizon

Analysts must make two major decisions when assessing environmental impacts. First, they must decide how far to look for environmental impacts—that is, they must determine the boundary of the economic analysis. When the internal benefits and costs of a project are assessed, the boundaries of the analysis are clear: if the benefits accrue to the project entity or if the costs are borne by the project entity, they enter into the analysis. When attempting to assess the externalities of a project to determine its impact on soci- q* Quantity ety, the boundaries become blurred. Identifying externalities implies expanding the conceptual cannot be quantified, they should be discussed and physical boundaries of the analysis. An oil- qualitatively. palm mill will generate wastewater that will In some cases, it is helpful to “internalize” ex- adversely affect downstream uses of water— ternalities by considering a package of closely drinking, irrigation, and fishing. Other impacts related activities as one project—that is, to draw on the environment may be more distant or more the “project boundary” to include them. For ex- difficult to identify: the effects of emissions from ample, in the case of the soot-emitting factory, a power plant on creation of acid rain, for ex- the externality could be internalized by treating ample. How far to expand the boundaries is a the factory and the neighboring buildings as if matter of judgment and depends on the indi- they belonged to the project entity. The additional vidual project. maintenance costs then become part of the main- The second decision concerns the time horizon. tenance costs of the project entity. If the factory Like the project’s physical boundaries, its time pays for the additional maintenance costs, or if horizon becomes blurred when moving from fi- the factory is forced to install a stack that does nancial to economic analysis. A project’s environ- not emit soot, the externality again becomes in- mental impact may not last as long as the project, ternalized. In these cases, the formerly external or it may outlive it. If the environmental impact cost becomes an internal cost that is reflected in is shorter-lived than the expected economic life the accounts of the factory. of the project, the effects can be included in the standard economic analysis. If, however, the ef- Environmental Externalities fects are expected to last beyond the lifetime of the project, the time horizon must be extended. Environmental externalities are a particular This can be done in two ways: by extending the form of externalities that good economic analy- cash flow analysis a number of years, or by add- sis should take into account. Environmental ing to the last year of the project the capitalized externalities are identified as part of the envi- value of the part of the environmental impact that ronmental assessment. They are quantified extends beyond the project’s life. The latter tech- where possible and are included in the eco- nique treats the environmental impact much as nomic analysis as project costs (e.g., increased one would treat a capital good whose life extends illness, or reduced productivity of nearby farm- beyond the project’s lifetime, by giving it a “sal- lands) or benefits (such as reduction in pollu- vage value.” 56 IMPLEMENTING POLICIES: SETTING PRIORITIES

Valuation of Environmental Impacts be instances where neither the market value nor the functional relationship between the level of The first step in assessing the costs or benefits of the activity and the environmental impact is environmental impacts is to determine the rela- known. Arriving at a monetary estimate of im- tionship between the project and the environmen- pacts in such cases is very difficult. A number of tal impact, that is, to determine a relationship functional relationships that relate the level of such as that depicted in Figure 2. This Handbook activity to the degree of physical damage (or ben- provides detailed information on the likely en- efit) have been developed for various environ- vironmental impacts of many classes of indus- mental impacts. Environmental damages include trial activities. The second step is to assign a changes in production (e.g., of crops or fisheries monetary value to the environmental impact. affected by polluted water), changes in health, These two steps are equivalent to determining damage to infrastructure due to air or water pol- the shape of the MSC curve shown in Figure 1 lution, and even loss of aesthetic benefits or rec- and its relationship to the MPC curve. For ex- reational opportunities. Various methods are ample, suppose an objective of the project is to available for valuing environmental externalities. reduce air pollution, perhaps through installation The choice of valuation technique depends on of scrubbers at the industrial facility or by replac- the impact to be valued, the data and time avail- ing an old bus or taxi fleet with new, less pollut- able for the analysis, financial resources, and the ing vehicles. First, the impact of the project on social and cultural setting within which the valu- air quality, as measured by some physical char- ation exercise is to be carried out. Some valua- acteristic, is determined. Second, the monetary tion approaches are more robust, and more likely value of the improvement in air quality is as- to be applied, than are others. Figure 3 presents sessed. In most cases, it is not necessary to esti- a menu of the more commonly used valuation mate the entire cost curve; it suffices to identify approaches. the cost (or benefit) of an externality at a given Although “objective” techniques rely on ob- level of activity. That is, it is enough to estimate servable environmental changes, use market the difference between the private and the social prices, and are more “concrete” and easier to cost for a given level of activity. present to decisionmakers, subjectively based In some cases, the market value of the exter- techniques (especially those using surrogate mar- nality is not readily available. There might also kets and hypothetical markets) are increasingly accepted for decisionmaking. These subjective Figure 2. Environmental Damage as a Function methods offer the only practical way of measur- of Activity Level ing certain categories of environmentally related benefits and costs. For example, suppose one Level of activity wishes to measure the recreational benefits from preventing damage to a marine park or a pris- tine forest area. Under the travel-cost approach, the time and cost of travel are used to develop estimates of the value of the park to its users. Under the various survey-based contingent valuation methods, users are asked to state the value they place on the “park experience,” permitting an estimate to be made of consumer’s surplus associated with park use. Both are fairly robust techniques if carefully applied. It is important to remember that the simplest techniques are usually the most useful. In most Bank projects, the most useful valuation techniques will be those that rely on actual changes in production, on replacement costs or preven- Environmental impact tive expenditures, or on information about im- Figure 3. A Simple Valuation Flowchart 57

