Environmental Economics: A Survey

Maureen L. Cropper; Wallace E. Oates

Journal of Economic Literature, Vol. 30, No. 2. (Jun., 1992), pp. 675-740.

Stable URL: http://links.jstor.org/sici?sici=0022-0515%28199206%2930%3A2%3C675%3AEEAS%3E2.0.CO%3B2-D

Journal of Economic Literature is currently published by American Economic Association.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/aea.html.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is an independent not-for-profit organization dedicated to and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact [email protected].

http://www.jstor.org Mon May 7 12:41:42 2007 journal of Economic Literature Vol. XXX ('June 1992), pp. 675-740

Environmental Economics: A Survey

BY MAUREENL. CROPPERAND WALLACEE. OATES University of Ma yhnd and Resources for the Future

Both authors are members of the Department of Economics, Uniuer- sity of Maryland, and are Fellows at Resources for the Future. We are grateful for many valuable comments on earlier drafts of this paper to a host of economists: Nancy Bockstael, Gardner Brown, Richard Carson, John Cumberland, Diane DeWitt, Anthony Fisher, A. Myrick Freeman, Tom Grigalunas, Winston Harrington, Robert Hahn, Charles Howe, Dale Jorgenson, Charles Kolstad, Ray Kopp, Allen Kneese, Alan Krupnick, Randolph Lyon, Ted McConnell, Albert McGartland, Robert Mitchell, Arun Malik, Roger Noll, Raymond Palmquist, John Pezzey, Paul Portney, V. Kery Smith, Tom Tieten- berg, and James Tobey. Finally, we want to thank Jonathan Dunn, Joy Hall, Dan Mussatti, and Rene Worley for their assistance in the preparation of the manuscript.

I. Introduction the analysis was of a fairly general charac- ter, there was at least some careful re- HEN THE ENVIRONMENTAL~~VO~U~~O~search underway exploring the applica- Warrived in the late 1960s, the eco- tion of economic solutions to certain nomics profession was ready and waiting. pressing environmental problems (e.g., Economists had what they saw as a coher- Allen Kneese and Blair Bower 1968). ent and compelling view of the nature The economist's view had-to the dis- of pollution with a straightforward set of may of the profession-little impact on policy implications. The problem of ex- the initial surge of legislation for the con- ternalities and the associated market fail- trol of pollution. In fact, the cornerstones ure had long been a part of microeco- of federal environmental policy in the nomic theory and was embedded in a United States, the Amendments to the number of standard texts. Economists Clean Air Act in 1970 and to the Clean saw pollution as the consequence of an Water Act in 1972, explicitly prohibited absence of prices for certain scarce envi- the weighing of benefits against costs in ronmental resources (such as clean air the setting of environmental standards. and water), and they prescribed the in- The former directed the Environmental troduction of surrogate prices in the form Protection Agency to set maximum limi- of unit taxes or "effluent fees" to provide tations on pollutant concentrations in the the needed signals to economize on the atmosphere "to protect the public use of these resources. While much of health"; the latter set as an objective the 676 Journal of Economic Literature, Vol. XXX (June 1992)

"elimination of the discharge of all [our der the new 1990 Ainendinents to the emphasis] pollutants into the navigable Clean Air Act. More broadly, an innova- waters by 1985."' tive report from within the U.S. Con- The evolution of environmental policy, gress sponsored by Senators Timothy both in the U. S. and elsewhere, has inev- Wirth and John Heinz, Project 88: Har- itably brought economic issues to the nessing Market Forces to Protect Our fore; environmental regulation has neces- Environment (Robert Stavins 1988) ex- sarily involved costs-and the question plores a lengthy list of potential applica- of how far and how fast to push for pollu- tions of economic incentives for environ- tion control in light of these costs has mental management. Likewise, there is entered into the public debate. Under widespread, ongoing discussion in Eu- Executive Order 12291 issued in 1981, rope of the role of economic measures many proposed environmental measures for pollution control. Most recently in have been subjected to a benefit-cost January of 1991, the Council of the Orga- test. In addition, some more recent nization for Economic Cooperation and pieces of environmental legislation, nota- Development (OECD) has gone on rec- bly the Toxic Substances Control Act ord urging member countries to "make (TSCA) and the Federal Insecticide, a greater and more consistent use of eco- Fungicide, and Rodenticide Act (FI- nomic instruments" for environinental FRA), call for weighing benefits against management. Of particular note is the costs in the setting of standards. At the emerging international concern with same time, economic incentives for the global environinental issues, especially containment of waste discharges have with planetary warming; the enornious crept into selected regulatory measures. challenge and awesome costs of policies In the United States, for example, the to address this issue have focused interest 1977 Amendments to the Clean Air Act on proposals for "Green Taxes" and sys- introduced a provision for "emission off- tems of tradable permits to contain global sets" that has evolved into the Emissions einissions of greenhouse gases. In short, Trading Program under which sources this seems to be a time when there is a are allowed to trade "rights" to emit air real opportunity for environmental econ- pollutants. And outside the United omists to make some valuable contribu- States, there have been some interesting tions in the policy arena-if, as we shall uses of effluent fees for pollution control. argue, they are willing to move from This is a most exciting time-and per- "purist" solutions to a realistic consider- haps a critical juncture-in the evolution ation of the design and implementation of econoinic incentives for environmental of policy measures. protection. The Bush Administration Our survey of environinental econom- proposed, and the Congress has intro- ics is structured with an eye toward its duced, a measure for the trading of sulfur policy potential. The theoretical founda- emissions for the control of acid rain un- tions for the field are found in the theory of externalities. And so we begin in Sec- 'Although standards were to be set solely on the tion I1 with a review of the theory of basis of health criteria, the 1970 Amendments to the environmental regulation in which we Clean Air Act did include economic feasibility among explore recent theoretical results regard- its guidelines for setting source-specific standards. Roger Noll has suggested that the later 1977 Amend- ing the choice among the key policy in- ments were, in fact, more "anti-economic" than any struments for the control of externalities: that went before. See Matthew McCubbins, Roger effluent fees, subsidies, and marketable Noll, and Barry Weingast (1989)for a careful analysis of this legislation. emission permits. Section I11 takes us Cropper and Oates: Environmental Economics 677

from the theory of externalities to policy mists have shown considerable ingenuity applications with a focus on the structur- in the development of techniques- ing and implementation of realistic mea- known as indirect market methods-that sures for environmental management. exploit the relationships between envi- This section reviews the work of environ- ronmental quality and various marketed mental economists in trying to move from goods. These methods allow us to infer formal theorems to measures that ad- the value of improved environmental dress the variety of issues confronting an amenities from the prices of the market environmental regulator. We describe goods to which they are, in various ways, and evaluate briefly, as part of this treat- related. Second, environmental econo- ment, the U.S. and European experi- mists have turned to an approach re- ences with economic incentives for pollu- garded historically with suspicion in our tion control. In addition, we explore a profession: the direct questioning of series of regulatory issues--centraliza- individuals about their valuation of en- tion versus decentralization of regulatory vironmental goods. Developing with authority, international effects of domes- considerable sophistication the so-called tic environmental policies, and enforce- "contingent valuation" approach, they ment-matters on which environmental have been able to elicit apparently reli- economists have had something to say. able answers to questions involving the In Section IV, we turn to the measure- valuation of an improved environment. ment of the benefits and costs of environ- In Section IV, we explore these various mental programs. This has been a partic- methods for the valuation of the benefits ularly troublesome area for at least two and costs of environmental programs and reasons. First, many of the benefits and present some empirical findings. costs of these programs involve elements In Section V, we try to pull together for which we do not have ready market our treatment of measuring benefits and measures: health benefits and aesthetic costs with a review of cases where bene- improvements. Second, policy makers, fit-cost analyses have actually been used perhaps understandably, have proved re- in the setting of environmental stan- luctant to employ monetary measures of dards. This provides an opportunity for such things as "the value of human life" an overall assessment of this experience in the calculus of environmental policy. and also for some thoughts on where such Environmental economists have, how- analyses are most needed. We conclude ever, made some important strides in the our survey in Section VI with some re- valuation of "nonmarket" environmental flections on the state of environmei~tal services and have shown themselves able economics and its potential contribution to introduce discussion of these measures to the formulation of public policy. in more effective ways in the policy Before turning to substantive matters, arena. we need to explain briefly how we have In a survey in this Journal some fifteen defined the boundaries for this survey. years ago, Anthony Fisher and Frederick For this purpose, we have tried to distin- Peterson (1976) justifiably contended guish between "environmental econom- that techniques for measuring the bene- ics" and "natural resource economics." fits of pollution control are "to be taken The distinguishing characteristic of the with a grain of salt" (p. 24). There has latter field is its concern with the inter- been considerable progress on two dis- temporal allocation of renewable and tinct fronts since this earlier survey. n6hrenewable resources. With its origins First, environmental (and other) econo- in the seminal paper by Harold Hotelling 678 Journal of Economic Literature, Vol. XXX (June 1992)

(1931), the theory of natural resource terizes pollutioi~as a public "bad" that economics typically applies dynamic con- results from "waste discharges" associ- trol methods of analysis to problems of ated with the production of private intertemporal resource usage. This has goods. The basic relationships can be ex- led to a vast literature on such topics as pressed in abbreviated form as: the management of fisheries, forests, minerals, energy resources, the extinc- tion of species, and the irreversibility of development over time. This body of work is excluded from our survey. The where the assumed signs of the partial precise dividing line between environ- derivatives are Ux> 0, UQ < 0, X, > mental economics and natural resource 0, X, > 0, XQ < 0, and QE > 0. The economics is admittedly a little fuzzy, but utility of a representative consumer in in order to keep our task a manageable equation (1) depends upon a vector of one, we have restricted our survey to goods consumed (X) and upon the level what we see as the two major issues in of pollution (Q). Pollution results from ei~vironmentaleconomics: the regulation waste emissioi~s(E) in the production of of polluting activities and the valuation X, as indicated in (2). Note that the pro- of enviroi~mentalamenities. duction function in (2) is taken to ii~clude as inputs a vector of conventional inputs 11. The Normative Theory of (L), like labor and capital, the quantity Environmental Regulation of waste discharges (E), and the level of The source of the basic economic prin- pollution (Q). In this formulation, waste ciples of environmental policy is to be emissions are treated simply as another found in the theory of externalities. The factor of production; this seems reason- literature on this subject is enormous; it able since attempts, for example, to cut encompasses hundreds of books and pa- back on waste discharges will involve the pers. An attempt to provide a compre- diversion of other inputs to abatement hensive and detailed description of the activities-thereby reducing the availa- literature on externalities theory reaches bility of these other inputs for the pro- beyond the scope of this survey. Instead, duction of goods. Reductions in E, in we shall attempt in this section to sketch short, result in reduced output. More- an outline of what we see as the central over, given the reasonable assumption results from this literature, with an em- of rising marginal abatement costs, it phasis on their implications for the design makes sense to assume the usual curva- of environmental policy. We shall not ad- ture properties so that we can legiti- dress a number of formal matters (e.g., mately draw isoquants in L and E space problems of existence) that, although im- and treat them in the usual way. portant in their own right, have little to say about the structure of policy mea- example, William Baumol (1972), Baurnol and Wal- sures for protection of the environment. lace Oates (1988), Paul Burrows (1979), and Richard Cornes and Todd Sandler (1986). We have not in- A. The Basic Theory of Enuironmental cluded in this survey a literature on conservation policy2 and development that has considered issues of irre- versibility in the time of development for which the The standard approach in the envi- seminal papers are John Krutilla (1967), and Kennetlt ronmental economics literature charac- Arrow and Anthony Fisher (1974). This literature is treated in the Anthony Fisher and Peterson survey For comprehensive and rigorous treatments of (1976) and, more recently, in Anthony Fisher (1981, the general ideas presented in this section, see, for ch. 5). Cropper and Oates: Environmental Economics 679

The production function also includes ity. This exercise produces a set of first- as an argument the level of pollution (Q), order conditions for a Pareto-efficient since pollution may have detrimental ef- outcome; of interest here is the condition fects on production (such as soiling the taking the form: output of the proverbial laundry or re- ducing agricultural output) as well as pro- ducing disutility to consumers. The level of pollution is itself some function of the vector of emissions (E) of all the produc- ing units. In the very simplest case, Q might be taken to equal the sum of the Equation (5) indicates that polluting emissions over all producers. firms should extend their waste dis- One extension of the model involves charges to the point at which the mar- the explicit introduction of "defensive" ginal product of these emissions equals activities on the part of "victims." We the sum of the marginal damages that might, for example, amend the utility they impose on consumers [the first sum- function: mation in (5)] and on producers [the sec- ond summation in (5)]. Or, put slightly differently, (5)says that pollution-control to indicate that individuals can employ measures should be pursued by each pol- a vector of inputs (L) to lessen, in some luting agent to the point at which the sense, their exposure to pollution. The marginal benefits from reduced pollution level of pollution to which the individual (summed over all individuals and all is actually exposed (F) would then de- firms) equal marginal abatement cost. pend upon the extent of pollution (Q) Another of the resulting first-order and upon the employment of inputs in conditions relates to the efficient level defensive activities (L). We could obvi- of defensive activities: ously introduce such defensive activities aua~- auax for producers as well. We thus have a (6) aF aL aL set of equations which, with appropriate ax subscripts, would describe the behavior which says simply that the marginal value of the many individual households and of each input should be equated in its firms that comprise the system. use in production and defensive activi- It is a straightforward exercise to maxi- ties. mize the utility of our representative in- The next step is to derive the first-or- dividual (or group of individuals) subject der conditions characterizing a competi- to (2) and (3) as constraints along with a tive market equilibrium, where we find further constraint on resource availabil- that competitive firms with free access to environmental resources will continue to engage in polluting activities until the This highly simplifed model, although useful for our analytical purposes, admittedly fails to encompass marginal return is zero, that is, until the complexity of the natural environment. There aXIaE = 0. We thus obtain the familiar is an important literature in environmental econom- result that because of their disregard for ics that develops the "materials-balance" approach the external costs that they impose on to environmental analysis (see Kneese, Robert Ayres, and Ralph d'Arge 1970; Karl-Coran Maler 1974, others, polluting agents will engage in 1985). This approach introduces explicitly the flows socially excessive levels of polluting ac- of environmental resources and the physical laws to tivities. which they are subject. Some of these matters will figure in the discussion that follows. The policy implication of this result is 680 Journal of Economic Literature, Vol. XXX (June 1992) clear. Polluting agents need to be con- (1960) argument is that in the absence fronted with a "price" equal to the mar- of transactions costs and strategic behav- ginal external cost of their polluting activ- ior, the distortions associated with exter- ities to induce them to internalize at the nalities will be resolved through volun- margin the full social costs of their pur- tary bargains struck among the interested suits. Such a price incentive can take the parties. No further inducements (such as form of the familiar "Pigouvian tax," a a Pigouvian tax) are needed in this setting lev), on the polluting agent equal to mar- to achieve an efficient outcome. In fact, ginal social damage. In the preceding for- as Ralph Turvey (1963) showed, the in- mulation, the tax would be set equal to troduction of a Pigouvian tax in a Coasian the expression in equation (5).Note fur- setting will itself be the source of distor- ther that the unit tax (or "effluent fee") tions. Our sense, however, is that the must be attached directly to the polluting Coasian criticism is of limited relevance activity, not to some related output or to most of the major pollution problems. input. Assuming some substitution Since most cases of air and water pollu- among inputs in production, the Pigou- tion, for example, involve a large number vian tax would take the form of a levy of polluting agents and/or victims, the per unit of waste emissions into the envi- likelihood of a negotiated resolution of ronment-not a tax on units of the firm's the problem is small-transactions costs output or an input (e.g., fossil fuel associ- are simply too large to permit a Coasian ated with p~llution).~ resolution of most major environmental The derivation of the first-order condi- problems. It thus seems to us that a tions characterizing utility-maximizing Nash or "independent adjustment" behavior by individuals yields a second equilibrium is, for most environmental result of interest. Inasmuch as defensive issues, the appropriate analytical frame- activities in the model provide only pri- work. In this setting, the Pigouvian cure vate benefits, we find that individual for the externality malady is a valid maximizing behavior will satisfy the first- one. 5 order conditions for Pareto efficiency for Second, there has been no mention such activities. Since they are confronted of any compensation to the victims of ex- with a given price for each input, Individ- ternalities. This is an important point- uals will allocate their spending so that and a source of some confusion in the a marginal dollar yields the same incre- literature-for Coase and others have ment to utility whether it is spent on suggested that in certain circumstances consumption goods or defensive activi- compensation of victims for damages by ties. There is no need for any extra in- polluting agents is necessary for an effi- ducement to achieve efficient levels of cient outcome. As the mathematics defensive activities. makes clear, this is not the case for our Although this is quite straightforward, model above. In fact, the result is even there are a couple of matters requiring stronger: compensation of victims is not further comment. First, the Pigouvian permissible (except through lump-sum solution to the problem of externalities transfers). Where victims have the op- has been the subject of repeated attack portunity to engage in defensive (or aloug Coasian lines. The Ronald Coase "averting") activities to mitigate the ef- fects of the pollution from which they 'Where it is not feasible to monitor emissions di- rectly, the alternative may be to tax an input or out- For comparative analyses of the bargaining and put that is closely related to emissions of the tax approaches to the control of externalities, see pollutal~t This gives rise to a standard sort of second- Daniel Bromley (1986), and Jonathan Hamilton, Ey- best problem in taxation. tan Sheshinski, and Steven Slutsky (1989). Cropper and Oates: Environmental Economics 681 suffer, compensation cannot be allowed. activities. There are, however, other ap- For if victims are compensated for the proaches to establishing the proper eco- damages they suffer, they will no longer nomic incentives for abatement activi- have the incentive to undertake efficient ties. Two alternative policy instruments levels of defensive measures (e.g., to lo- have received extensive attention in the cate away from polluting factories or em- literature: unit subsidies and marketable ploy various sorts of cleansing devices). emission permits. As is clear in the preceding formulation, It was recognized early on that a sub- the benefits from defensive activities are sidy per unit of emissions reduction could private in nature (they accrue solely to establish the same incentive for abate- the victim that undertakes them) and, ment activity as a tax of the same mag- as a result, economic efficiency requires nitude per unit of waste discharges: a no incentives other than the benefits thev subsidy of 10 cents per pound of confer on the ~ictim.~ sulfur emissions reductions creates the The basic theoretical result then (sub- same opportunity cost for sulfur emis- ject to some qualifications to be discussed sions as a tax of 10 cents per unit of later) is that the efficient resolution of sulfur discharges. From this perspec- environmental externalities calls for pol- tive, the two policy instruments are luting agents to face a cost at the margin equivalent: the regulator can use ei- for their polluting activities equal to the ther the stick or the carrot to create value of the damages they produce and the desired incentive for abatement ef- for victims to select their own levels of forts. defensive activities with no compensa- It soon became apparent that there are tion from polluters. We consider next some important asymmetries between some policy alternatives for achieving these two policy instruments (e.g., Mor- this result. ton Kamien, Nancy L. Schwartz, and F. B. The Choice Among Policy Trenery Dolbear 1966; D. Bramhall and lnstruments7 Edwin Mills 1966; Kneese and Bower 1968). In particular, they have quite dif- The analysis in the preceding section ferent implications for the profitability of has run in terms of a unit tax on polluting production in a polluting industry: subsi- dies increase profits, while taxes de- There may, of course, exist cases where defensive crease them. The policy instruments thus activities have "publicness" properties-where the have quite different implications for the actions of one victim to defend himself against pollu- long-run, entry-exit decisions of firms. tion also provide defense for others. In such cases, there is clearly an externality present so that individ- The subsidy approach will shift the indus- ual maximizing behavior will not yield the efficient try supply curve to the right and result levels of defensive activities. For a careful and thor- in a larger number of firms and higher ough examination of defensive activities, see Richard Butler and Michael Maher (1986). Incidentally, the industry output, while the Pigouvian tax general issue of compensation of victims from pollu- will shift the supply curve to the left with tion obviously has much in common with the moral a consequent contraction in the size hazard problem in insurance. 'A further policy instrument not discussed in this of the industry. It is even conceivable section but with some potentially useful applications that the subsidy approach could result in environmental policy is deposit-refund systems in an increase in the total amount of pol- (Peter Bohm 1981). Such systems can shift some of the responsibility for monitoring and effectively place lution (Baumol and Oates 1988, ch. 14, the burden of proof on the source. For under this Stuart Mestelman 1982; Robert Kohn approach, the source, to recoup its deposit, must 1985). demonstrate that its activities have not damaged the environment. See Robert Costanza and Charles Per- The basic point is that there is a further rings (1990) for a policy proposal under this rubric. condition, an entry-exit condition, that Journal of Economic Literature, Vol. XXX (June 1992) long-run equilibrium must satisfy for an set either "price" or "quantity" and efficient outcome (William Schulze and achieve the desired result. d'Arge 1974; Robert Collinge and Oates This symmetry between the price and 1982; Daniel Spulber 1985). To obtain quantity approaches is, however, criti- the correct number of firms in the long cally dependent upon the assumption of run, it is essential that firms pay not only perfect knowledge. In a setting of imper- the cost of the marginal damages of their fect information concerning the marginal emissions, but also the total cost arising benefit and cost functions, the outcomes from their waste emissions. Only if firms under the two approaches can differ in bear the total cost of their emissions will important ways. the prospective profitability of the enter- C. Environment Policy Under prise reflect the true social net benefit Uncertainty of entry and exit into the industry.8 In sum, unit subsidies are not a fully In a seminal paper, Martin Weitzrnan satisfactory alternative to Pigouvian (1974) explored this asymmetry between taxes (Donald Dewees and W. A. Sims price and quantity instruments and pro- 1976). duced a theorem with important policy In contrast, in a world of perfect knowl- implications. The theorem establishes edge, marketable emission permits are, the conditions under which the expected in principle, a fully equivalent alternative welfare gain under a unit tax exceeds, to unit taxes. Instead of setting the is equal to, or falls short of that under a proper Pigouvian tax and obtaining the system of marketable permits (quotas). efficient quantity of waste discharges as In short, the theorem states that in the a result, the environmental authority presence of uncertainty concerning the could issue emission permits equal in the costs of pollution control, the preferred aggregate to the efficient quantity and policy instrument depends onthe rela- allow firms to bid for them. It is not hard tive steepness of the marginal benefit and to show that the market-clearing price cost curves.1° will produce an outcome that satisfies the 'The discussion glosses over some quite trouble- first-order conditions both for efficiency some matters of implementation. For example, the in pollution abatement activities in the effects of the emissions of a particular pollutant on short run and for entry-exit decisions in ambient air or water quality will often depend impor- tantly on the location of the source. In such cases, the long run. The regulator can, in short, the optimal fee must be tailored to the damages per unit of emissions source-by-source. Or, alternatively, in a market for emission permits, the rate at which In an intriguing qualification to this argument, permits are traded among any two sources will vary Martin Bailey (1982) has shown that not only subsi- with the effects of their respective emissions. In such dies to polluters, but also compensation to victims, a setting, programs that treat all sources uniformly will result in no distortions in resource use where can forego significant efficiency gains (Eugene Seskin, benefits and damages are capitalized into site rents. Robert Anderson, and Robert Reid 1983; Charles For a discussion of the Bailey argument, see Baumol Kolstad 1987). More on all this shortly. and Oates (1988, pp. 23034). In another interesting lo This result assumes linearity of the marginal ben- extension, Gene Mumy (1980) shows that a combined efit and cost functions over the relevant range and charges-subsidy scheme can be fully efficient. Under that the error term enters each function additively. this approach, sources pay a unit tax for emissions Uncertainty in the benefits function, interestingly, above some specified baseline, but receive a unit is not enough in its own right to introduce any asym- subsidy for emissions reductions below the baseline. metries; while it is the source of some expected wel- The key provision is that the right to subsidy pay- fare loss relative to the case of perfect information, ments is limited to existing firms (i.e., new sources there is no difference in this loss as between the have a baseline of zero) and that this right can either two policy instruments. For useful diagrammatic be sold or be exercised even if the firm chooses to treatments of the Weitzman analysis, see'Zvi Adar exit the industry. For a useful development of and James Griffin (1976), Gideon Fishelson (1976), Mumy's insight, see John Pezzey (1990). and Baumol and Oates (1988, ch. 5). Cropper and Oates: Environmental Economics 683