Source: Adapted from Dixon and Böjo in Dixon and others 1994. 58 IMPLEMENTING POLICIES: SETTING PRIORITIES pacts on human health (cost of illness). All these may require more frequent painting as a result deal with physical changes that can be valued of emission of pollutants by a nearby factory. The using market prices, and all are included in the higher maintenance costs should be included as objectively based set of techniques. These ap- a cost of the factory in the economic analysis. proaches are discussed in the following section. (For more detailed information on these and the Dose-Response Relationships other techniques, see Dixon et al. 1994.) and Health Outcomes

Loss in Productivity Some investment projects yield important health benefits from reduced mortality and morbidity; A project may raise or lower the productivity of examples are an increased potable water supply, another system. In these cases the valuation is improved sewage collection and treatment, and fairly straightforward. For example, in Fiji con- reduction of vehicular pollution. Some invest- version of a coastal wetland to an industrial site ments however, may have unintended but im- resulted in lower catches in a coastal fishery that portant negative impacts on health. For example, was partly dependent on the wetland. The mon- expanded industrial production or new thermal etary value of the reduced catch was an economic power plants produce important economic ben- externality attributable to the industrial devel- efits but also result in some undesirable environ- opment project and hence an economic cost of mental externalities. These health impacts should the project. The loss in production had an assess- be identified and incorporated in the economic able market value. Because the lower production analysis, either qualitatively or quantitatively. was accompanied by lower costs of production, For air pollution, a dose-response relationship the change in net benefits yielded the net impact (DRR) is commonly used to link changes in am- of the externality. Box 1 illustrates the use of the bient pollution levels to health outcomes. The change-in-production approach in a geothermal DRR is a statistically estimated relationship be- project in the Philippines. tween levels of certain pollutants in the air and In some cases, the impact of the project is not different health outcomes—illness, lost work- on the level of production but on the costs of pro- days, and so on. Although the DRR approach was duction or consumption. For example, buildings developed in the United States and Europe, there

Box 1. Assessing Disposal Alternatives for × P346/hectare = P 2,768,000) and the net return from Geothermal Wastewater in the Philippines one crop of unirrigated rice (4,000 hectares × 1 crop × P324/hectare = P1,296,000). The difference repre- The change-in-production approach was used to as- sents an annual loss of about P1.47 million. sess the impacts of various means of disposing of toxic The change-in-production approach was also ap- geothermal wastewater from a geothermal power de- plied to a coastal fishery. Various disposal options velopment project on the island of Leyte in the Philip- that did not include treatment of wastewater would pines. The analysis considered seven different cause heavy-metal pollution of coastal waters and disposal options, including reinjection of geothermal lead to closing of the coastal fishery. The cost of this wastewater, untreated disposal in local rivers, and use loss was calculated by multiplying the value of the of ocean outfalls. It estimated the economic costs of annual catch (P39.4 million) by the net return to fish- these options for irrigated rice production and an off- ing, estimated at 29%, for an annual loss valued at shore fishery. P11.4 million. Pollution of surface water would prevent its use for Both of these annual costs were then capitalized irrigation of 4,000 hectares in the dry season. Rain- to represent the economic damage to rice and fishery fed crop production would continue during the wet sea- production from environmental pollution. Other envi- son, but with lower average yields. The net return per ronmental costs were also calculated (some qualita- hectare was estimated at 346 pesos (P) for irrigated tively), and this information was used to help assess rice and P324 for rain-fed rice. The economic cost of the total benefits and costs of the various wastewater the loss of agricultural production on 4,000 hectares disposal management alternatives. would therefore be the difference between the net return from two irrigated crops (4,000 hectares × 2 crops Source: Balagot and Grandstaff 1994. Economic Analysis of Environmental Externalities 59