The intuition of the Weitzman proposi- Spence (1976) have shown, the regulator tion is straightforward. Consider, for ex- can set the quantity of permits at the ample, the case where the marginal ben- level that equates expected marginal efits curve is quite steep but marginal benefits and costs and then offer a sub- control costs are fairly constant over the sidy for emissions reductions' in excess relevant range. This could reflect some of those required by the permits and also kind of environmental thresh'old effect a unit tax to provide a kind of "escape where, if pollutant concentrations rise hatch" in case control costs turn out to only slightly over some range, dire envi- be significantly higher than anticipated. ronmental consequences follow. In such In this way, a combination of price and a setting, it is clearly important that the quantity instruments can;' in a setting of environmental authority have a close imperfect information, provide a larger control over the quantity of emissions. expected welfare gain than an approach If, instead, a price instrument were em- relying on either policy instrument alone ployed and the authority were to under- (see also Weitzman 1978)." estimate the true costs of pollution con- trol, emissions might exceed the critical D. Market Imperfections range with a resulting environmental di- The efficiency properties of the policy saster. In such a case, the Weitzman measures we have discussed depend for theorem tells us, quite sensibly, that the their validity upon a perfectly competi- regulator should choose the quantity in- tive equilibrium. This is a suspect as- strument (because the marginal benefits sumption, particularly since many of the curve has a greater absolute slope than major polluters in the real world are large the marginal cost curve). firms in heavily concentrated industries: Suppose, next, that it is the marginal oil refineries, chemical companies, and abatement cost curve that is steep and auto manufacturers. This raises the issue that the marginal benefits from pollution of the robustness of the results to the control are relatively constant over the presence of large firms that are not price relevant range. The danger here is that takers in their output markets. because of imperfect information, the James Buchanan (1969) called attention regulatory agency might, for example, to this issue by showing that the imposi- select an overly stringent standard, tion of a Pigouvian tax on a monopolist thereby imposing large, excessive costs could conceivably reduce (rather than on polluters and society. Under these cir- raise) social welfare. A monopolist re- cumstances, the expected welfare gain stricts output below socially optimal lev- is larger under the price instrument. Pol- els, and a tax on waste emissions will luters will not get stuck with inordinately lead to yet further contractions in output. high control costs, since they always have The net effect is unclear. The welfare the option of paying the unit tax on emis- gains from reduced pollution must be off- sions rather than reducing their dis- set against the losses from the reduced charges further. output of the monopolist. The Weitzman theorem thus suggests The first-best response to this conun- the conditions under which each of these two policy instruments is to be preferred I' Butler and Maher (1982) show that in a setting to the other. Not surprisingly, an even of economic growth, the shifts in the marginal dam- better expected outcome can be obtained age and marginal control cost schedules are likely to be such as to increase substantially the welfare by using price and quantity instruments loss from a fixed fee system relative to that from a in tandem. As Marc Roberts and Michael system of marketable permits. 684 Journal of Economic Literature, Vol. XXX Uune 1992) drum is clear. The regulatory authority regulator will have either the information should introduce two policy measures: a needed or the authority to determine and Pigouvian tax on waste emissions plus a impose a set of taxes on waste emissions unit subsidy to output equal to the differ- that is differentiated by the degree of mo- ence between marginal cost and marginal nopoly power. Suppose that the environ- revenue at the socially optimal level of mental authority is constrained to levying output. Since there are two distortions, a uniform tax on waste discharges and two policy instruments are required for suppose that it determines this tax in a a full resolution of the problem. Environ- Pigouvian manner by setting it equal to mental regulators, however, are unlikely marginal social damages from pollution, to have the authority (or inclination) to completely ignoring the issue of market subsidize the output of monopolists. In imperfections. How badly are things the absence of such subsidies, the agency likely to go wrong? Oates and Diana might seek to determine the second-best Strassmann (1984) have explored this tax on effluents. Dwight Lee (1975) and question and, using some representative Andy Barnett (1980) have provided the values for various parameters, conclude solution to this problem by deriving for- that the complications from monopoly mally the rule for the second-best tax on and other noncompetitive elements are waste emissions. The rule calls for a unit likely to be small in magnitude; the losses tax on emissions that is somewhat less from reduced output will typically be than the unit tax on a perfectly competi- "swamped" by the allocative gains from tive polluter (to account for the output reduced pollution. They suggest that, effect of the tax): based on their estimates, it is not unrea- sonable simply to ignore the matter of incremental output distortions from ef- fluent fees.12 Their analysis suggests fur- ther that the failure of polluting agents Equation (7) indicates that the second- to minimize costs because of more com- best tax per unit of waste emissions (t*) plex objective functions (a la William- equals the Pigouvian tax on a perfectly son), public agencies of the Niskanan competitive firm (t,) minus the welfare sort, or because of regulatory constraints loss from the reduced output of the mon- on profits need not seriously undermine opolist expressed as the difference be- the case for pricing incentives for pollu- tween the value of a marginal unit of out- tion control. This subject needs further put and its cost times the reduction in study, especially since many of the prin- output associated with a unit decrease cipal participants in the permit market in waste emissions. It can be shown by for trading sulfur allowances under the the appropriate manipulation of (7) that new Amendments to the Clean Air Act the second-best tax on the monopolist will be regulated firms. varies directly with the price elasticity E. On the Robustness of the Pigouvian of demand. The rationale is clear: where Prescription: Some Further Matters demand is more price elastic, the price distortion (i.e., the divergence between Although the literature has estab- price and marginal cost) tends to be lished certain basic properties of the Pi- smaller so that the tax on effluent need not be reduced by so much as where de- '*For more on this issue, see Peter Asch and Jo- mand is more price inelastic. seph Seneca (1976), Walter Misiolek (1980), and Bur- It seems unlikely, however, that the rows (1981). Cropper and Oates: Environmental Economics 685 gouvian solution to the problem of exter- come.14 Under such circumstances, nalities, there are some remaining trou- equilibrium prices may tell us nothing blesome matters. One concerns the about the efficiency of current output information requirements needed to im- or the direction in which to seek improve- plement the approach. Developing reli- ment. able measures of the benefits and costs There are thus reasons for some real of environmental amenities is, as we shall reservations concerning the direct appli- see shortly, a difficult undertaking. To cation of the Pigouvian analysis to the determine the appropriate Pigouvian formulation of environmental policy. It levy, moreover, we not only need mea- is to this issue that we turn next. sures of existing damages and control costs, but we need to develop measures 111. The Design and Implementation of of the incremental costs and benefits over Environmental Policy a substantial range. For the proper Pi- gouvian levy is not a tax equal to marginal A. Introduction: From Theory to Policy social damages at the existing level of pol- lution; it is a tax equal to marginal dam- Problems of measurement and the ages at the optimal outcome. We must breakdown of second-order conditions effectively solve for the optimal level of (among other things) constitute formida- pollution to determine the level of the ble obstacles to the determination of a tax. As an alternative, we might set the truly first-best environmental policy. In tax equal to the existing level of damages response to these obstacles, the litera- and then adjust it as levels of pollution ture has explored some second-best ap- change in the expectation that such an proaches to policy design that have ap- iterative procedure will lead us to the pealing properties. Moreover, they try socially optimal outcome. But even this to be more consistent with the proce- is not guaranteed (Baumol and Oates dures and spirit of decision making in 1988, ch. 7). the policy arena. There is, moreover, a closely related Under these approaches, the determi- problem. In the discussion thus far, we nation of environmental policy is taken have examined solely the first-order con- to be a two-step process: first, standards ditions for efficient outcomes; we have or targets for environmental quality are not raised the issue of satisfying any sec- set, and, second, a regulatory system is ond-order conditions. As Baumol and designed and put in place to achieve David Bradford (1972) have shown, this these standards. This is often the way is a particularly dangerous omission in environmental decision making pro- the presence of externalities.13 In fact, ceeds. Under the Clean Air Act, for ex- they demonstrate that if a detrimental ample, the first task of the EPA was to externality is of sufficient strength, it set standards in the form of maximum must result in a breakdown of the convex- ity-concavity conditions required for an I4This problem is further compounded by the optimal outcome. As a result, there may presence of defensive activities among victims of pol- easily exist a multiplicity of local maxima lution. The interaction among abatement measures by polluters and defensive activities by victims can from which to choose-with no simple be a further source of nonconvexities (Hirofumi Shi- rule to determine the first-best out- bata and Steven Winrich 1983; Oates 1983). Yet an- other source of nonconvexities can be found in the structure of subsidy programs that offer payments l3 See also Richard Portes (1970), David Starrett for emissions reductions to firms in excess of some (1972), J. R. Gould (1977), and Burrows (1986). minimum size (Raymond Palmquist 1990). 686 Journal of Economic Literature, Vol. XXX (June 1992) permissible concentrations of the major the least-cost levels. (See Thomas Tieten- air pollutants. The next step was to de- berg 1985, ch. 3, for a useful survey of sign a regulatory plan to attain these stan- these cost studies.) dards for air quality. The source of these large cost savings In such a setting, systems of economic is the capacity of economic instruments incentives can come into play in the sec- to take advantage of the large differentials ond stage as effective regulatory instru- in abatement costs across polluters. The ments for the achievement of the pre- information problems confronting regula- determined environmental standards. tors under the more traditional CAC ap- Baumol and Oates (1971) have described proaches are enormous-and they lead such a system employing effluent fees as regulators to make only very rough and the "charges and standards" approach. crude distinctions among sources (e.g., But marketable permit systems can also new versus old firms). In a setting of per- function in this setting-a so-called "per- fect information, such problems would, mits and standards" approach (Baumol of course, disappear. But in the real and Oates 1988, ch. 12) l5 world of imperfect information, eco- The chief appeal of economic incen- nomic instruments have the important tives as the regulatory device for achiev- advantage of economizing on the need ing environmental standards is the large for the environmental agency to acquire ~otentialcost-savings that they promise. information on the abatement costs of in- There is now an extensive body of empir- dividual sources. This is just another ical studies that estimate the cost of example of the more general principles achieving standards for environmental concerning the capacity of markets to quality under existing command-and- deal efficiently with information prob- control (CAC) regulatory programs (e.g., lems.16 Scott Atkinson and Donald Lewis 1974; The estimated cost savings in the stud- Seskin, Anderson, and Reid 1983; Alan ies cited above result from a more cost Krupnick 1983; Adele Palmer et al. 1980; effective allocation of abatement efforts Albert McGartland 1984). These are typi- within the context of existing control cally programs under which the environ- technologies. From a more dynamic per- mental authority prescribes (often in spective, economic incentives promise great detail) the treatment procedures additional gains in terms of encouraging that are to be adopted by each source. the development of more effective and The studies compare costs under CAC less costly abatement techniques. As programs with those under a more cost John Wenders (1975) points out in this effective system of economic incentives. context, a system that puts a value on The results have been quite striking: they any discharges remaining after control indicate that control costs under existing (such as a system of fees or marketable programs have often been several times permits) will provide a greater incentive to R&D efforts in control technology than will a regulation that specifies some given l5 This is admittedly a highly simplified view of the policy process. There is surely some interplay level of discharges (see also Wesley Ma- in debate and negotiations between the determina- gat 1978, and Scott Milliman and Ray- tion of standards and the choice ofpolicy instruments. mond Prince 1989). More broadly, there is an emerging literature on the political economy of environmental policy that seeks to provide a better understanding of the pro- l6 There is also an interesting literature on incen- cess of instrument choice-see, for example, McCub- tive-compatible mechanisms to obtain abatement cost bins, Noll, and Weingast (1989), and Robert Hahn information from polluters-see, for example, Evan (1990). Kwerel (1977). Cropper and Oates: Environmental Economics 687

B. The Choice of Policy Instruments do so later in the presence of economic gain" growth and a rising price level. The regu- latory agency will have to enact periodic Some interesting issues arise in the (and unpopular) increases in effluent choice between systems of effluent fees' fees. In contrast, a system of marketable and marketable emission permits in the permits automatically accommodates it- policy arena (John H. Dales 1968; De- self to growth and inflation. Since there wees 1983; David Harrison 1983). There can be no change in the aggregate quan- is, of course, a basic sense in which they tity of emissions without some explicit are equivalent: the environmentaI au- action on the part of the agency, in- thority can, in principle, set price (i.e., creased demand will simply translate it- the level of the effluent charge) and then self into a higher market-clearing price adjust it until emissions are reduced suffi- for permits with no effects on levels of ciently to achieve the prescribed envi- waste discharges. ronmental standard, or, alternatively, is- Polluters (that is, existing polluters), sue the requisite number of permits as well as regulators, are likely to prefer directly and allow the bidding of pollu- the permit approach because it can in- ters to determine the market-cleating volve lower levels of compliance costs. price. If the permits are auctioned off, then of However, this basic equivalence ob- course polluters must pay directly for the scures some crucial differences between right to emit wastes as they would under the two approaches in a policy setting; a fee system. But rather than allocating they are by no means equivalent policy the permits by auction, the environmen- instruments from the perspective of a tal authority can initiate the system with regulatory agency. A major advantage of a one-time distribution of permits to ex- the marketable permit approach is that isting sources-free of charge. Some it gives the environmental authority di- form of "grandfathering" can be used to rect control over the quantity of emis- allocate permits based on historical per- sions. Under the fee approach, the regu- formance. Existing firms thus receive a lator must set a fee, and if, for example, marketable asset, which they can then the fee turns out to be too low, pollution use either to validate their own emissions will exceed permissible levels. The or sell to another polluter.18 ,4nd finally, agency will find itself in the uncomforta- the permit approach has some advan- ble positio of having to adjust and re- tages in terms of familiarity. Regulators adjust the fee to ensure that the environ- have long-standing experience with per- mental standard is attained. Direct mits, and it is a much less radical change control over quantity is to be preferred to make permits effectively transferable since the standard itself is prescribed in than to introduce a wholly new system quantity terms. of regulation based on effluent fees. Mar- This consideration is particularly im- portant over time in a world of growth '"n an interesting simulation study, Randolph and inflation. A nominal fee that is ade- Lyon (1982) finds that the cost of permits to sources quate to hold emissions to the requisite under an auction system can be quite high; for one of the auction simulations, he finds that aggregate levels at one moment in time will fail to payments for permits will exceed treatment costs. Lyon's results thus suggest potelltially large gains to polluting firms from a free distribution of permits li For a useful, comprehensive survey of the instead of their sale through an auction. These gains, strengths and weaknesses of alternative policy instru- of course, are limited to current sources. Polluting ments for pollution control, see Bohm and Clifford firms that arrive on the scene at a later date will Russell (1985). have to purchase permits from existing dischargers. 688 Journal of Economic Literature, Vol. XXX (June 1992)

ketable permits thus have some quite ap- "Polluter Pays Principle" on the grounds pealing features to a regulatory agency- that those who use society's scarce envi- features that no doubt explain to some ronmental resources should compensate degree the revealed preference for this the public for their use. approach (in the U.S. at least) over that There exists a large literature on the of fees. design of fee systems and permit markets Effluent charges have their own ap- to attain predetermined levels of envi- peal. They are sources of public revenue, ronmental quality. This work addresses and, in these days of large budget defi- the difficult issues that arise in the design cits, they promise a new revenue source and functioning of systems of economic to hard-pressed legislators. From an eco- incentives-issues that receive little or nomic perspective, there is much to be only perfunctory attention in the purely said for the substitution of fees for other theoretical literature but are of real con- sources of revenues that carry sizable ex- cern in the operation of actual policy cess burdens (Lee and Misiolek 1986). In measures. For example, there is the a study of effluent charges on emissions tricky matter of spatial differentiation. of particulates and sulfur oxides from sta- For most pollutants, the effect of dis- tionary sources into the atmosphere. Da- charges on environmental quality typi- vid Terkla (1984) estimates, based on as- cally has important spatial dimensions: sumed levels of tax rates, that revenues the specific location of the source dictates in 1982 dollars would range from $1.8 the effects that its emissions will have to $8.7 billion and would, in addition, on environmental quality at the various provide substantial efficiency gains ($630 monitoring points. While, in principle, million to $3.05 billion) if substituted for this simply calls for differentiating the revenues from either the federal individ- effluent fee according to location, in prac- ual income tax or corporation income tax. tice this is not so easy. The regulatory Moreover, the charges approach does agency often does not have the authority not depend for its effectiveness on the or inclination to levy differing tax rates development of a smoothly functioning on sources according to their location. market in permits. Significant search Various compromises including the con- costs, strategic behavior, and market im- struction of zones with uniform fees have perfections can impede the workings of been investigated (Tietenberg 1978; Ses- a permit market (Hahn 1984; Tietenberg kin, Anderson, and Reid 1983; Kolstad 1985, ch. 6). In contrast, under a system 1987). of fees, no transfers of permits are Similarly, problems arise under sys- needed-each polluter simply responds tems of transferable permits where (as directly to the incentive provided by the is often the case) the effects of the emis- existing fee. There may well be circum- sions of the partners to a trade are not stances under which it is easier to realize the same. (The seminal theoretical paper a cost-effective pattern of abatement ef- is W. David Montgomery 1972.) Several forts through a visible set of fees than alternatives have been proposed includ- through the workings of a somewhat dis- ing zoned systems that allow trades only torted permit market. And finally, there among polluters within the specified is an equity argument in favor of fees zones, ambient permit systems under (instead of a free distribution of permits which the terms of trade are determined to sources). The Organization for Eco- by the relative effects of emissions at nomic Cooperation and Development binding monitors, and the pollution-off- (OECD), for example, has adopted the set system under which trades are sub- Cropper and Oates: Environmental Economics 689 ject to the constraint of no violations of control.20 The major program of this the prevailing standard at any point in genre is the EPA's Emission Trading Pro- the area (Atkinson and Tietenberg 1982; gram for the regulation of air quality. But Atkinson and Lewis 1974; Hahn and No11 there are also three other programs wor- 1982; Krupnick, Oates, and Eric Van de thy of note: the Wisconsin system of Verg 1983; McGartland and Oates 1985; Transferable Discharge Permits (TDP) McGartland 1988; Tietenberg 1980, for the management of water quality, the 1985; Walter Spofford 1984; Baumol and lead trading program (known formally as Oates 1988, ch. 12). For certain pollu- "interrefinery averaging"), and a recent tants, these studies make clear that a sub- program for the trading of rights for phos- stantial portion of the cost-savings from phorus discharges into the Dillon Reser- economic-incentive approaches will be voir in color ad^.^^ lost if spatial differentiation is not, at least By far the most important of these pro- to some degree, built into the program grams in terms of scope and impact, (Robert Mendelsohn 1986). Emissions Trading has undergone a fairly The actual design of systems of eco- complicated evolution into a program nomic incentives inevitably involves that has several major components. Un- some basic compromises to accommodate der the widely publicized "Bubble" pro- the range of complications to the regula- vision, a plant with many sources of emis- tory problem (Albert Nichols 1984). sions of a particular air pollutant is It is instructive to see how some of subjected to an overall emissions limita- these issues have been dealt with in tion. Within this limit, the managers of practice. the plant have the flexibility to select a C. Experience with Economic lncentives set of controls consistent with the aggre- for Environmental Managementlg gate limit, rather than conforming to specified treatment procedures for each In the United States proposals for ef- source of discharges with the plant. Un- fluent fees have met with little success; der the "Netting" provision, firms can however, there has been some limited avoid stringent limitations on new experience with programs of marketable sources permits for the regulation of air and water quality. In Europe, the experience (at 20 One case in which there has been some use of fees in the U.S.is the levying of charges on industrial least until quite recently) has been the emissions into municipal waste treatment facilities. reverse: some modest use of effluent In some instances these charges have been based charges but no experience with transfera- not only on the quantity but also on the strength or quality of the effluent. The charges are often related ble permits. We shall provide in this sec- to "average" levels of discharges and have had as tion a brief summary of these measures their primary objective the raising of funds to help along with some remarks on their finance the treatment plants. Their role as an eco- nomic incentive to regulate levels of emissions has achievements and failures. apparently been minor (see James Boland 1986; Bau- Largely for the reasons mentioned in mol and Oates 1979, pp. 2.5843). There are also a the preceding section, policy makers in variety of taxes on the disposal of hazardous wastes, including land disposal taxes in several states. the U.S. have found marketable permits Tietenberg's book (1985)is an excellent, compre- preferable to fees as a mechanism for pro- hensive treatment of the Emissions Trading Program. viding economic incentives for pollution Robert Hahn and Gordon Hester have provided a series of recent and very valuable descriptions and assessments of all four of these programs of market- l9 The OECD (1989) has recently provided a useful able permits. See Hahn and Hester (1989a, 1989b), "catalog" and accompanying discussion of the use of and Hahn (1989). For analyses of the Wisconsin TDP economic incentives for environmental protection in system, see William O'Neil (1983), and O'Neil et the OECD countries. al. (1983). Journal of Economic Literature, Vol. XXX (June 1992)

of discharges by reducing emissions from regulations which effectively prohibit other sources of the pollutant within the certain kinds of trades. Many potentially facility. Hahn and Hester (1989b) report profitable trades simply have not come that to date there have been over 100 to pass.22 approved Bubble trapsactions in the U.S. Likewise, the experience under the and a much larger 'number of Netting Wisconsin TDP system has involved lit- "trades" (somewhere between 5,000 and tle external trading. The program estab- 12,000). The estimated cost savings from lishes a framework under which the these trades have been quite substantial; rights to BOD discharges can be traded although the estimates exhibit a very among sources. Since the program's in- wide range, the cost savings probably ception in 1981 on the Fox River, there amount to several billion dollars. has been only one trade: a paper mill There are provisions under Emissions which shifted its treatment activities to Trading for external trades across firms- a municipal wastewater treatment plant mainly under the Offset provision which transferred its rights to the municipal fa- allows new sources in nonattainment ar- cility. The potential number of trades is eas to "offset" their new emissions with limited since there are only about twenty reductions in discharges by existing major sources (paper mills and municipal sources. Offsets can be obtained through waste treatment plants) along the banks either internal (within plant) or external of the river. But even so, preliminary trades. Hahn and Hester (1989b) indicate studies (O'Neil 1983; O'Neil et al. 1983) that there have been about 2,000 trades indicated several potentially quite profit- under the Offset policy; only about 10 able trades involving large cost savings. percent of them have been external A set of quite severe restrictions appears trades-the great bulk of offsets have to have discouraged these transfers of been obtained within the plant or facility. permits. Trades must be justified on the Emissions Trading, as a whole, re- basis of "needD-and this does not in- ceives mixed marks. It has significantly clude reduced costs! Moreover, the increased the flexibility with which traded rights are granted only for the sources can meet their discharge limita- term of the seller's discharge permit (a tions-and this has been important for maximum period of five years) with no it has allowed substantial cost savings. assurance that the rights will be re- The great majority of the trades, how- newed. The Wisconsin experience seems ever, have been internal ones. A real and to be one in which the conditions needed active market in emissions rights involv- for the emergence of a viable market in ing different firms has not developed un- discharge permits have not been estab- der the program (in spite of the efforts lished. of an active firm functioning as a broker In contrast, EPA's "interrefinery aver- in this market). This seems to be largely aging" program for the trading of lead the result of an extensive and compli- rights resulted in a very active market cated set of procedures for external over the relatively short life of the pro- trades that have introduced substantial gram. Begun in 1982, the program al- levels of transactions costs into the mar- lowed refiners to trade the severely lim- ket and have created uncertainties con- cerning the nature of the property rights 22 In an interesting analysis of the experience with that are being acquired. In addition, the Emissions Trading, Roger Raufer and Stephen Feld- man (1987)argue that some of the obstacles to trading program has been grafted onto an elabo- could be circumvented by allowing the leasing of rate set of. command-and-control style rights.. Cropper and Oates: Environmental Economics 691 ited rights to lead additives to gasoline. permits is thus a limited one with quite The program expired in 1986, although mTxed results. In the one case where the refiners were permitted to make a use market was allowed to function free of of rights that were "banked" through heavy restrictions, vigorous trading re- 1987. Trading became brisk under the sulted with apparently large cost savings. program: over the first half of 1987, for In contrast, under Emissions Trading example, around 50 percent of all lead and the Wisconsin TDP systems, strin- added to gasoline was obtained through gent restrictions on the markets for trad- trades of lead rights, with substantial cost ing emissions rights appear to have ef- savings reported from these trades. Al- fectively increased transaction costs though reliable estimates of cost-savings and introduced uncertainties, seriously for the lead-trading program are not impeding the ability of these markets to available, Hahn and Hester (198913) sur- realize the potentially large cost savings mise that these savings have run into the from trading. Even so, the cost savings hundreds of millions of dollars. As they from Emissions Trading (primarily from point out, the success of the program the Netting and Bubble provisions) have stemmed largely from the absence of a run into several billion dollars. Finally, large body of restrictions on trades: refin- it is interesting that these programs seem ers were essentially free to trade lead not to have had any significant and ad- rights and needed only to submit a quar- verse environmental effects; Hahn and terly report to EPA on their gasoline pro- Hester (1989a) suggest that their impact duction and lead usage. There were, on environmental quality has been moreover, already well established mar- roughly "neutral. " kets in refinery products (including a In light of this experience, the pros- wide variety of fuel additives) so that re- pects, we think, appear favorable for the finery managers had plenty of experience functioning of the new market in sulfur in these kinds of transaction^.^^ allowances that is being created under Finally, there is an emerging program the 1990 Amendments to the Clean Air in Colorado for the trading of rights to Act. This measure, designed to address phosphorous discharges into the Dillon the acid rain problem by cutting back Reservoir. This program is noteworthy annual sulfur emissions by 10 million in that among those that we have dis- tons, will permit affected power plants cussed, it is the only one to be designed to meet their emissions reduction quotas and introduced by a local government. by whatever means they wish, including The plan embodies few encumbrances to the purchase of "excess" emissions re- trading; the one major restriction is a ductions from other sources. The market 2 :1trading ratio for pointlnonpoint trad- area for this program is the nation as a ing, introduced as a "margin of safety" whole so that there should be a large because of uncertainties concerning the number of potential participants in the effectiveness of nonpoint source controls. market. At this juncture, plans for the The program is still in its early stages: structure and functioning of the market although no trades have been approved, do not appear to contain major limitations some have been requested. that would impede trading in the sulfur The U. S. experience with marketable allowances. There remains, however, the possibility that state governors or public utility commissions will introduce some '3 We should also note that various irregularities and illegal procedures were discovered in this mar- restrictions. There is the further concern ket-perhaps because of lax oversight. that regulated firms may not behave Journal of Economic Literature, Vol. XXX (June 1992) in a strictly cost-minimizing fashion, ence with systems of fees for pollution thereby compromising some of the cost- control, mainly of water pollution, these effectiveness properties of the trading systems have not, for the most part, been scheme. But as we suggested earlier, this designed in the spirit of economic incen- may not prove to be a serious distortion. tives for the regulation of water quality. The use of effluent fees is more preva- Their role has been more that of a reve- lent in Europe where they have been nue device to finance programs for water employed extensively in systems ofwater quality management. quality management and to a limited ex- These systems, it is worth noting, have tent for noise abatement (Ralph Johnson addressed almost exclusively so-called and Gardner Brown, Jr. 1976; Bower et "point-source" polluters. Non-point al. 1981; Brown and Hans Bressers 1986; source pollution (including agricultural Brown and Johnson 1984; Tietenberg and urban runoff into waterways) has 1990). There are few attempts to use proved much more difficult to encompass them for the control of air pollution. within systems of charges or permits. France, Germany, and the Netherlands, Winston Harrington, Krupnick, and for example, have imposed effluent fees Henry Peskin (1985) provide a useful on en~issionsof various water pollutants overview of the potential role for eco- for over two decades. It should be nomic incentives in the management of stressed that these fee systems are not non-point sources. This becomes largely pure systeins of economic incentives of a matter of seeking out potentially effec- the sort discussed in economics texts. tive second-best measures (e.g., fees on Their primary intent has not been the fertilizer use), since it is &ult to mea- regulation of discharges, but rather the sure and monitor "discharges" of pollu- raising of funds to finance projects for tants from these sources. Kathleen Seg- water quality management. As such, the erson (1988) has advanced an ingenious kes have typically been low and have proposal whereby such sources would be tended to apply tb "average" or "ex- subject to a tax (or subsidy payment) pected" discharges rather than to provide based, not on their emissions, but on the a clear cost signal at the margin. More- observed level of environmental quality; over, the charges are overlaid on an ex- although sources might find themselves tensive command-and-control system of with tax payments resulting from circum- regulations that mute somewhat further stances outside their control (e.g., ad- their effects as economic incentives. verse weather conditions), Segerson The Netherlands has one of the oldest shows that such a scheme can induce effi- and most effectively managed systems of cient abatement and entrylexit behavior charges-and also the one with relatively on the part of non-point sources. high levels of fees. There is some evi- D. Legal Liability as an Economic dence suggesting that these fees have, Instrument for Environmental in fact, had a measurable effect in reduc- Protection ing emissions. Some multiple regression work by Hans Bressers (1983) in the An entirely different approach to reg- Netherlands and surveys of industrial ulating sources is to rely on legal liability polluters and water board officials by for damages to the environment. Al- Brown and Bressers (1986) indicate that though we often do not include this ap- firms have responded to the charges with proach under the heading of economic significant cutbacks in discharges of wa- instruments, it is clear that a system of ter borne pollutants. "strict liability," under which a source In sum, although there is some experi- is financially responsible for damages, Cropper and Oates: Environmental Economics 693 embodies important economic incen- earlier, such payments to victims can re- tives."The imposition of such liability sult in inefficient levels of defensive ac- effectively places an "expected price" tivities. Strict liability thus does not get on polluting activities. The ongoing perfect marks on efficiency grounds, suits, for example, following upon the mas- even in principle, for although it internal- sive Exxon-Valdez oil spill suggest that izes the social costs of the polluter, it such penalties will surely exert pres- can be a source of distortions in victims' sures on potential polluters to engage in behavior. preventive measures. The more important concern, in prac- Under this approach, the environmen- tice, is the effectiveness of legal liability tal authority, in a setting of uncertainty, in disciplining polluter behavior. Even need not set the values of any price or if the basic rule is an efficient one in qiiantity instruments; it simply relies on terms of placing liability on the source the liability rule to discipline polluters. of the environmental damage, the actual Two issues are of interest here. he Arst "price" paid by the source may be much is the capacity, in principle, for strict lia- less than actual damages because of im- bility to mimic the effects of a Pigouvian perfections in the legal system: failures tax. And the second is the likely effective- to impose liability on responsible parties ness, in practice, of strict liability as a resulting from uncertainty over causa- substitute for other forms of economic tion, statutes of limitation, or high costs incentives. There is a substantial litera- of prosecution.25There is the further pos- ture in the economics of the law that ad- sibility of bankruptcy as a means of avoid- dresses these general issues and a grow- inglarge payments for damages. The evi- ing number of studies that explore this dence on these matters is mixed (see matter in the context of environmental Segerson 1990), but it seems to suggest management (see, for example, Steven that legal liability has functioned only Shave11 1984a, 198413; Segerson 1990). very imperfectly. It is clear that strict liability can, in An interesting area of application in principle, provide the source of potential the environmental arena involves various damages with the same incentive as a pieces of legislation that provide strict Pigouvian tax. If a polluter knows that liability for damages from accidental he will be held financially accountable spills of oil or leakage of hazardous for any damages his activities create, then wastes. The Comprehensive Environ- he will have the proper incentive to seek mental Responses, Compensation, and methods to avoid these damages. Strict Liability Act (CERCLA) of 1980 and its liability serves to internalize the external later amendments (popularly known as costs-just as does an appropriate tax. "Superfund") are noteworthy for their Strict liability is unlike a tax, however, broad potential applicability (Thomas in that it provides compensation to vic- Grigalunas and James Opaluch 1988). tirns. The Pigouvian tax possesses an im- Such measures may well provide a useful portant asymmetry in a market sense: it framework for internalizing the external is a charge to the polluter-but not a payment to the victim. And, as noted =AS one reviewer noted, in these times of height- ened environmentalBensitivity, liability determina- tions could easily exceed actual damages in some 24 The major alternative to strict liability is a negli- instances. However, this seems not to have happened gence rule under which a polluter is liable only if in the recent Exxon-Valdez case. The case was settled he has failed to comply with a "due standard of care" out of court with Exxon agreeing to pay some $900 in the activity that caused the damages. Under strict million over a period of several years. Some observers liability, the party causing the damages is liable irre- believe that this falls well short of the true damages spective of the care exercised in the polluting activity. from the Exxon-Valdez oil spill in Alaska. 694 Journal of Economic Literature, Vol. XXX (June 1992) costs of spills (Opaluch and Grigalunas tions, the optimal level of effluent fees 1984). In particular, the liability ap- (or quantities of marketable permits) will proach appears to have its greatest appeal also vary (Sam Peltzman and T. Nicolaus in cases like those under Superfund Tideman 1972). The first-best outcome where damages are infrequent events must therefore be one in which fees or and for which monitoring the level of care quantities of permits are set in accord a firm takes under conventional regula- with local circumstances, suggesting that tory procedures would be diffic~lt.~~ an optimal regulatory system for pollu- tion control will be a form of environmen- E . Environmental Federalism tal federalism. In addition to the choice of policy in- Some environmental economists have strument, there is the important issue raised an objection to this general pre- of the locus of regulatory authority. In sumption. John Cumberland (1981), the case of fees, for example, should a among others, has expressed the concern central environmental authority establish that in their eagerness to attract new a uniform fee applicable to polluters in business and jobs, state or local officials all parts of the nation or should decentral- will tend to set excessively lax environ- ized agencies set fee levels appropriate mental standards-fees that are too low to their own jurisdictions? U. S. environ- or quantities of permits that are too high. mental policy exhibits considerable am- The fear is that competition among de- bivalence on this matter. Under the centralized jurisdictions for jobs and in- Clean Air Act in 1970, the U. S. Congress come will lead to excessive environmen- instructed the Environmental Protection tal degradation. This, incidentally, is a Agency to set uniform national standards line of argument that has appeared else- for air quality-maximum permissible where in the literature on fiscal federal- concentrations of key air pollutants appli- ism under the title of "tax competition." cable to all areas in the country. But two The difficulty in assessing this objection years later under the Clean Water Act, to decentralized policy making is that the Congress decided to let the individ- there exists little systematic evidence on ual states determine their own standards the issue; most of the evidence is anec- (subject to EPA approval) for water qual- dotal in character, and, until quite re- ity. The basic question is "Which ap- cently, there has been little theoretical proach, centralized decision making or work addressing the phenomenon of in- environmental federalism, is the more terjurisdictional omp petition.^' promising?" In a pair of recent papers, Oates and Basic economic principles seem to sug- Robert Schwab (1988a, 1988b) have set gest, on first glance, a straightforward an- forth a model of such competition in swer to this question. Since the benefits which "local" jurisdictions compete for and costs of reduced levels of most forms a mobile national stock of capital using of pollution are likely to vary (and vary both tax and environmental policy instru- substantially) across different jurisdic- ments. Since the production functions