Box 2. Using Dose-Response Relationships standards and the more stringent WHO standards. The to Estimate Health Outcomes in Jakarta estimated numbers of lives saved and illnesses avoided per year in the population of 8.2 million are This case study illustrates the use of dose-response shown below. relationships (DRRs) to estimate the health impacts of reducing air pollution. The health impact can be Medium estimated by the following relationship: Health effect estimate

dHi = bi × POPi × dA Premature mortality (deaths) 1,200 Hospital admissions 2,000 where dH stands for the change in population risk of i Emergency-room visits 40,600 health effect i; b for the slope from the dose-response i Restricted-activity days 6,330,000 curve for health impact i; POP for the population at i Lower-respiratory illness (cases) 104,000 risk of health effect i; and dA for the change in ambi- Asthma attacks (cases) 464,000 ent air pollutant under consideration. Respiratory symptoms (cases) 31,000,000 Foreign dose-response functions were applied to Chronic bronchitis (cases) 9,600 local conditions in Jakarta to assess the benefits of reducing airborne pollution to meet both Indonesian Source: Ostro 1994.

is increasing acceptance of its transferability to methods are used in practice: estimation of will- other countries. Recent Bank work in Jakarta ingness to pay to avoid premature death, wage (Ostro 1994) and Chile (Ostro et al. 1995) illus- differential approaches, and, although not eco- trate what can be done (see Box 2). nomically sound, a “human capital” approach Whereas everyone breathes the same air in a lo- that estimates the present value of the future earn- cation, actual exposure to polluted water is the key ings of an individual that would be lost due to variable in determining whether a person becomes premature mortality. The difficulty arises when ill. Individuals can “self-insure” themselves from one compares estimates for different countries, the effects of contaminated water by, for example, especially countries with very different income boiling their water or using bottled water. Epide- levels. For example, a common value for a “sta- miologic studies are therefore usually required in tistical life” in the United States s now US$3 mil- order to estimate the impacts of changes in water lion–$5 million or more, as determined by income quality on health outcomes. Such studies take into levels and willingness to pay to avoid premature account the important social and economic factors death (see Box 3). that determine the links between contaminated Clearly this same value cannot be applied di- water and illness and death. rectly to another country with a per capita in- Once the impacts on health have been identi- come one twentieth the size of the U.S. economy. fied, they can be quantified in physical terms and, Yet, deflating the U.S. value by the relative dif- where feasible, assigned a monetary value. Sick- ference in income levels also ignores important ness is much easier than death to measure in eco- dimensions, including purchasing power parity. nomic terms. For illness, it is possible to estimate, In the absence of careful national studies of the for example, the costs of medical treatment and value of a statistical life, it is often best to present hospitalization (doctors’ visits, medicine, hospi- mortality data in terms of the number of lives tal costs, and lost work time). It is more difficult lost or saved rather than in terms of dollar value. to estimate the “cost” of pain and suffering to the sick individual, relatives, and others. The Measuring Intangibles measured costs of illness based on direct expenditures (or their appropriate shadow prices) are One of the most difficult valuation areas is mea- a minimum estimate of the true costs of illness suring subtle or dramatic changes in ecosystems, and, in turn, of the potential benefits from pre- destruction of nonsubstitutable goods (such as venting morbidity. biodiversity), effects on historical or cultural For death, we do not have an equivalent, sites, and recreational benefits. Although these equally applicable valuation approach. Various considerations are seemingly distant from this 60 IMPLEMENTING POLICIES: SETTING PRIORITIES