"A more complicated and problematic issue re- 27 TWOrecent studies, one by Virginia McConnell lates to the permission of the courts to sue under and Schwab (1990), and the other by Timothy Bartik Superfund for damages from toxic substances using (198&), find little evidence of strong effects of exist- "the joint and several liability doctrine." Under this ing environmental regulations on the location deci- provision, each defendant is potentially liable for an sions of firms within the U.S. This, of course, does amount up to the entire damage, irrespective of his not preclude the possibility that state and local offi- individual contribution. For an analysis of this doc- cials, in fear of such effects, will scale down standards trine in the Superfund setting, see Tietenberg (1989). for environmental quality. Cropper and Oates: Environmental Economics 695 are neoclassical in character, an increase not immune to various "imperfections. " in a jurisdiction's capital stock raises the If, for example, local governments are level of wages through an associated in- constrained constitutionally to use taxes crease in the capital-labor ratio. In the on capital to finance various local public model, local officials simultaneously em- goods, then it is easy to show that not ploy two policy tools to attract capital: a only will the tax rate on capital be posi- tax rate on capital itself which can be tive, but officials will select socially ex- lowered or even set negative (a subsidy) cessive levels of pollution. Likewise, if to raise the return to capital in the juris- Niskanen bureaucrats run the local pub- diction, and a level of allowable pollutant lic sector, they will choose excessively emissions (or, alternatively, an effluent lax environmental standards as a mecha- fee). By increasing the level of permissi- nism to attract capital so as to expand ble waste discharges either directly or the local tax base and public revenues. by lowering the fee on emissions, the Finally, there can easily be conflicts local authority increases the marginal among local groups of residents with dif- product of capital and thereby encour- fering interests (e.g., workers vs. non- ages a further inflow of capital. The workers) that can lead to distorted out- model thus involves two straightforward comes (although these distortions may tradeoffs: one between wage income and involve too little or too much pollution). tax revenues, and the other between The basic model does at least suggest wage income and local environmental that there are some fundamental forces quality. The analysis reveals that in a set- promoting efficient decentralized envi- ting of homogeneous worker-residents ronmental decisions. If the regions se- making choices by simple majority rule, lected for environmental decision making jurisdictions select the socially optimal are sufficiently large to internalize the levels of these two policy instruments. polluting effects of waste discharges, the The tax rate on capital is set equal to zero, case for environmental federalism has and the level of environmental quality some force. Exploration of this issue is is chosen so that the willingness to pay admittedly in its infancy-in particular, for a cleaner environment is equal to there is a pressing need for some sys- marginal abatement cost. The analysis tematic empirical study of the effects of thus supports the case for environmental "local" competition on environmental federalism: decentralized policy making choices.29 is efficient in the F. Enforcement Issues In one sense, this is hardly a surprising result. Since local residents care about The great bulk of the literature on the level of environmental quality, we the economics of environmental regula- should not expect that they would wish tion simply assumes that polluters com- to push levels of pollution into the range ply with existing directives: they either where the willingness to pay to avoid en- keep their discharges within the pre- vironmental damage exceeds the loss in scribed limitation or, under a fee wage income from a cleaner environ- scheme, report accurately their levels of ment. At the same time, this result is emissions and pay the required fees.

'%sing an alternative analytical framework in '"or some other recent theoretical studies of in- which local jurisdictions "bid" against one another terjurisdictional fiscal competition, see Jack Mintz for polluting firms in terms of entry fees, William and Henry Tulkens (1986), John Wilson (1986), David Fischel (1975) likewise finds that local competition Wildasin (1989), and George Zodrow and Peter Mi- produces an efficient outcome. eszkowski (1986). 696 Journal of Economic Literature, Vol. XXX (June 1992)

Sources, in short, are assumed both to charges is independent both of the level act in good faith and to have full control of the fine for underreporting and of the over their levels of discharges so that vio- probability of punishment (so long as the lations of prescribed behavior do not oc- slope of the expected penalty function cur. with respect to the size of the violation Taking its lead from the seminal paper is increasing and the probability of pun- by Gary Becker (1968) on the economics ishment is greater than zero). The pollu- of crime and punishment, a recent litera- ter sets the level of actual wastes such ture has addressed enforcement issues that marginal abatement cost equals the as they apply to environmental reg- effluent fee-the efficient level! But he ulation~.~~As this literature points out, then, in general, underreports his dis- violations of environmental regulations charges with the extent of underreport- can have two sources: a polluter can will- ing varying inversely with the level of fully exceed his discharge limitation (or fines and the probability of punishment. under-report his emissions under a fee Arun Malik (1990) has extended this system) to reduce compliance costs or a line of analysis to the functioning of sys- stochastic dimension to discharges may tems of marketable permits. He estab- exist so that the polluter has only imper- lishes a result analogous to Harford's: un- fect control over his levels of emissions. der certain circumstances, noncompliant In such a setting, the regulatory problern polluters will ernit precisely the same becomes a more complicated one. Not level of wastes for a given permit price only must the regulatory agency set the as that discharged by an otherwise identi- usual policy parameters (emissions lirni- cal compliant firm. The conditions, how- tations or fees), but it must also decide ever, for this equivalence are fairly strin- upon an enforcement policy which in- gent ones. More generally, Malik shows volves both monitoring procedures and that noncompliant behavior will have ef- levels of fines for violations. fects on the market-clearing price in the The early literature explored these en- permit market-effects that will compro- forcement issues in a wholly static frame- mise to some extent the efficiency prop- work. The seminal papers, for example, erties of the marketable permit system. by Paul Downing and William Watson One implication of this body of work (1974) and by Jon Harford (1978), estab- is the expectation of widespread noncom- lished a number of interesting results. pliance on the part of polluters. But as Downing and Watson show that the in- Harrington (1988) points out, this seems corporation of enforcement costs into the not to be the case. The evidence we have analysis of environmental policy suggests from various spot checks by EPA and that optimal levels of pollution control GAO suggests that most industrial pollu- will be less than when these costs are ters seem to be in compliance most of ignored. Harford obtains the especially the time.31Substantial compliance seems interesting result that under a system of effluent fees, the level of actual dis- 31 Interestingly, noncompliance seems to be more widespread among municipal waste treatment plants than among industrial sources! (Russell 1990, p. 256). 30 Russell, Harrington, and William Vaughan Some of the most formidable enforcement problems (1986, ch. 4) provide a useful survey of the enforce- involve federal agencies. The GAO (1988),for exam- ment literature in environmental economics up to ple, has found the Department of Energy's nuclear 1985. Harrington (1988)presents a concise, excellent weapons facilities to be a source of major concern; overview both of the more recent literature and of the costs of dealing with environmental contarnina- the "stylized facts" of actual compliance and enforce- tion associated with these facilities are estimated at ment behavior. See also Russell (1990). more than $100 billion. Cropper and Oates: Environmental Economics 697

to exist in spite of modest enforcement lar reduction in the minimum resources efforts: relatively few "notices of viola- required to achieve a given level of com- tion" have been issued and far fewer pol- pliance" (p. 47). In sum, the dynamic luters have actually been fined for their game-theoretic approach can produce violations. Moreover, where such fines compliance in cases in which the ex- have been levied, they have typically pected penalty is insufficient to pre- been quite small. And yet in spite of such vent violations in a purely static model. modest enforcement efforts, "cheating" Moreover, it suggests some potentially is not ubiquitous-violations are cer- valuable guidelines for the design of tainly not infrequent, but they are far cost-effective enforcement procedures. from universal. Enforcement is an area where economic This finding simply doesn't square at analysis may make some quite useful con- all well with the results from the static tributions. models of polluter behavior.32 An alter- G. The Eflects of Domestic native line of modeling (drawing on the Environmental Policy on Patterns of tax-evasion literature) seems to provide International Trade a better description of polluter behavior; it also has some potentially instructive The introduction of policy measures normative implications. This approach to protect the environment has potential puts the problem in a dynamic game- implications not only for the domestic theoretic framework. Both polluters and economy but also for international trade. regulators react to the activities of one Proposed environmental regulations are, another in the previous period. In a pro- in fact, often opposed vigorously on the vocative paper, Harrington (1988) mod- grounds that they will impair the "inter- els the enforcement process as a Markov national competitiveness" of domestic in- decision problem. Polluters that are de- dustries. The increased costs associated tected in violation in one period are with pollution control measures will, so moved to a separate group in the next the argument goes, result in a loss of ex- period in which they are subject to more port markets and increased imports of frequent inspection and higher fines. products of polluting industries. Polluting firms thus have an incentive These potential effects have been the to comply in order to avoid being moved subject of some study. It is clear, for ex- into the second group (from which they ample, that the adoption of costly control can return to the original group only after measures in certain countries will, in a period during which no violations are principle, alter the international struc- detected). In such a framework, firms ture of relative costs with potential effects may be in compliance even though they on patterns of specialization and world would be subject to no fine for a viola- trade. These trade effects have been ex- tion. Following up on Russell's analysis plored in some detail, making use of stan- (Russell, Harrington, and Vaughan 1986, dard models of international trade (Ka- pp. 199-216), Harrington finds that the zumi Asako 1979; Baumol and Oates addition of yet a third group, an absorb- 1988, ch. 16; Anthony Koo 1974; Martin ing state from which the polluter can McGuire 1982; John Merrifield 1988; Rii- never emerge, can result in a "spectacu- diger Pethig 1976; Pethig et al. 1980; Horst Siebert 1974; James Tobey 1989; Ingo Walter 1975). In particular, there 32 Perhaps public opprobrium is a stronger discipli- nary force than economists are typically inclined to has been a concern that the less devel- believe! oped countries, with their emphasis on 698 Journal of Economic Literature, Vol. XXX (June 1992)

economic development rather than envi- same issue in a large econometric study ronmental protection, will tend over of international trade patterns in "pollu- time to develop a comparative advantage tion-intensive" goods. After controlling in pollution-intensive industries. In con- for the effects of relative factor abun- sequence, they will become the "havens" dance and other trade determinants, To- for the world's dirty industries; this con- bey cannot find any effects of various cern has become known as the "pollu- measures of the stringency of domestic tion-haven hypothesis" (Walter and Ju- environmental policies. Tobey estimates dith Ugelow 1979; Walter 1982). two sets of equations that explain, respec- Some early studies made use of exist- tively, patterns of trade in pollution- ing macro-econometric models to assess intensive goods and changes in trade pat- the likely magnitudes of these effects. terns from 1970 to 1984. In neither set These studies used estimates of the costs of equations do the variables measuring of pollution control programs on an in- the stringency of dornestic environmen- dustry basis to get some sense of the ef- tal policy have the predicted effect on fects of these programs on trade and pay- trade patterns. ments flows. Generally, they found Why have dornestic environmental small, but measurable, effects (d'Arge measures not induced "industrial flight;" and Kneese 1971; Walter 1974). and the development of "pollution ha- We are now in a position to examine vens?" The primary reason seems to be historically what has, in fact, happened. that the costs of pollution control have To what extent have environmental mea- not, in fact, loomed very large even in sures influenced the pattern of world heavily polluting industries. Existing es- trade? Have the LDC's become the ha- timates suggest that control costs have vens of the world's dirty industries? Two run on the order of only 1to 2% percent recent studies, quite different in charac- of total costs in most pollution-intensive ter, have addressed this issue directly. industries; H. David Robison (1985, p. H. Jeffrey Leonard (1988), in what is 704), for example, reports that total largely a case study of trade and foreign- abatement costs per dollar of output in investment flows for several key indus- 1977 were well under 3 percent in all tries and countries, finds little evidence industries with the sole exception of elec- that pollution-control measures have ex- tric utilities where they were 5.4 per- erted a systematic effect on international cent. Such small increments to costs are trade and investment. After examining likely to be swamped in their impact on some aggregate figures, the policy international trade by the much larger stances in several industrialized and de- effects of changing differentials in labor veloping countries, and the operations costs, swings in exchange rates, etc. of multinational corporations, Leonard Moreover, nearly all the industrialized concludes that "the differentials in the countries have introduced environmental costs of complying with environmental measures-and at roughly the same regulations and in the levels of environ- time-so that such measures have not mental concern in industrialized and in- been the source of significant cost differ- dustrializing countries have not been entials among major competitors. There strong enough to offset larger political seems not to have been a discernible and economic forces in shaping aggregate movement in investment in these indus- international comparative advantage" (p. tries to the developing countries because 23 1). major political and economic uncertain- Tobey (1989, 1990) has looked at the ties have apparently loorned much larger Cropper and Oates: Environmental Economics 699

in location decisions than have the mod- fashion "CAC" methods of regulatory est savings from less stringent environ- control and to contrast them as a class mental controls. with the least-cost outcornes typically as- In short, domestic environmental poli- sociated with systems of economic incen- cies, at least to this point in time, do tives. In fact, the compromises and "irn- not appear to have had significant effects perfections" inherent in the design and on patterns of international trade. From implementation of incentive-based sys- an environmental perspective, this is a tems virtually guarantee that they also comforting finding, for it means that will be unable to realize the formal least- there is little force to the argument that cost result. we need to relax environmental policies Empirical studies contrasting the cost to preserve international competitive- effectiveness of the two general ap- ness. proaches have typically examined the cost under each system of attaining a H. Command-and-Control us. specified standard of environmental Economic Incentives: Some quality-which typically means ensuring Concluding Observations that at no point in an area do pollutant Much of the literature in environ- concentrations exceed the maximum mental economics, both theoretical and level permissible under the particular empirical, contrasts in quite sharp and standard. As Atkinson and Tietenberg uncompromising terms the properties of (1982) and others have noted, CAC sys- systerns of economic incentives with the tems typically result in substantial "over- inferior outcornes under existing systerns control" relative to incentive-based sys- of command-and-control regulations. In tems. Since it effectively assigns a zero certain respects, this literature has been shadow price to any environmental im- a bit misleading and, perhaps, unfair. provements over and above the standard, The term command-and-control encom- the least-cost algorithm attempts to make passes a very broad and diverse set of use of any "excess" environmental capac- regulatory techniques-some admittedly ity to increase emissions and thereby re- quite crude and excessively costly. But duce control costs. The less cost-sensitive others are far more sophisticated and cost CAC approaches generally overly restrict sensitive. In fact, the dividing line be- emissions (relative to the least-cost solu- tween so-called CAC and incentive- tion) and thereby produce pollutant con- based policies is not always so clear. A centrations at nonbinding points that are program under which the regulator spec- less than those under the least-cost out- ifies the exact treatment procedures to come. In sum, at most points in the area, be followed by polluters obviously falls environmental quality (although subject within the CAC class. But what about a to the same overall standard) will be policy that establishes a fixed emissions higher under a CAC system than under limitation for a particular source (with no the least-cost solution. So long as there trading possible) but allows the polluter is some value to improved environmental to select the form of compliance? Such quality beyond the standard, a proper flexibility certainly allows the operation comparison of benefits and costs should of economic incentives in terms of the give the CAC system credit for this incre- search for the least-cost method of con- ment to environmental quality. One re- trol. cent study (Oates, Paul Portney, and The point here is that it can be quite McGartland 1989) which does just this misleading to lump together in a cavalier for a major air pollutailt finds that a rela- 700 Journal of Economic Literature, Vol. XXX Wzme 1992)