Box 3. Use of Statistical Techniques a particular life. However, the current consensus is that for Valuing Life the societal value of reducing risk of death cannot be based on such an estimate. Although most economists The use of loss of earnings to value the cost associ- do not favor using this method for policy analysis pur- ated with premature mortality is referred to as the poses, it is often used to establish ex-post values for human capital approach. It is similar to the change- court settlements related to the death of a particular in-production approach in that it is based on a dam- individual. age function relating pollution to production, except An alternative method of valuing reductions in risk that in this case the loss in productivity of human of death—the wage differential approach—uses infor- beings is measured. In essence, this method is an mation on the “wage premium” commonly paid to indi- ex-post, exogenous valuation of the life of a particu- viduals with risky jobs (e.g., coal miners and steel lar individual using as an approximation the present construction workers) to impute a value for an value of the lost (gross or net) market earnings of individual’s implicit valuation of a statistical death. This the deceased. value is found by dividing the wage premium by the This approach has many shortcomings. By reduc- increased chance of death; for example, a US$100 ing the value of life to the present value of an per year premium to undertake a job with a chance of individual’s income stream, the human capital ap- accidental death of 1 in 10,000 is equivalent to a value proach to the valuation of life suggests that the lives of US$1 million for a statistical death. Information on of those with high earnings are worth more than the self-insurance and other measures also gives an indi- lives of those with low earnings. As a direct conse- cation of an individual’s willingness to pay to avoid quence, the lives of residents of rich countries would premature death. be rated as more valuable than the lives of people in In many cases, a project’s impact on the environ- poor countries. Narrowly applied, the human capital ment is not apparent, but the market value of the ex- approach implies that the lives of subsistence work- ternality is assessable, albeit sometimes indirectly. For ers, the unemployed, and retirees have a low or zero example, property values decrease with the proximity value and that the lives of the underemployed have a of houses to a highway. The increased noise from traf- very low value. The very young are also valued low, fic creates a project externality that should be included since their future discounted earnings are often off- in assessing the costs of the highway. The exact rela- set by education and other costs incurred before they tionship between the highway and the noise level may entered the labor force. Furthermore, the approach be unknown, but the value of quiet surroundings can ignores substitution possibilities that people may make be assessed indirectly. For example, information from in the form of preventive health care, and it excludes another neighborhood may be used to compare the nonmarket values such as pain and suffering. value of houses that are close to a highway with the At best, this method provides a first-order, lower- value of houses farther away, controlling for differences bound estimate of the lost production associated with in other characteristics of the properties.

Handbook’s main concern, industrial pollution, in damage to buildings, equipment, and other capi- many cases an important benefit from control- tal goods as a result of pollution. Cleaner air will ling pollution may be the protection or enhance- also improve visibility—an important but ment of a recreational site or important natural unpriced benefit. Ideally, the visibility benefits habitat. It is possible, although difficult, to esti- should also be entered into the economic analy- mate economic values for, say, the consumer’s sis. Because of data and measurement difficul- surplus of visitors to parks and protected areas, ties, however, these measures are usually entered by using the travel cost approach or conducting into the analysis only qualitatively. contingent valuation studies. Recent work in East Africa is incorporating the results of such Preventing and Mitigating studies in the analysis of projects (see Box 4). In- Environmental Impacts tangible benefits often include important environmental benefits that are secondary to the Sometimes a project can go ahead only if the primary benefits produced by a project. Air pol- implementing agency takes measures to prevent lution control projects in Santiago and Mexico or mitigate its environmental impact. If the im- City, for example, will yield primary benefits pact is completely prevented, the costs of pre- from reduced health effects and reductions in vention are taken into consideration in the Economic Analysis of Environmental Externalities 61