tively sophisticated CAC approach pro- so at the lowest cost. A general realiza- duces results that compare reasonably tion of this point seems to be emerging well to the prospective outcome under with a consequent renewed interest in a fully cost effective system of economic many countries in the possibility of inte- incentives. grating incentive-based policies into en- Our intent is not to suggest that the vironmental regulations-a matter to economist's emphasis on systems of eco- which we shall return in the concluding nomic incentives has been misplaced, section. but rather to argue that policy structure and analysis is a good deal more compli- IV. Measuring the Benejits and Costs cated than the usual textbooks would sug- of Pollution Control gest (Nichols 1984). The applicability of systems of economic incentives is to some As we suggested in the previous sec- extent limited by monitoring capabilities tions, effluent fees and transferable per- and spatial complications. In fact, in any mits are capable, in principle, of achiev- meaningful sense the "optimal" structure ing a given pollution standard at least of regulatory programs for the control of cost. Eventually, however, economists air and water pollution is going to involve must ask whether environmental stan- a combination of policy instruments- dards have been set at appropriate levels: some making use of economic incentives does the marginal cost of achieving the and others not. Careful economic analy- ozone standard in the Los Angeles basin sis has, we believe, an important role exceed the marginal benefits? The an- to play in understanding the workings swer to this question requires that we of these systems. But it can make its best measure the benefits and costs of pollu- contribution, not through a dogmatic tion control. commitment to economic incentives, but While the measurement of control rather by the careful analysis of the whole costs is itself no simple task, environmen- range of policy instruments available, in- tal economists have turned most of their suring that those CAC measures that are attention to the benefit side of the ledger. adopted are effective devices for control- Of central concern has been the develop- ling pollution at relatively modest cost ment of methodologies to measure the (Kolstad 1986). benefits of goods-such as clean air or At the same time, it is our sense that water-that are not sold in markets. incentive-based systems have much to These techniques fall into two categories: contribute to environmental protec- indirect market methods, which attempt tion-and that they have been much ne- to infer from actual choices, such as glected in part because of the (under- choosing where to live, the value people standable) predisposition of regulators to place on environmental goods; and direct more traditional policy instrument^.^^ questioning approaches, which ask peo- There are strong reasons for believing, ple to make tradeoffs between environ- with supporting evidence, that this ne- mental and other goods in a survey con- glect has seriously impaired our efforts text. We shall review both approaches, both to realize our objectives for im- and then discuss the application of these proved environmental quality and to do methods to valuing the benefits of pollu- tion control. In particular, we will try '"See Steven Kelman (1981) for a fascinating-if to highlight areas where benefits have solnewhat dismaying-study of the politics and ideol- ogy of economic incentives for environmental protec- been successfully measured, as well as tion. areas where good benefit estimates are Cropper and Oates: Enzjironmental Economics 701

most needed. But first we must be clear For the case of a firm, the value of a about the valuation of changes in envi- change in Q (or S) is the change in the ronmental quality. firm's profits when Q (or S) is altered. This amount is the same whether we are A. De5ning the Value of a Change in talking about the firm's willingness to pay Enzjironmental Quality (WTP) for an improvement in environ- We noted at the beginning of this re- mental quality or its willingness to accept view that pollution may enter both con- (WTA) compensation for a reduction in sumers' utility functions and firms' pro- environmental quality. duction functions. (See equations (1) and For a consumer, in contrast, the value (2).)To elaborate on how this might occur of a change in Q (or S) depends on the we introduce a damage function that initial assignment of property rights. If links pollution, Q, to something people consumers are viewed as having to pay value, S, for an improvement in environmental quality, for example, from Q' to Q1, the most they should be willing to pay for For a consumer, S might be time spent this change is the reduction in expendi- ill or expected fish catch; for a firm it ture necessary to achieve their original might be an input into production, such utility level when Q improves. Formally, as the stock of halibut. We assume that if e(P,S(QU),0') denotes the minimum ex- S replaces Q in the utility and production penditure necessary to achieve pre-im- functions (equations (1) and (2)). provement utility UO at prices P and envi- There are two cases of interest here. ronmental quality QO, then the most First, if the consumer (or firm) views S people would be willing to pay (WTP) as out of his control, we can define the for the improvement in environmental value of a change in S (which may be quality to Q1 is easier to measure than the value of a change in Q), and then predict the WTP = ~(P,s(Q'),@) change- in S resulting- from a change- in - e(P,S(Q1), (10) Q. For example, if people view reduc- If, on the other hand, consumers are tions in visibility associated with air pol- viewed as having rights to the higher lution as beyond their control, one can level of environmental quality and must predict the reduction in visibility from be compensated for a reduction in Q, (8) and concentrate on valuing visibility. then the smallest amount they would be This is commonly known as the damage willing to accept is the additional amount function approach to benefit estimation. they must spend to achieve their original The second case is more complicated. utility level when Q declines. Formally, It may sometimes be possible to mitigate willingness to accept (WTA) compensa- the effects of pollution through the use tion for a reduction in Q from Q' to QO of inputs, 2. For example, medicine may is given by exist to alleviate respiratory symptoms associated with air pollution. In this in- WTA = e(p,~(~~),~l) stance, equation (8)must be modified to - e(P,S(Q1),U1), (11) S = S(Q,Z), ") where U' is the utility level achieved at and it is Q rather than S that must be the higher level of environmental qual- valued, because S is no longer exoge- ity. nous. In general, willingness to accept corn- 702 Journal of Economic Literature, Vol. XXX (June 1992)

pensation for a reduction in Q will be ture, the difference between the two wel- higher than willingness to pay for an in- fare measures will be small. Moreover, crease in Q of the same magnitude. As approximating WTP or WTA by con- W. Michael Hanemann (1991) has re- sumer surplus-the area to the left of cently shown, the amount by which WTA the Marshallian demand function will exceeds WTP varies directly with the in- produce an error of no more than 5 per- come elasticity of demand for S and in- cent in most cases (Robert Willig 1976).34 versely with the elasticity of substitution One problem with the definitions of between S and private goods. If the in- the value of a change in environmental come elasticity of demand for S is zero quality (equations (10) and (11)) is that or if S is a perfect substitute for a private not all environmental benefits can be good, WTP should equal WTA. If, how- viewed as certain. Reducing exposure to ever, the elasticity of substitution be- a carcinogen, for example, alters the tween S and private goods is zero, the probability that persons in the exposed difference between WTA and WTP can population will contract cancer, and it be infinite. It is therefore important to is this probability that must be valued. determine which valuation concept, To define the value of a quality change WTP or WTA, is appropriate for the under uncertainty, suppose that the problem at hand. value of S associated with a given Q is The preceding definitions of the value uncertain. Specifically, suppose that two of a change in environmental quality do values of S are possible: SO and S1. For not by themselves characterize all of the example, SO might be 360 healthy days welfare effects of environmental policies. per year and S1 no healthy days (death). Improvements in environmental quality Q no longer determines S directly, but may alter prices as well as air or water affects IT, the probability that So occurs. quality, and these price changes must be If the individual is an expected utility valued in addition to quality changes. maximizer and if V(M,Si), i = 0,1, de- In contrast to valuing quality changes, notes his expected utility in each state valuing price changes is relatively (M being income), willingness to pay straightforward. WTP for a reduction in for a change in Q from QO to Q' is the price is just the reduction in expenditure most one can take away from the indi- necessary to achieve UO (the consumer's vidual and leave him at his original ex- original utility level) when prices are re- pected utility level (Michael Jones-Lee duced. As is well known, this is just the 1974). area to the left of the relevant compen- sated demand function (i.e., the one that holds utility at UO) between the two prices. Willingness to accept compensa- tion for a price increase is the increase For a small change in Q, WTP is just in expenditure necessary to achieve ul, the difference in utility between the two the utility level enjoyed at the lower states, divided by the expected marginal price, when price is increased. utility of money, Unlike the case of a quality change,- WTA compensation for a price increase exceeds WTP for a mice decrease onlv 34 Sufficient conditions for this to hold are that (1) by the amount of income effect. consumer surplus is no more than 90 percent of i'n- come; (2) the ratio of consumer surplus to income, long as On the good in ques- multiplied by one-half the income elasticity of de- tion is a small fraction of total expendi- mand, is no more than 0.05. Cropper and Oates: Environmental Economics 703

[v(M,sO)- v(M,S1)l the value of a small change in pollution WTP = can be measured by the value of the in- nqf+ (1- 54q4 puts used to compensate for the change a.rr .-d~.(13) in pollution. If, for example, a reduction aQ in one-hour maximum ozone levels from An important point to note here is that 0.16 parts per million (ppm) to 0.11 ppm the value of the change in Q is an ex reduces the number of days of respiratory ante value: changes in Q are valued be- symptoms from 6 to 5, and if an expendi- fore the outcomes are known. For exam- ture on medication of $20 has the same ple, suppose that reducing exposure to effect, then the value of the ozone reduc- an environmental carcinogen is expected tion is $20. to save two lives in a city of 1,000,000 Somewhat more formally, if S = persons. The ex ante approach views this S(Q,Z),willingness to pay for a marginal as a 2-in-one-million reduction in the change in Q may be written as the mar- probability of death for each person in ginal rate of substitution between an the population. The ex post approach, averting good and pollution, times the by contrast, would value the reduction price of the averting good (Paul Courant in two lives with certainty. and Richard Porter 1981). We are now in a position to discuss as/aQ the principal methods that have been WTP = -p,- (14) used to value changes in pollution. dS/dzl' B. Indirect Methods for Measuring the where zl is medication. Marginal WTP Benefits of Environmental Quality can thus be estimated from the produc- tion function alone. Economists have employed three ap- To value a nonmarginal change in pol- proaches to valuing pollution that rely lution, one must know both the cost func- on observed choices: the averting behav- tion for the good affected by pollution ior approach, the weak complementarity and the marginal value function for that approach, and the hedonic price ap- good. For example, in the case of health proach. damages, a large improvement in air 1. The Averting Behavior Approach. quality will shift the marginal cost of The averting behavior approach relies on healthy days to the right (see Figure 1) the fact that in some cases purchased in- and the value of the change is given by puts can be used to mitigate the effects the area between the two marginal cost of pollution. 35 For example, farmers can curves, bounded by the marginal value increase the amount of land and other of healthy time. When the good in ques- inputs to compensate for the fact that tion is not sold in markets, as is the case ozone reduces soybean yields. Or, for an- for health, estimating the marginal value other, residents of smoggy areas can take function is, however, difficult. 36 medicine to relieve itchy eyes and runny noses. 36 If S were sold in markets, estimation of the mar- As long as other inputs can be used ginal value function would be simple, assuming one to compensate for the effects of pollution, could observe the price of S and assuming that the price was exogenous to any household The problem is that, for a good produced by the household itself, one cannot observe the price (marginal co5t) of the 35 In terms of the notation above, either (9)applies, good-it must be estimated from the marginal cost or other inputs can he <~~hstitutedfor S in production; function. Furthermore, the price is endogenou5, see equation (2). since it depends on the level of S. 704 Journal of Economic Literature, Vol. XXX uune 1992)

pollution in the first place, or by mitigat- ing the effects of exposure once they oc- cur. For example, the deleterious effects of water pollution can be avoided by pur- chasing bottled water (V. Kerry Smith and William Desvoi~sges1986b), and pol- lutants in outdoor air may be filtered by running an air-conditioner (Mark Dickie A and Shelby Gerking 1991). I of Healthy Time I Two problems, however, arise in ap- 1 I plying the averting behavior method in Healthy Time these cases. First, in computing the Figure 1. Morbidity Benefits of a Non~~larginal right-hand-side of (14), the researcher Pollution Reduction must know what the household imagined the benefit of purchasing water asl la^,) An alternative approach, suggested by to be, since it is the perceived benefits Bartik (1988a), is to use the change in of averting behavior that the household the cost of producing the original level equates to the marginal cost of this be- of S, i.e., the area between the marginal havior. Second, when the averting input cost functions to the left of SO (area ABD produces joint products, as in the case in Figure l), to approximate the value of running an air-conditioner, the cost of the environmental quality change. For of the activity cannot be attributed solely an improvement in Q, this understates to averting behavior. Inputs that mitigate the value of the change because it does the effects of pollution include medicine not allow the individual to increase his and doctors' visits (Gerking and Linda chosen value of S. When the marginal Stanley 1986); however, use of the latter cost of S increases, the relevant area will often runs into the joint product prob- overstate the value of the welfare de- lem-a doctor's visit may treat ailments crease. The advantage of this approxima- unrelated to pollution, as well as pollu- tion is that it can be estimated from tion related illness. knowledge of the cost function alone. 2. The Weak Complementarity Ap- The usefulness of the averting behavior proach. While the averting behavior ap- approach is clearly limited to cases where proach exploits the substitutability be- other inputs can be substituted for pollu- tween pollution and other inputs into tion. Most pollution damages suffered by production, the weak complementarity firms occur in agriculture, forestry, and approach values changes in environmen- fishing. In the case of agriculture, irriga- tal quality by making use of the comple- tion can compensate for the effects of mentarity of environmental quality, e.g., global warming on crop yields. Likewise, cleaner water, with a purchased good, capital (boats and gear) and labor can e.g., visits to a lake. Suppose that a speci- compensate for fish populations depleted fied improvement in water quality at a as a result of water pollution. lake resort results in an increase in a In the case of pollution damages suf- household's demand for visits to the re- fered by households, averting behavior sort from ED to AB (see Figure 2). One has been used to value health damages can view the value of access to the lake and the soiling damages caused by air at the original quality level QO as the va- pollution. Households can avoid health lue of being able to visit the lake at a damages either by avoiding exposure to cost of C rather than at some cost E. Cropper and Oates: Environmental Economics 705

bute, such as fish catch.38 Although site visits do not have a market price, their cost can be measured by summing the cost of traveling to the site, including the time cost, as well as any entrance fees. A problem in measuring the demand for site visits as a function of site quality is that there is no variation in site quality among persons who visit a site. A popular solution to this problem is the varying 1 parameters model, which assumes that Site Visits site quality enters recreation demand Figure 2. The Effect of a Change in functions multiplied by travel cost or in- Environmental Quality on the Demand for Visits to a Recreation Site come, both of which vary across ho~seholds.~~In the first stage of the model, the demand for visits to site i is The value of access to the lake is thus regressed on the cost of visiting the site the area EDC.~~The increase in the and on income. In the second stage the value of access when Q changes (area coefficients from stage one are regressed ABDE) is the value of the water quality on quality variables at site i. This is improvement. equivalent to estimating a set of demand For area ABDE to measure the value functions in which visits to site i depend of the water quality improvement, envi- on the quality of the ith site, the cost of ronmental quality must be weakly com- visiting the ith site, income, and interac- plementary to the good in question tions between travel cost and quality, and (Maler 1974; Nancy Bockstael and Ken- income and quality. neth McConnell 1983). This means that One drawback of this approach is that (1)the marginal utility of environmental it allows visits to a given site to depend quality (water quality) must be zero if only on the cost of visiting that site- none of the good is purchased (no visits the cost of visiting substitute sites is not are made to the lake); (2) there is a price considered. This is equivalent to assum- above which none of the good is pur- ing that, except for the quality variables chased (no visits are made). If (1)did not that enter the model in stage two, all hold, three would be additional benefits sites are perfect substitutes. The varying to a change in water quality not reflected parameters model may, therefore, give in the demand for visits. misleading results if one wishes to value In practice, the weak complementarity quality changes at several sites. approach has been used most often to A second approach to valuing quality value the attributes of recreation sites- changes is to use a discrete choice model. either water quality, or a related attri- This approach examines the choice of

38 Surveys of recreation demand models may be 37 Strictly speaking EDC should be measured using found in Mendelsohn (1987) and also in John Braden the consumer's compensated demand function. and Kolstad (1991). Bockstael, Hanemann, and Cath- When measuring the value of access to a good, use erine Kling (1987) discuss their application to valuing of the Marshallian demand function may no longer environmental quality at recreation sites. provide a good approximation to the welfare triangle 39 This solution was first used by Vaughan and Rus- since the choke prices of the Marshallian and corn- sell (1982) and has also been used by V. Kerry Smith, pensated demand functions may vary substantially. Desvousges, and Matthew McGivney (1983), and V. The Willig bounds do not apply in this case. Kerry Smith and Desvousges (1986a). 706 Journal of Economic Literature, Vol. XXX (June 1992)

which site to visit on a given day as a third method used by economists to function of the cost of visiting each site, value environmental quality, or a related and the quality of each site. If the choice output such as mortality risk, exploits the of which site to visit on the first recre- concept of hedonic prices-the notion ation day can be viewed as independent that the price of a house or job can be of which site to visit on the ith, a simple decomposed into the prices of the attri- discrete choice model, such as the mul- butes that make up the good, such as tinomial logit, can be applied to the air quality in the case of a house (Ronald choice of site, conditional on partici- Ridker and John Henning 1967), or risk pation (Clark Binkley and Hanemann of death in the case of a job (Richard 1978; Daniel Feenberg and hlills 1980). Thaler and Sherwin Rosen 1976). The he- The choice of whether to participate donic price approach has been used pri- and, if so, on how many days, is made marily to value environmental disameni- by comparing the maximum utility re- ties in urban areas (air pollution, ceived from taking a trip with the utility proximity to hazardous waste sites), of the best substitute activity on that which are reflected both in housing day.40 prices and in wages. It has also been used The advantage of the discrete choice to value mortality risks by examining the model is that the probability of visiting compensation workers receive for volun- any one site depends on the costs of visit- tarily assuming job risks. Finally, the he- ing all sites and the levels of quality at donic travel cost approach has been used all sites. The drawback of the model is to value recreation sites. We discuss each that the decision to take a trip or not approach in turn. and, if so, which site to visit, is made independently on each day of the season. Urban Amenities. Air quality and other The number of trips made to date influ- environmental amenities can be valued ence neither which site the individual in an urban setting by virtue of being chooses to go to on a given day, nor tied to residential location: they are part whether he takes a trip at all.41 Thus, of the bundle of amenities-public these models must be combined with schools, police protection, proximity to models that predict the total number of parks-that a household purchases when trips taken. buying a house. 3. Hedonic Market Methods. The The essence of the hedonic approach is to try to decompose the price of a house

40 If one estimates a discrete choice model of recre- (or of residential land) into the prices of ation decisions, the value of a change in environmen- individual attributes, including air qual- tal quality at site i is no longer measured as indicated ity. This is done using an hedonic price in Figure 2 (Hanemann 1984). Because utility is ran- dom from the viewpoint of the researcher, compen- function, which describes the equilib- sating variation for a change in quality at a recreation rium relationship between house price, site on a given day equals the change in utility condi- p, and attributes, A = (al, a2, . . . , tional on visiting the site times the probability that the site is visited, plus the change in the probability a,). The marginal price of an attribute of visiting- the site times the utility received from in the market is simply the partial deriva- the site. tive of the hedonic price function with 41 One solution to this problem, proposed by Ed- ward Morev 11984). is to estimate a share model, respect to that attribute. In selecting a which allocates the recreation budget for a season house, consumers equate their marginal among different sites. The drawback of this model willingness to pay for each attribute to is that the share of the budget going to each site is assumed to be positive, whereas, in reality, a house- its marginal price (S. Rosen 1974; A. hold may not visit all sites. Myrick Freeman 1974). This implies that Cropper and Oates: Environmental Economics 707

the gradient of the hedonic price func- Marginal Attribute Bid tion, evaluated at the chosen house, gives the buyer's marginal willingnesses to pay for each attribute. Somewhat more formally, utility maxi- mization in an hedonic market calls for the marginal price of an attribute to equal the household's marginal willingness to pay for the attribute,

Figure 3. The Identification Problem in an where 0 is the household's bid function, Hedonic Market the most one can take away from the household in return for the collection of amenities, A, and keep its utility con- ity changes, which requires estimating stant. Equation (15) implies that, in equi- marginal bid functions. S. Rosen origi- librium, the marginal willingness to pay nally suggested that this be done by re- for an attribute can be measured by its gressing marginal attribute price, com- marginal price, computed from the he- puted from the gradient of the hedonic donic price function. price function, on the arguments of the If a large improvement in environmen- marginal bid function. This procedure, tal quality is contemplated in one section however, may encounter an identifica- of a city-an improvement large enough tion problem which is caused by the fact to alter housing prices-the derivative that the arguments of the marginal attri- of the hedonic price function no longer bute bid function determine marginal at- measures the value of the amenity tribute price as well. change. In the short run, before house- An example of the identification prob- holds adjust to the amenity change and lem, provided by James Brown and Har- prices are altered, the value of the amen- vey Rosen (1982), occurs when the he- ity change is the area under the house- donic price function is quadratic and the hold's marginal bid function-the right marginal value functions are linear in at- hand side of (15)-between the old and tributes. In the case of a single amenity, new levels of air quality. To value the a,, amenity change in the long run, how- = + (16) ever, one must take into account the dplda1 Po Plal household's adjustment to the amenity dOlda, = bo + blal + b2M. (17) change and to any price changes that may In this case regressing Po + Plal on result. The area under the marginal bid al and M will reproduce the parameters function (the short-run welfare measure) of the marginal price function, i.e., 6, is, however, a lower bound to the long- = Po, 6, = Pl and 6, = 0. This is illus- run benefits of the amenity change (Bar- trated graphically in Figure 3. The prob- tik 1988b). lem is that the marginal price function Empirical applications of the hedonic does not shift independently of the mar- approach have typically focused either on ginal bid function. Shifts in the latter, valuing marginal amenity changes, which due, say, to differences in income, thus requires estimating only the hedonic trace out points on the marginal price price function, or on computing the function. short-run benefits of nonmarginal amen- To achieve identification in this ex- 708 Journal of Economic Literature, Vol. XXX (June 1992) ample, one can introduce functional' form compensating wage differentials across restrictions, such as adding a: to the mar- cities can be used to infer the value of ginal price function, but not to the mar- environmental amenities (Glenn Blom- ginal value function, which will cause d quist, Mark Berger, and John Hoehn pldai to shift independently of dOlda, 1988; Maureen Cropper and Amalia Arri- (Mendelsohn 1984). Another solution is aga-Salinas 1980; V. Kerry Smith 1983). to estimate hedonic price functions for Intuitively, the value people attach to ur- several markets, so that the coefficients ban amenities should be reflected in the of the marginal price function vary across higher wages they require to live in less cities (Palmquist 1984; Robert Ohsfeldt desirable cities. and Barton Smith 1985; Ohsfeldt 1988). When migration is possible, consum- For this to work, households in all cities ers choose the city in which they live to must have identical preferences; how- maximize utility; however, wage income, ever, the distribution of measured house- as well as amenities, vary from one city hold characteristics andlor the supply of to another (S. Rosen 1979; Jennifer Ro- amenities must vary across cities so that back 1982).42Household equilibrium re- the hedonic price function and its gra- quires that utility be identical in all cities. dient vary from one city to another. In The fact that consumers in all cities the case of several ai's, one can impose must enjoy the same level of utility im- exclusion restrictions on the ai's that en- plies that wages and land rents must ad- ter each marginal value function (Dennis just to compensate for amenity differ- Epple 1987) so that marginal prices vary ences. The marginal value of an amenity independently of the variables that enter change to a consumer is thus the sum the marginal value function. of the partial derivatives of an hedonic In view of the problems in estimating wage function and an hedonic property marginal attribute bid functions, it is im- value function (Roback 1982). portant to note that an upper bound to Hedonic Labor Markets. The fact that the long-run benefits of an amenity im- risk of death is a job attribute traded in provement can be obtained from the he- hedonic labor markets has provided donic price function alone. Yoshitsugu economists with an alternative to the Kanemoto (1988) has shown that the averting behavior approach as a means change in prices in the improved area of valuing mortality risk (Thaler and S. predicted by the hedonic price function Rosen 1976). The theory behind this ap- is an upper bound to the long-run bene- proach is simple: other things equal, fits of an amenity improvement. Thus, workers in riskier jobs must be compen- from knowledge of the hedonic price sated with higher wages for bearing this function alone one can obtain (1) the ex- risk. As in the case of hedonic housing act value of a marginal attribute change, markets, the worker chooses his job by and (2) an upper bound to the long-run equating the marginal cost of working in value of an attribute change. a less risky. -job--the derivative of the he- donic price function-to the marginal Wage-Amenity Studies. The analysis of benefit, the value (in dollars) of the re- hedonic housing markets, by focusing on sulting increase in life expectancy. housing market equilibrium within a There are three problems in using the city, implicitly ignores migration among compensating wage approach. One is cities. If one takes a long-run- view and that workers can move 42 In most models wages, lot size, and amenities from one city to another, then data on vary among, but not within, cities. Cropper and Oates: Environmental Economics 709

that compensating wage differentials ex- C. The Contingent Valuation Method ist only if workers are informed of job risks. Thus, the absence of compensating While the indirect market ap- differentials need not mean that workers proaches we have described above can do not value reducing the risk of death. be used to value many of the benefits A second problem is that compensating of pollution reduction, there are impor- differentials appear to exist only in union- tant cases in which they cannot be used. ized industries (William Dickens 1984; When no appropriate averting or mitigat- Douglas Gegax, Gerking, and Schulze ing behavior exists, indirect methods 1985). This suggests that the wage differ- cannot be used to estimate the morbidity ential approach may provide estimates benefits of reducing air pollution. Recre- of the value of a risk reduction only for ation benefits may be difficult to measure certain segments of the population. This since there may not be enough variation problem is compounded by the fact that in environmental quality across sites in the least risk averse individuals work in a region to estimate the value of water risky jobs. Third, if workers have biased quality using the travel cost approach. estimates of job risks, or if the objective There is, in addition, an entire cate- measures of job risk used in most wage gory of benefits-nonuse values-which studies over- or understate workers' risk cannot even in principle be measured by perceptions, market wage premia will indirect market methods. Nonuse values yield biased estimates of the value of a refer to the benefits received from know- risk reduction. ing that a good exists, even though the individual may never experience the The Hedonic Travel Cost Approach. Yet good directly. Examples include preserv- another area in which the hedonic ap- ing an endangered species or improving proach has been applied is in valuing the visibility at the Grand Canyon for per- attributes of recreation sites (G. Brown sons who never plan to visit the Grand and Mendelsohn 1984). In valuing sites, Canyon. the analog to the hedonic price function This suggests that direct questioning is obtained by regressing the cost of trav- can play a role in valuing the benefits elling to a recreation site on the attri- of pollution control. Typically, direct butes of the site, such as expected fish questioning or contingent valuation stud- catch, clarity of water, and water color. ies ask respondents to value an output, However, because this relationship is not such as a day spent hunting or fishing, the result of market forces, there is noth- rather than a change in pollution concen- ing to guarantee that the marginal cost trations per se. Examples of commodities of an attribute is positive. More desirable that have been valued using the contin- sites may be located closer to population gent valuation method (CVM) include centers rather than farther away from improvements in water quality to the them.43 In this case, the individual's point where the water is fishable or choice of site will not be described by swimmable (Richard Carson and Robert (13), and care must be taken when infer- Mitchell 1988), improvements in visibil- ring values from marginal attribute costs ity resulting from decreased air pollution (V. Kerry Smith, Palmquist, and Paul Ja- (Alan Randall, Berry Ives, and Clyde kus 1990). Eastman 1974; Schulze and David Brook- shire 1983; Decision Focus 1990), the 43 The problem may be reduced by using only sites value of preserving endangered species actually visited from a given origin in estimating the hedonic travel cost function. (James Bowker and John Stoll 1988; 710 Journal of Economic Literature, Vol. XXX (June 1992)

Kevin Boyle and Richard Bishop 1987), contrast, a much easier task, and one that and days free of respiratory symptoms parallels decisions made when purchas- (George Tolley et al. 1986b; Dickie et ing goods sold in conventional markets. al. 1987). It must be acknowledged that, despite Any contingent valuation study must advances made in contingent valuation incorporate (1) a description of the com- methodology during the last 15 years, modity to be valued; (2) a method by many remain skeptical of the method. which payment is to be made; and (3) a Perhaps the most serious criticism is that method of eliciting values. In studies that responses to contingent valuation ques- value recreation-related goods, hypoth- tions are hypothetical-they represent etical payment may take the form of a professed, rather than actual, willingness user fee or an increase in taxes; in the to pay. This issue has been investigated case of improved visibility, a charge on in at least a dozen studies that compare one's utility bill, since power plant pollu- responses to contingent valuation ques- tion can contribute to air quality degrada- tions with actual payments for the same tion. To determine the maximum a per- commodity. son is willing to pay for an improvement How close hypothetical values are to in environmental quality, the inter- actual ones depends on whether the com- viewer may simply ask what this amount modity is a public or private good, on is (an open-ended survey), or he may ask the elicitation technique used, and on whether or not the respondent is willing whether it is willingness to pay (WTP) to pay a stated amount (a closed-ended for the good or willingness to accept com- survey). The yeslno answer does not pensation (WTA) that is elicited. Most yield an estimate of each respondent's experiments comparing hypothetical and willingness to pay; however, the fraction actual WTP for a private good (straw- of respondents willing to pay at least the berries or hunting permits) have found stated amount gives a point on the cumu- no statistically significant difference be- lative distribution function of willingness tween mean values of hypothetical and to pay for the commodity (Trudy Cam- actual willingness to pay (Dickie, Ann eron and Michelle James 1987). Fisher, and Gerking 1987; Bishop and There seems to be general agreement Thomas Heberlein 1979; Bishop, Heber- that closed-ended questions are easier for lein, and Mary Jo Kealy 1983). Such is respondents to answer and therefore not the case when hypothetical and actual yield more reliable information than WTA are compared. In three experi- open-ended questions, especially when ments involving willingness to accept the commodity valued is not traded in compensation for hunting permits, conventional markets. Asking an open- Bishop and Heberlein (1979) and Bishop, ended question about a good that respon- Heberlein, and Kealy (1983) found that dents have never been asked to value, actual WTA was statistically significantly such as improved visibility, often yields lower than hypothetical WTA in two out a distribution of responses that has a large of three cases. Hypothetical and actual number of zero values and a few very WTP have also been found to differ when large ones. This may reflect the fact that the commodity valued is a public good respondents have nothing to which to an- (Kealy, Jack Dovidio, and Mark L. chor their responses, and are unwilling Rockel 1987). , to go through the reasoning necessary Other criticisms of the CVM have fo- to discover the value they place on the cused on: (1) the possibility that individu- good. Answering a yeslno question is, by als may behave strategically in answering Cropper and Oates: Environmental Economics 711

questions-either overstating WTP if this that willingness to pay for an environ- increases the likelihood that an improve- mental improvement is usually many ment is made, or understating WTP if it times lower than willingness to accept reduces their share of the cost (the free- compensation to forego the same im- rider problem); (2) the fact that individu- provement (Judd Hammack and G. als may not be sufficiently familiar with Brown 1974; Bishop and Heberlein 1979; the commodity to have a well-defined Robert Rowe, d'Arge, and Brookshire value for it; and (3) the fact that WTP 1980; Jack Knetsch and J. A. Sinden for a commodity is often an order of mag- 1984). This is sometimes interpreted as nitude less than willingness to accept evidence that the method of eliciting re- (WTA)compensation for the loss of the sponses is unsatisfactory; however, as we commodity. noted above, there is no reason why The possibility that respondents be- WTA for a quality (public good) decrease have strategically has been tested in should not exceed WTP for an increase laboratory experiments by examining of the same magnitude, provided that whether announced WTP for a public there are few substitutes for the public good varies with the method used to fi- good.44An alternative explanation for the nance the public good. Studies by Bohm WTAIWTP discrepancy that has been of- (1972), Bruce Scherr and Emerson Babb fered by some economists (Donald Cour- (1975), and Vernon Smith (1977, 1979) sey, John Hovis, and Schulze 1987; suggest that strategic behavior is not a Brookshire and Coursey 1987) is that in- problem, possibly because of the effort dividuals are simply not as familiar with that effective strategic behavior requires. the sale of an item as with its purchase. If the commodity to be valued is not These authors find that, in experiments well understood, contingent valuation re- where individuals were allowed to sub- sponses are likely to be unreliable: re- mit bids or offers for the same commod- sponses tend to exhibit wide variation, ity, WTA approached WTP after several and respondents may even prefer less of rounds of transactions. 45 a good to more! One interpretation of D. Applications of Valuation Techniques this result is that people really do not have values for the commodity in ques- Having described the main tech- tion-they are created by the researcher niques used to value environmental in the course of the survey (Thomas amenities, we now wish to give the Brown and Paul Slovic 1988). This is a reader a feel for the way in which these serious criticism: Do people really know enough about groundwater contamina- 4"he explanation of the discrepancy between tion or biodiversity to place a value on WTA and WTP offered by psychologists-that mone- tary losses from some reference point are valued either good? more highly than monetary gains (Daniel Kahneman Fortunately, it is possible to defend and Amos Tversky 1979talso suggests that this dis- against this criticism by seeing how re- parity has nothing to do with flaws in the contingent valuation method. sponses vary with the amount of informa- 45 None of these explanations, however, seems to tion that is provided about the commod- account for results obtained by Kahneman, Knetsch, ity being valued. If values are well and Thaler (1990). They find that, even for common items such as coffee mugs and ballpoint pens, sellers defined, they should not, on average, have reservation prices that are higher, much higher vary with small changes in the amount on average, than buyers' bid prices. This disparity of information. does not disappear after several rounds of trading. The initial distribution of property rights (the "en- One of the most striking and challeng- dowment effect") may, therefore, matter, even for ing findings emerging from this work is goods with many substitutes. 712 Journal of Economic Literature, Vol. XXX (lune 1992)

TABLE 1 TOTALANNUALIZEDENVIRONMENTAL COMPLIANCE COSTS, BY MEDIUM,1990 (Millions of 1986 dollars)

Mediuin Costs Major Statutes Air and Radiation, Total 28,029 Air 27,588 Clean Air Act (CAA) Radiation 441 Radon Pollution Control Act Water, Total Water Quality Clean Water Act (CWA) Drinking Water Safe Drinking Water Act Land, Total RCRA Resource Conservation and Recovery Act (RCRA) Superfund Comprehensive Environmental Response, Compensation and Liability Act (CERCLA) Chemicals, Total Toxic Substances Toxic Substances Control Act (TSCA) Pesticides Federal Insecticide, Fungicide and Rodenticide Act (FIFRA) Total Costs 100,167

Note: These represent the costs of complying with all federal pollution control laws, assuming full implementation of the law (USEPA 1990). techniques have been used to value the ing Water Act is, similarly, to provide benefits of pollution control. We shall be- a margin of safety in protecting the gin with an overview of the types of bene- country's drinking water supplies from fits associated with the major pieces of toxic substances, while the goal of environmental legislation. We then turn the Federal Insecticide, Fungicide, and to a description and assessment of actual Rodenticide Act (FIFRA) is to prevent benefit estimation. adverse effects to human health and to Table 1 lists the major pieces of envi- the environment from the use of pesti- ronmental legislation in the U. S. and the cides. estimated costs of complying with each Each of the statutes in Table 1 also statute in 1990. With the exception of results in certain nonhealth benefits. The the Clean Water Act, the primary goal Clean Air Act provides important aes- of U. S. environmental legislation is to thetic benefits in the form of increased protect the health of the population. visibility, and the 1990 Amendments to According to the Clean Air Act, ambi- the Act, designed to reduce acid rain, ent standards for the criteria air pollu- may yield ecological and water quality tants are to be set to protect the health benefits. The Clean Water Act-whose of the most sensitive persons in the goal is to make all navigable water bodies population.46 The goal of the Safe Drink- fishable and swimmable-yields recre- ational and ecological benefits. Both Acts yield benefits to firms in agriculture, for- 4fi The criteria air pollutants are particulate matter, sulfur oxides, nitrogen oxides, carbon monoxide, estry, and commercial fishing. FIFRA, lead, and ozone. the primary law governing pesticide us- Cropper and Oates: Environmental Economics 713