Box 4. Valuing Consumer Surplus of average increase in willingness to pay per trip of $24 International Tourists in Madagascar per tourist. If 3,900 foreign tourists visit the new park— a conservative assumption that uses the same num- This example presents an application of the travel cost ber as currently visit the Perinet Reserve—the annual and contingent valuation methods for estimating some “benefit” to foreign tourists would be $93 600. of the benefits associated with the creation of a new The contingent valuation method was used to es- park in Madagascar. A strong point of the study is that timate directly the value of the proposed park to for- it used questionnaires based on two different valua- eign tourists. Visitors to the Perinet Forest Reserve tion techniques to estimate consumer surplus and were provided with information about the new park. compared the estimated results. Using a discrete choice format, they were asked how Questionnaires were prepared and administered much more they would have been willing to pay for to visitors to the small Perinet Forest Reserve adja- their trip to Madagascar if in the new national park (a) cent to the proposed Mantadia National Park. Visitors they saw twice as many lemurs, and (b) they saw the tended to be well off and well educated, with, on aver- same number of lemurs as on their current visit. Since age, annual income of $59,156 and 15 years of edu- most of these visitors are expected to visit Madagas- cation. The average stay in Madagascar was 27 days. car only once, their response represents a one-time, Data from the visitor survey, supplemented with data lump-sum payment they are willing to make in order from tour operators, was used in an econometric to have the park available. Mean willingness to pay analysis that employed the travel cost approach. To for the park (conditional on seeing the same number estimate demand by international tourists, traditional of lemurs) was $65. Assuming current visitation pat- travel cost models have to be reformulated because terns, the total annual willingness-to-pay for the park people who travel to a country like Madagascar en- would be $253,500. gage in a variety of activities, of which the visit to the The information from these two estimates could proposed national park would be only one. The travel be used to help design policies to capture part of cost model was then used to predict the benefits to this willingness to pay and compensate nearby vil- tourists (the increase in consumer surplus) under the lagers for income lost when the establishment of assumption that the new national park would result in the park prevents their traditional activities within a 10% increase in the quality of local guides, educa- the park. tional materials, and facilities for interpreting natural areas in Madagascar. The estimation indicated an Source: Kramer 1993; Kramer et al. 1993. economic and financial analysis of the project. If References and Sources a factory is required to install equipment to elimi- nate air pollution, there is no environmental im- Balagot, B., and S. Grandstaff. 1994. “Tongonan Geo- pact. If the factory is merely required to mitigate thermal Power Plant: Leyte, Philippines.” Case the environmental impact, the cost of the miti- Study 5 in Dixon et al. (1994). gating action is a direct and identifiable cost of Dixon, John A., and Paul B. Sherman. 1990. Econom- the project, but the value of the residual environ- ics of Protected Areas: A New Look at Benefits and mental impact also needs to be considered in the Costs. Washington, D.C.” Island Press. costs of the project. If a dam reduces fish catch Dixon, John A., Louise F. Scura, Richard A. Carpenter, downstream, despite mitigating measures, the and Paul B. Sherman. 1994. Economic Analysis of Envi- reduction of the catch is still a cost of the project. ronmental Impacts. London: Earthscan Publications. Care however, must be taken to avoid double Kramer, Randall A. 1993. “Tropical Forest Protection counting. If the favored solution to an environ- in Madagascar.” Paper presented at Williams Col- mental impact is to let the damage occur, tax the lege, Williamstown, Mass. culprit, and then repair the damage, the cost of Kramer, R. A., Mohan Munasinghe, N. Sharma, E. the project should include the environmental cost Mercer, and P. Shyamsundar. 1993. “Valuation of only once—as the cost of repairing the environ- Biophysical Resources in Madagascar.” In Mohan mental damage or as the tax (if the tax is exactly Munasinghe, ed., Environmental Economics and Sus- equal to the cost of repairing the environment), tainable Development. World Bank Environment Pa- but not both. per 3. Washington, D.C. 62 IMPLEMENTING POLICIES: SETTING PRIORITIES

Ostro, Bart. 1994. “Estimating the Health Effects of Air Ostro, Bart, José Miguel Sanchez, Carlos Aranda, and Pollution: A Methodology with an Application to Gunnar S. Eskeland. 1995. “Air Pollution and Mor- Jakarta.” Policy Research Working Paper 1301. tality: Results from Santiago, Chile.” Policy Re- World Bank, Policy Research Department, Washing- search Working Paper 1453. World Bank, Policy ton, D.C. Research Department, Washington, D.C.