age, is designed to protect animal as well giardiasis) may also cause acute illness. as human health. Reductions in risk of death have been In addition to the pollution problem valued using three methods: averting be- addressed by the major environmental havior, hedonic analysis, and contingent statutes, there is increasing concern valuation. The most common approach about the effects of emissions of green- to valuing changes in risk of death due house gases, including carbon dioxide, to environmental causes is hedonic wage chlorofluorocarbons (CFCs) and meth- studies. The results of these studies are ane. Studies suggest that emissions of typically expressed in terms of the value these gases may contribute to increases per "statistical life" saved. If reducing ex- in mean temperature, especially in the posure to some substance reduces cur- Northern Hemisphere, changes in pre- rent probability of death by 10-'for each cipitation, and sea level rises that could of 200,000 persons in a population, it average 65 cm by the end of the next will save two statistical lives (lop5 x century. The main effects of these 200,000). If each person is willing to pay changes are likely to be felt in agricul- $20 for the lo-' risk reduction, then the ture, in animal habitat, and in human value of a statistical life is the sum of comfort. these willingnesses to pay ($20 x In light of the preceding discussion, 2OO,OOO), divided by the number of sta- we review empirical work for four catego- tistical lives saved, or $2,000,000. ries of nonmarket benefits: health, recre- Recent compensating wage studies ation, visibility, and ecological benefits. (Ann Fisher, Daniel Violette, and Lau- We also discuss the benefits of pollution raine Chestnut 1989) generate mean esti- control to agriculture. mates of the value of a statistical life that 1. The Health Benefits of Pollution fall within an order of magnitude of one Control. The statutes listed in Table 1 another: $1.6 million to $9 million contribute to improved human health in ($1986), with most studies yielding mean several ways. By reducing exposure to estimates between $1.6 million and $4.0 carcinogens-in the air, in drinking wa- million. Contingent valuation studies ter, and in food-environmental legisla- that value reductions in job-related risk tion reduces the probability of death at of death (Gerking, Menno DeHaan, and the end of a latency period-the time Schulze 1988) or reductions in risk of auto that it takes for cancerous cells to de- death (Jones-Lee, M. Hammerton, and velop. Mortality benefits are also associ- P. R. Philips 1985) fall in the same range. ated with control of noncarcinogenic air Averting behavior studies-based on pollutants, which reduces mortality espe- seat belt use (Blomquist 1979) or the use cially among sensitive persons in the pop- of smoke detectors (Rachel Dardis ulation, e.g., angina sufferers or persons 1980)-yield estimates of the value of a with chronic obstructive lung disease. statistical life that are an order of magni- Lessening children's exposure to lead in tude lower than the studies cited above. gasoline or drinking water avoids learn- These studies, however, estimate the ing disabilities and other neurological value of a risk reduction for the person problems associated with lead poisoning. who just finds it worthwhile to undertake Finally, controlling air pollution reduces the averting activity. This is because illness-ranging from minor respiratory buckling a seat belt or purchasing a symptoms associated with smog (runny smoke detector are 0-1 activities. They nose, itchy eyes) to more serious respira- are undertaken provided that their mar- tory infections, such as pneumonia and ginal benefit equals or exceeds their mar- influenza. Water borne disease (e.g., ginal cost, with equality of marginal ben- 714 Journal of Economic Literature, Vol. XXX (June 1992) efit and marginal cost holding only for amount to reduce risks to life makes pos- the marginal purchaser. If 80 percent of sible rational debate and analysis in the all persons use smoke detectors, the policy arena over tradeoffs in risk reduc- value of the risk reduction to the mar- tion. Moreover, estimates of the value ginal purchaser may be considerably of a statistical life are in sufficiently close lower than the mean value. agreement to permit their use in actual There are, however, other problems benefit-cost calculations (subject, per- in using the indirect market approaches haps, to some sensitivity analysis). we have reviewed here to value changes The valuation of morbidity has been in environmental risks. One problem is less successful. Estimates of the value of that the risks valued in labor market and reductions in respiratory symptoms come averting behavior studies are more vol- from two sources: averting behavior stud- untary than many environmental risks. ies and contingent valuation studies. The Work by Slovic, Baruch Fischhoff, and averting behavior approach has been Sarah Lichtenstein (1980, 1982) suggests used to value illnesses associated with that willingness to pay estimates ob- both water and air pollution. It has been tained in one context may not be transfer- more successful in the case of water pol- able to the other. Second, death due to lution because an averting behavior ex- an industrial accident is often instanta- ists (buying bottled water) that is closely neous, whereas death resulting from en- linked to water pollution (Abdalla 1990; vironmental contaminants may come Harrington, Krupnick, and Walter Spof- from cancer and involve a long latency ford 1989). By contrast, the averting be- period. Deaths due to cancer thus occur haviors used to value air pollution-run- in the future and cause fewer years of ning an air-conditioner in one's home or life to be lost than deaths in industrial car-are in most cases not undertaken accidents. At the same time, however, primarily because of pollution. The use cancer is one of the most feared causes of doctor visits (purpose unspecified) to of death. mitigate the effects of air pollution suffers In a study designed to value reductions from a similar shortcoming. in chemical contaminants (trihalometh- Contingent valuation studies of respi- anes) in drinking water, Mitchell and ratory symptoms (coughing, wheezing, Carson (1986) found that the former ef- sinus congestion) have encountered two fect seems to be important: the value of problems. The first concerns what is to a statistical life associated with a reduc- be valued. Ideally, one would like to tion in risk of death 30 years hence was value a change in air pollution which, only $181,000 ($1986). This is lower than after defensive behavior is undertaken, the value of a statistical life associated might cause a change in the level of with current risk of death for two reasons: the symptom experienced. The individ- (1)the number of expected life years lost ual's willingness to pay for the pollution is smaller if the risk occurs 20 years change includes the value of the change hence, and (2) the individual may dis- in illness after mitigating behavior is un- count the value of future life years lost dertaken, plus the cost of the mitigating (Cropper and Frances Sussman 1990; behavior. This suggests that a symptom Cropper and Paul Portney 1990). day be valued after mitigating actions In spite of these difficulties, valuing have been taken. A second problem is mortality risks is an area in which econo- that the respondent must be encouraged mists have made important contribu- to consider carefully his budget con- tions. The notion that, ex ante, individu- straint. Failure to handle these problems als are willing to spend only a certain has led to unbelievably high average Cropper and Oates: Environmental Economics 715

values of a symptom day. In more careful multiple symptoms, and since the value studies, mean willingness to pay to elimi- of jointly reducing several symptoms is nate one day of coughing range from generally less than the sum of the values $1.39 ($1984) (Dickie et al. 1987) to of individual symptom reductions. In the $42.00 ($1984) (Edna Loehmann et al. case of mortality risks, more refined esti- 1979); for a day of sinus congestion $1.88 mates are needed that take into account (Dickie et al.) to $52.00 (Loehmann et the timing of the risk, the degree of al.). voluntariness, and the cause of death. An alternative approach to valuing The timing issue is especially crucial morbidity is to use the cost of illness- here: the benefits of environmental pro- the cost of medical treatment plus lost grams to reduce exposure to carcinogens, earnings-which, as Harrington and such as asbestos, are not realized until Portney (1976) have shown, is a lower the end of a latency period-perhaps 40 bound to willingness to pay for the years in the case of asbestos. Since the change in illness. Mean willingness to exposed population is 40 years older, pay for symptom reduction is usually fewer life-years are saved, compared three to four times higher than the tradi- with programs that save lives immedi- tional cost of illness. Berger et al. (1987) ateIy.47 report a mean WTP of $27 to eliminate 2. The Recreation Benefits of Pollution a day of sinus congestion, compared with Control. Reductions in water pollution an averge cost of illness of $7. The corre- may enhance the quality of recreation ex- sponding figures for throat congestion are periences by allowing (or improving) $44 and $14. swimming, boating, or fishing. Most Studies of willingness to pay to reduce studies of the recreation benefits of water the risk of chronic disease are few (W. pollution control have focused on fishing- Kip Viscusi, Magat, and Joel Huber related benefits, and it is on them that 1988, is a notable exception), and cost we concentrate our attention. of illness estimates are more prevalent Travel cost studies have taken one of in valuing chronic illness (Ann Bartel and three approaches to valuing the fishing Paul Taubman 1979; Barbara Cooper and benefits of improved water quality. In Dorothy Rice 1976). Viscusi, Magat, and some studies (V. Kerry Smith and Des- Huber estimate the value of a statistical vousges 1986a), measures of water qual- case of chronic bronchitis to be $883,000, ity such as dissolved oxygen are valued approximately one-third of the value of directly. That is, water quality variables a statistical life. This may be contrasted directly enter equations that describe the with cost of illness estimates of $200,000 choice of recreation site or demand func- per case of chronic lung disease (Cropper tions for site visits." This approach is and Krupnick 1989). clearly useful if one wishes to link the As the preceding discussion indicates, valuation study to pollution control poli- more work is needed in the area of both morbidity and mortality valuation. Be- 47 While some studies have attempted to take the cause of the in finding activities latency period and number of life-years saved illto that mitigate the effects of air pollution, account (Josephine Mauskopf 1987), this is not the contingent valuation studies would seem general practice (Cropper and Portney 1990). "This approach is also used when the recreation to be a more promising approach to valu- activity studied is swimming or viewing, activities ing morbidity. If new studies are done, where perceptions of water quality are likely to be they should value of symp- linked to water clarity and odor. It has, for example, been applied in studies of beach visits in Boston toms rather than symptoms, (Bockstael, Hanemann, and Kling 1987) and lake vis- since pollution exposures often tri~ger its in Wisconsin (George Parsons and Kealy 1990). 716 Journal of Economic Litel -ature, Vol. XXX (June 1992)

cies, such as policies to reduce biochemi- models are typically used in conjunction cal oxygen demand (BOD), a measure with models that predict the total num- of the oxygen required to neutralize or- ber of trips. Treating changes in water ganic waste. A second approach is to re- quality as altering the supply of available late site visits (or choice of site) to fish sites captures participation effects but not catch. Fish catch is clearly more closely improvements in quality at existing sites. associated with motives for visiting a site In addition to travel cost models, con- than is dissolved oxygen; however, it tingent valuation studies have been used must be linked to changes in the fish pop- to value improvements in fish catch or ulation, which must, in turn, be linked water quality. Because it is difficult to to changes in ambient water quality. ask consumers to value changes in dis- A third approach is to treat changes solved oxygen levels or fecal coliform in water quality as effectively eliminating count-another measure of water qual- or creating recreation sites. This ap- ity-without linking these water quality proach has been used in valuing the ef- measures to the type of activities they fects of acid rain on fishing in Adirondack support, many CVM studies use the RFF lakes: reductions in pH below certain Water Quality Ladder (Vaughan and thresholds have been treated as eliminat- Russell 1982), which relates a water qual- ing acres of surface area for fishing of ity index to the type of water use-boat- particular species (John Mullen and ing, fishing (rough fish), fishing (game Frederic Menz 1985). It is also the ap- fish), swimming-that can be supported proach used by Vaughan and Russell by various levels of the index. It is these (1982) in valuing the benefits of the Clean activity levels that are valued by respon- Water Act. They treat the benefits of dents. The water quality ladder has been moving all point sources to the Best Prac- used both to value water quality at spe- tical Control Technology Currently cific sites (e.g., the Monongahela River, Available (BPT) as an inciease in the by V. Kerry Smith and Desvousges number of acres of surface water that sup- 1986a) and at all sites throughout the port game fish (bass, trout) as opposed country (Carson and Mitchell 1988). to rough fish (carp, catfish). The Clean It is interesting to compare estimates Water Act is thus viewed as increasing of the value of water quality improve- the number of recreation sites, rather ments obtained by the travel cost and than raising fish catch at existing sites. contingent valuation approaches. Carson Regardless of the form of water recre- and Mitchell (1988) report that house- ation valued, an improvement in water holds are, on average, willing to pay $80 quality has two effects: it increases the per year (in 1983 dollars) for an improve- utility of people who currently use the ment in water quality throughout the resource, and it may increase participa- U.S. from boatable to fishable (capable tion rates (number of days spent fishing). of supporting game fish). V. Kerry Smith Varying parameter models that value and Desvousges (1986a) report a mean changes in water quality or fish catch us- value of $25 per household for the same ing the shift in demand for site visits (see improvement in a five-county region in Figure 2) capture both effects. Discrete western Pennsylvania. The difference be- choice models measure the effect of a tween these estimates reflects the fact quality improvement on a given recre- that non-use values are important: house- ation day, but do not estimate the effect holds care about clean water in areas of quality changes on the total number where they do not live. Even the $25 of days spent fishing; however, these estimate for western Pennsylvania re- Cropper and Oates: Environmental Economics 717

fleets nonuse values, since only one-third a day of coughing independently of loca- of the households surveyed engaged in tion, it is harder to value a generic fishing some form of water based recreation. day. Because they do not capture nonuse This raises important questions con- values, travel cost estimates of the value cerning priorities for research in the area of improving water quality are not di- of recreation benefits.4g Future research rectly comparable with those obtained can proceed using a contingent valuation using the CVM. Using a varying parame- approach in which use and nonuse values ter model, V. Kerry Smith and Desvous- are elicited simultaneously for sites in the ges (1986a) find the value of an improve- respondent's region. The problem here ment in water quality from boatable to is to have the respondent value an im- fishable to be between $0.06 and $30.00 provement to recreation that is suffi- per person per day ($1983) for 30 Army ciently specific that it can be related to Corps of Engineers sites. This value may changes in pH levels from acid rain or be contrasted with estimates of $5 to $10 changes in levels of dissolved oxygen as- per person per day ($1983) obtained by sociated with the adoption of BPT. The Vaughan and Russell. advantage of this approach is that it The preceding discussion suggests two would capture both use and nonuse val- problems that arise in valuing water qual- ues. The advantage of the travel cost ap- ity benefits that do not arise in valuing proach is that it could use endpoints health effects. The first is an aggregation more closely related to pollution (such problem. Suppose that one wishes to as dissolved oxygen); however, it would value the benefits of water quality im- not yield estimates of nonuse values. provements in a river basin, and suppose 3. The Visibility Benejts of Pollution that the travel cost approach is used to Control. Reductions in air pollution, by measure use values associated with an increasing visibility, may improve the improvement in dissolved oxygen or fish quality of life in urban areas as well as catch. The nonuse values associated with at recreation sites. Since the number of these improvements could be measured persons affected by improvements in visi- using a contingent valuation study. How- bility is large-at least as great as the ever, while the responses of nonusers number of persons whose health is af- could be added to values obtained from fected by air pollution-the potential the travel cost approach, it would, in value of such benefits is great. practice, be hard to separate use from One can view the results of hedonic nonuse values in the responses of fisher- property value studies performed in the men. 1970s and early 1980s as evidence that The second problem is one of transfer- people value the visibility benefits of pol- ring results from a water quality study lution control. In these studies housing done in one geographic area to another prices were regressed on measures of am- area. While one can easily control for dif- bient air quality such as particulates or ferences in willingness to pay in the two sulfates, which are negatively correlated regions associated with differences in in- come and population, the value of water 49 It should be emphasized that, while there exist several dozen studies of water quality benefits in a quality improvements is also likely to recreation context, Inany studies analyze the same vary with the particular aesthetic and data. Thus, empirical estimates of water quality bene- other characteristics of the region-and fits exist for only a few areas of the country-lakes in Wisconsin and the Adirondacks, beaches in Boston such characteristics are intrinsically hard and on the Chesapeake Bay, recreation sites in west- to measure. Thus, whereas one can value ern Pennsylvania. 718 Journal of Economic Literature, Vol. XXX (June 1992)

with visibility. The studies, most of are computed may be considerably larger which found significant negative effects than for urban visibility benefits. The of air pollution on housing prices, thus EastIWest distinction is important both provide indirect evidence that people are because of differences in baseline visibil- willing to pay for improved visibility.50 ity and because of qualitative differences For example, John Trijonis et al. (1984) in the nature of visibility impairments, estimated based on differences in hous- e.g., haze versus brown cloud. ing prices that households in San Fran- There are two key problems in any cisco were willing, on average, to pay contingent valuation study of visibility. $200 per year for a 10 percent improve- One is presenting changes in visibility ment in visibility. that are both meaningful to the respon- The difficulty in using these studies to dent and that can be related to pollution estimate benefits, however, is that the control policies. The other is separating coefficient of air pollution (or visibility) the respondent's valuation of health ef- captures all reasons why households may fects from his valuatioil of visibility prefer to live in nonpolluted areas-in- changes. eluding both improved health and re- Most CVM studies define increased duced soiling. Indeed, the reason why visibility as an improvement in visual property value studies have become less range-the distance at which a large, popular as a method of valuing the bene- black object disappears from view. Visual fits of pollution control is that it is difficult range is both correlated with people's to know what the pollution coefficient perceptions of visibility and with ambient captures and, therefore, difficult to ag- concentrations of certain pollutants (fine gregate benefit estimates obtained from nitrate and sulfate aerosols). Differences these studies with those obtained from in visual range are presented in a series other approaches. Such aggregation is of pictures in which all other condi- necessary because residential property tions-weather, brightness, the objects value studies capture benefits only at photographed-are, ideally, kept con- home and not at the other locations the stant. household frequents. It has long been recognized (Brook- For these reasons contingent valuation shire et al. 1979) that, in responding to seems the most promising method for such pictures, people assume that the valuing visibility. Because visibility ben- health effects of pollution diminish as efits vary regionally, CVM studies can visibility improves. Health effects are most usefully be classified according to therefore inherently difficult to separate whether they measure urban visibility from visibility changes. The best way to benefits or benefits at recreation sites, handle this problem is to ask respondents and according to whether the locations what they assume health effects to be studied are in the Eastern or in the West- and then to control for these effects. ern United States. The former distinction Unfortunately, existing CVM studies is important because visibility benefits of visibility benefits-especially those for at recreation sites-especially national urban areas-have failed to treat the is- parks-are likely to have a substantial sues raised above in a satisfactory man- nonuse component; consequently, the ner. With this limitation in mind, it is relevant population for which benefits nonetheless of interest to contrast the magnitude of benefits associated with im- provements in urban air quality with esti- Freeman (1979a) provides an excellent summary of early studies. mates obtained from hedonic property Cropper and Oates: Environmental Economics 719 value studies. Studies of visibility im- et al. (1986a) found respondents were provements in eastern U.S. cities (Tolley willing to pay only $22 annually for the et al. 1986a; Douglas Rae 1984) have esti- same visibility improvement at the mated that households would pay ap- Grand Canyon when this was valued at proximately $26 annually for a 10 percent the same time as visibility improvements improvement in visibility. Loehmann in Chicago (the site of the interviews) Boldt, D., and Chaikin, K. (1981) reports and throughout the East coast. an annual average willingness to pay per 4. The Ecological Benefits of Pollution household of $101 for a 10 percent im- Control.52 By the ecological benefits of provement in visibility in San Francisco. pollution control, we mean reduced pol- Both figures are considerably lower than lution of animal and plant habitats, such estimates implied by property value as rivers, lakes, and wetlands. Because studies. the benefits of clean water to recreational Studies in recreation areas have fo- fisherman or larger populations of deer cused on major national parks, including to hunters are captured in recreation the Grand Canyon (Decision Focus 1990; studies, the benefits discussed in this sec- Schulze and Brookshire 1983), because tion are the nonuse benefits associated of the possibility of large nonuse values with reduced pollution of ecosystems. attached to visibility benefits at these It should be clear to the reader that sites. Two conclusions emerge from valuing this category of benefits poses se- these studies. First, nonuse values ap- rious conceptual problems. One is defin- pear to be large relative to use values. ing the commodity to be valued. Does Use values associated with an improve- one value reductions in pollution concen- ment in visibility at the Grand Canyon trations, increases in animal populations, from 70 to 100 miles are under $2.00 or some more subtle index of the health per visitor party per day ($1988) (Schulze of an ecosystem? Two approaches can be and Brookshire 1983; K. K. MacFarland taken here. The "top down" approach et al. 1983). By contrast, Schulze and asks the respondent to value the preser- Brookshire found that a random sample vation of an ecosystem, such as 100 acres of households were willing to pay $95 of wetland (John Whitehead and Blom- per year ($1988) to prevent a deteriora- quist 1991). The "bottom up" approach tion in visibility at the Grand Canyon values the preservation of particular spe- from the 50th percentile to the 25th per- cies inhabiting the wetland, such as geese centile. and other birds. Second, the embedding, or superaddi- Regardless of the approach taken, sev- tivity, problem is potentially quite seri- eral problems must be faced. One diffi- ous. This refers to the fact that, in gen- culty is defining what substitutes are as- eral, an individual's willingness to pay sumed to exist, whether for a particular for simultaneous improvements in visi- species or for a wetland (Whitehead and bility at several sites should be less than Blomquist 1991). Presumably the value the sum of his willingness to pay for iso- lated improvements at each site (Hoehn "Outside environmental econo~nics,there is a and Randall 1989). In a follow-up study considerable literature in environmental ethics that to Schulze and Brookshire (1983), Tolley explores the issue of nonhuman rights and their pol- icy implications. From this perspective, the econo- mist's benefit-cost calculation with its wholly anthro- This figure, reported by Chestnut and Rowe pocentric orientation is an excessively narrow and (1989), is an average of mean willingness to pay for illegitimate framework for analysis. Kneese and each city surveyed by Tolley and Rae, based on Schulze (1985)provide an excellent treatment of this Chestnut and Rowe's reanalvsis of the data. set of issues. 720 Journal of Economic Literature, Vol. XXX (June 1992)

placed on the preservation of 10,000 sized the nonmarket benefits of pollution geese depends on the size of the goose control, some benefits accrue directly to population. A related problem arises firms, and can be measured by examining when programs are valued one at a time; shifts in the supply curves for the affected in general, the value attached to preserv- outputs. The industries that are most ing several species at the same time is subject to ambient air and water pollu- less than the sum of the values attached tion are forestry, fishing, and agriculture. to preserving each species in isolation. We focus on agriculture because it is the This implies that the totality of what is sector that is likely to experience the to be preserved should be valued: one largest benefits from pollution control. cannot compute this by summing the Reductions in ozone concentrations values attached to individual compo- and, possibly, in acid rain, should in- nents. crease the yields of field crops such as To date, most studies of endangered soybeans, corn, and wheat. In addition, species have valued individual species in reductions in greenhouse gases, to the isolation. For example, Bowker and Stoll extent that they prevent increases in (1988) estimate that households are, on temperature and decreases in precipita- average, willing to pay $22 per year tion in certain areas, should also increase ($1983) to preserve the whooping crane, crop yields. while Boyle and Bishop (1987) find that In measuring the effects on agricultural non-eagle watchers are willing to spend output of changes in pollution concentra- $11 per year to preserve the bald eagle tions or climate, two approaches can be in the state of Wisconsin. These values taken. The damage function approach are appropriate if one is considering a translates a change in environmental con- program to preserve either of these spe- ditions into a yield change, assuming that cies in isolation; however, the values farmers take no actions to mitigate the should not be added together if one is effects of the change. The yield change contemplating preserving both species. shifts the supply curve for the crop in Even if one decides to value a wetland question, and the corresponding changes (of given size) and defines the nature of in consumer and producer surpluses are substitutes, an important question re- ~alculated.~~This is the predominant ap- mains: do people really have well-de- proach used thus far to analyze the effects fined, or in the terminology of psycholo- of global climate change (Sally Kane, gists, "crystallized" values for these John Reilly, and Tobey 1991). It has also commodities? Since respondents in CVM been used in some studies of the effects studies are likely to be less familiar with of ozone on field crops (Richard Adams, ecological benefits than with health and Thomas Crocker, and Richard Katz 1984; recreation benefits, responses are likely Raymond Kopp et al. 1985; Kopp and to depend critically on the information Krupnick 1987). given to respondents in the survey itself The averting behavior approach allows (Karl Samples, John Dixon, and Marcia farmers to adjust to the change in pollu- Gown 1986). This problem, however, is tionlclimate by altering their input mix widely recognized, and recent studies and/or by adjusting the number of acres have taken pains to see how responses are influenced by the amount of informa- 53 In calculating the welfare effects of a shift in tion provided. supply, one must be careful to take into account the effects of agricultural price support programs, which 5. The Agricultural Benejts of Pollu- distort market prices. See Erik Lichtenberg and Da- tion Control. Although we have empha- vid Zilberman (1986). Cropper and Oates: Environmental Economics 721

planted. In some applications, a profit these damages are borne by consumers. function is estimated in which the envi- Producers in most cases gain from yield ronmental pollutant enters as a parame- decreases due to the resulting increases ter (James Mjelde et al. 1984; Philip Gar- in prices! cia et al. 1986). The value of the change Kane, Reilly, and Tobey (1991) obtain in Q can then be computed directly from similar results when estimating the wel- the profit function. If the resulting shift fare effects of global climate change on in supply is big enough to alter market agriculture: reductions in the yields of price, the welfare effects of these price field crops (wheat, corn, soybeans, and changes must also be computed. rice) in the U. S., Canada, China, and A more common approach is to solve the USSR benefit producers worldwide for the effect of the change in pollution due to increases in commodity prices. on output using a mathematical program- Consumers, however, lose. Thus, al- ming model whose coefficients have not though the aggregate losses to producers been econometrically estimated (Adams, and consumers worldwide are small Scott Hamilton, and Bruce McCarl 1986; (about one-half of one percent of world Scott Hamilton, McCarl, and Adams GDP), food-importing countries such as 1985). The effect of output changes on China suffer large welfare losses (equal price is then computed separately. to 5.5 percent of GDP) while food export- While benefit estimates that allow ers such as Argentina enjoy welfare gains. farmers to adjust to changes in pollution E. Measuring the Costs of Pollution are clearly preferable on theoretical Control grounds to estimates that do not allow such adjustments, it is important to ask Table 1, which lists the costs of the how much of a difference this is likely major environmental statutes, may give to make empirically, especially as the the reader the impression that measuring damage function approach is much easier the costs of pollution control is a straight- to implement. For changes in tempera- forward matter. Such is not the case. ture and precipitation, damages are To begin with, the costs of pollution likely to be greatly overstated if opportu- control must be measured using the same nities for mitigating behavior (e.g., irri- concepts that are used to measure the gation) are ignored.54On the other hand, benefits of pollution control: the change mitigating behavior does not seem to in consumer and producer surpluses as- make a great deal of difference in the sociated with the regulations and with case of ozone damage (Scott Hamilton, any price and/or income changes that McCarl, and Adams 1985). may result. The figures in Table 1repre- Estimates of annual damage to field sent, for the most part, expenditures on crops from a 25 percent increase in ozone cleaner fuels or abatement control equip- are in the neighborhood of $2 billion ment by firms. They do not represent ($1980)-not negligible, but small rela- the change in firms' profits, and thus ig- tive to estimates of health damages. It nore any adjustments firms may make is also interesting to note that most of to these expenditures. The figures also ignore the price and output effects associ-

54 We base this statement on the results of the ated with reducing emissions. At the very RFF MINK project (Norman Rosenberg et al. 1990), least, one would want to take into ac- which examines damages associated with climate count the price changes likely to result change-specifically, a return to the climate of the dust bowl-in Missouri, Iowa, Nebraska, and Ken- within a sector because of environmental tucky, under alternate adjustment scenarios. regulations-for example, one would 722 Journal of Economic Literature, Vol. XXX uune 1992)

want to measure the welfare effects of In the long run, however, the social an increase in electricity prices resulting costs of the Clean Air and Clean Water from the 10 million ton reduction in SO2 Acts exceed simple expenditure esti- emissions by electric utilities projected mates because of the effects of decreases under the 1990 Amendments to the in income on saving and investment. In Clean Air Act. their analysis of the effects of environ- We note that, at least in the short run, mental regulation on U.S. economic the effect of ignoring these adjustments growth, Dale Jorgenson and Peter Wil- is to overstate the cost of environmental coxen (1990a) measure this effect. Using regulations. Abatement expenditures a CGE model of the U. S. economy, they overstate the loss in firms' profits if firms estimate that mandated pollution con- can pass on part of their cost increase trols reduced the rate of GNP growth to consumers. Consumers in turn can by .I91 percentage points per annum avoid some of the welfare effects of price over the period 1973-85. increases of "dirty" goods by substituting "clean" goods for "dirty" ones. V. The Costs and Benefits of When environmental regulations affect Environmental Programs sectors, such as electricity production, that are important producers of interme- The value of a symptom-day or a statis- diate goods, it may be important to mea- tical life is, of course, only one compo- sure the impacts that environmental reg- nent in evaluating a pollution control ulations have throughout the economy. strategy. To translate unit benefit values Computable general equilibrium mod- into the benefits of an environmental pro- els, preferably those in which supply and gram requires three steps: (1) the emis- demand functions have been economet- sions reduction associated with the pro- rically estimated, may be needed to mea- gram must be related to changes in sure correctly the social costs of environ- ambient air or water quality; (2) the mental regulation. change in ambient environmental quality Michael Hazilla and Kopp (1990) have must be related to health or other used an econometrically estimated CGE outcomes through a dose-response func- model of the U.S. economy to compute tion; (3) the health or nonhealth out- the social costs of the Clean Air and comes must be valued. The informa- Clean Water Acts, as implemented in tion required for the first two tasks is 1981. The effects of these regulations on considerable, especially if one wants to firms are modeled as an upward shift in evaluate a major piece of legislation firms' cost functions, to which firms can such as the Clean Air Act or Clean Water adjust by altering their choice of inputs Act. and outputs. It is interesting to contrast In this section we review attempts to the estimates of social costs obtained estimate the benefits and costs of envi- from this approach with EPA's estimates ronmental programs. Of central interest of compliance costs. The EPA estimated are cases in which benefit-cost analyses the costs of complying with the Clean have actually been used in setting envi- Air and Clean Water Acts in 1981 to be ronmental standards; in addition, we dis- $42.5 billion (1981 dollars). Hazilla and cuss instances in which such analyses Kopp estimate the costs to be $28.3 bil- have not been used but should be. This lion; the lower figure reflects the substi- leads naturally to a discussion of priorities tution possibilities that the expenditure for research in the area of benefit and approach ignores. cost measurement. Cropper and Oates: Environmental Economics 723

A. The Use of Bene$t-Cost Analysis in matter, and carbon monoxide-effluent Setting Environmental Standards standards for water pollutants in the iron and steel and and plastics in- Executive Order 12291, signed in them-icals dustries, and regulations to ban lead in 1981, requires that benefit-cost analyses gasoline, as well as certain uses of be performed for all major regulations asbestos.56 Regulatory Impact Analyses (defined as those having annual costs in (RIAs) were prepared for 15 of these excess of $100 million). Furthermore, the rules. order requires, to the extent permitted In five of the RIAs, both benefits and by law, that regulations be undertaken costs were monetized; however, benefits only if the benefits to society exceed the could legally be compared with costs only costs. in the case of lead in gasoline. In this One consequence of Executive Order case, the benefits in terms of engine 12291 is the undertaking of benefit-cost maintenance alone were judged to ex- analyses for all major environmental reg- ceed the costs by $6.7 billion over the ulations; however, the extent to which period 1985-92, and the regulation was benefits and costs can be considered in issued. In two other cases-the PM stan- making regulations is limited by the en- dard and effluent limitations for iron and abling statutes. Of the major environ- steel plants-the benefits exceeded the mental statutes only two, the Toxic Sub- costs of the proposed regulation and the stances Control Act (TSCA) and the regulation was implemented, although Federal Insecticide, Fungicide, and Ro- EPA denied that it weighed benefits denticide Act (FIFRA) explicitly require against costs in reaching its decision. The that benefits and costs be weighed in set- remaining cases are more difficult to eval- ting standards." Some standards-spe- uate. The clean water benefits of pro- cifically, those pertaining to new sources posed effluent guidelines for chemicals under the Clean Air Act and to the set- and plastics manufacturers were judged ting of effluent limitations under the to exceed regulatory costs in some sec- Clean Water Act-allow costs to be taken tions of the country but not in others. into account, but do not suggest that ben- EPA recommended that these guidelines efits and costs be balanced at the margin. be implemented. Of several alternative In contrast, the National Ambient Air standards for emissions of particulate Quality Standards and regulations for the matter by surface coal mines, only one disposal of hazardous waste under RCRA was found to yield positive net benefits, and CERCLA are to be made without and these were small ($300,000). Eventu- regard to compliance costs. ally, no regulation was issued by EPA. In spite of these limitations, benefit- The preceding review suggests that cost analyses have been used in EPA's benefit-cost analysis has not entirely rulemaking process since 1981. Between been ignored in setting environmental February of 1981 and February of 1986, standards, but its use has been selective. EPA issued 18 major rules (USEPA In part, this is the result of law-EPA 1987), including reviews of National Am- was allowed to weigh benefits against bient Air Quality Standards for three pol- costs for only 5 of the 18 major regula- lutants-nitrogen dioxide, particulate tions that it issued between 1981 and

"Some portions of the Clean Air Act, specifically, sh A complete listing of the regulations may be those pertaining to aircraft emissions, motor vehicle found in USEPA (1987). Also included were regula- standards and fuel standards, also require that mar- tions governing the disposal of used oil, and standards ginal benefits and costs be balanced. regarding land disposal of hazardous waste. 724 Journal of Economic Literature, Vol. XXX (June 1992)

1986.'~One could argue that the govern- gests that the recreational use values as- ment should not invest resources in a sociated with the adoption of BPT are full blown benefit-cost analysis if the re- small, relative to the costs presented in sults of such an analysis cannot be used Table 1. Justification for these standards in regulating the polluting activity. But must then rest on other grounds. A sec- this would be a mistake. Even where the ond example where costs inay exceed explicit use of a benefit-cost test is pro- benefits involves the extent of cleanup hibited, such studies can be informative of Superfund sites under CERCLA. and useful. In their own way, they are While the cost of cleaning up these sites likely to influence the views of legislators is predicted to run into the hundreds of and regulators. In particular, the issue billions of dollars, the health benefits of is often one of amending standards-ei- these cleanups are thought by many to ther raising them or lowering thein. Ben- be inodest (Curtis Travis and Carolyn efit-cost information on such adjust- Doty 1989). Current law does not require ments, although not forinally admissible, an explicit benefit-cost analysis of reme- may well have some impact on decisions dial alternatives at each Superfund site, to revise standards. In addition, siinply but, in our view, it probably should. demonstrating the feasibility and poten- The second general class of cases in tial application of such studies may lead which careful benefit-cost analyses are to their explicit introduction into the pol- needed is where environmental stan- icy process at a later time. dards are sufficiently stringent to push control efforts onto the steep portion of B. The Need for Benefit-Cost Analyses of Environmental Standards the marginal cost of abatement curve. Even though the total costs of these stan- We turn now to a set of priorities for dards may exceed their total benefits (see benefit-cost analyses of environinental Figure 4), society might experience a regulation: which of existing environ- gain in welfare from relaxing the standard mental programs require closest scrutiny if the inarginal benefits of abatement are and what benefit techniques must be de- considerably below the inarginal costs at veloped in order to perforin these analy- the level of the standard. In terms of Fig- ses? We begin with an enumeration of ure 4, we need to know whether the mar- these programs, as we see them, and ginal benefit function is MB, or MB,. then offer some thoughts on the analysis There are several instances of actual poli- of each of thein. cies that appear to fall within this class: There are, broadly, two areas in which (1) the ground-level ozone standard, in careful benefit-cost analyses are most areas that are currently out of compliance needed. One is for statutes whose total with the standard; (2) certain provisions costs are thought to exceed their total in RCRA for disposal of hazardous waste; benefits. A widely cited example is the and (3)the 1990 acid rain amendments Clean Water Act (CWA), which will soon to the Clean Air Act. In addition to these be up for renewal. Freeman (1982) sug- existing laws, proposals for significant re- ductions in Coo emissions may entail high marginal costs, suggesting a close j7 For the other four regulations where a compari- son of costs and benefits was allowed-the three toxic scrutiny benefits. substances (TSCA) regulations and the setting of Turning first to the Clean Water Act, emission standards for light duty trucks-benefits we note that evaluating the CWA will were quantified but not monetized. In the case of PCB's the cost per catastrophe avoided was com- require the use puted; in the case of asbestos, the cost per life saved. and nonuse (ecological) benefits of im- Cropper and Oates: Environmental Economics

$ I

MB,

Environlnental 0 C Quality (Standard) Figure 4. Welfare Loss froin Setting Incorrect Standards proved water quality. As we noted above, there is no current route of exposure to one can either use a contingent valuation toxic waste. People could, however, be approach that captures both values, or exposed to contaminated soils or ground- one can attempt to capture use values water if substances were to leak from using travel cost methods and measure storage containers in the future. This in- nonuse values separately. Whichever ap- volves valuing future risks to persons cur- proach is used, we emphasize the re- rently alive as well as to persons yet un- gional character of the costs and benefits born. While some research has been of improved water quality; benefit esti- done in this area (Mauskopf 1987; Crop- mates must, in consequence, be available per and Portney 1990; Cropper and Suss- at this level of disaggregation. The con- man 1990), there are few empirical stud- tingent valuation method avoids two ies that examine either the value that problems inherent in the use of travel people place on reducing future risks to cost models. First, unless the transfera- themselves or the rate at which they dis- bility problem can be solved, travel cost count lives saved in future generations. models will have to be estimated for each Estimates of these values are also crucial river or lake throughout the U.S.! And, if one is to analyze regulations governing second, if a contingent valuation survey the current disposal of hazardous waste of nonuse values is to be added to travel under RCRA, as well as other regulations cost measures of use values, it may be that affect exposure to carcinogens (e.g., hard to get users to separate use from air toxics and pesticide regulations). nonuse values. An additional problem is how to incor- A key issue in valuing the benefits of porate uncertainty regarding estimates of Superfund cleanups is how to value health risks into the analysis. While most health risks-usually risks of cancer- valuation studies treat the probability of that will not occur until the distant fu- an adverse outcome as certain, in reality ture. Many Superfund sites pose very there is great uncertainty about health low health risks today, primarily because risks, especially the risk of contracting 726 Journal of Economic Literature, Vol. XXX (June 1992)

cancer from exposure to environmental reational, resulting from an increase in carcinogens. This uncertainty has two the pH of lakes." There are also likely sources: uncertainty about actual expo- to be visibility benefits (reduced haze) sures received, and uncertainty about in the Eastern U.S. This underscores the the effects of a given exposure.58 The need for better estimates of the value of standard procedure in risk assessments improved visibility, especially in urban is to "correct" for this uncertainty by areas. It will also be necessary to measure presenting a point estimate based on very the ecological benefits associated with re- conservative assumptions (Nichols and duced acid rain, especially as these are Richard Zeckhauser 1986). It would, likely to differ qualitatively from the eco- however, be more appropriate to incor- logical benefits associated with the CWA. porate the distribution of cancer risk into Finally, we note that in the area of the analysis. global climate change, considerable at- Existing estimates of the marginal costs tention has been devoted to measuring and marginal benefits of achieving the the costs of reducing greenhouse gas one-hour ozone standard in areas that are emissions, especially through the use of currently out of attainment suggest that a tax on the carbon content of fuels (Jor- marginal costs exceed marginal benefits genson and Wilcoxen 1990b). Little, (Krupnick and Portney 1991). Estimates however, is known about the benefits of of the health benefits of ozone control reducing greenhouse gases, even if one have, however, focused on the value of assumes that the link between CO, and reducing restricted-activity or symptom climate change is certain.60 days. There is some evidence that ozone The benefits of preventing these cli- may exacerbate the rate at which mate changes differ from the benefits as- lung tissue deteriorates, contributing sociated with conventional air and water to chronic obstructive lung disease pollutants in two respects. First, many- (COPD). Since, for healthy individuals, though by no means all-of the effects the probability of contracting COPD is of climate change are likely to occur uncertain, what must be valued is a through markets. These include effects change in the risk of contracting chronic on agriculture and forestry, as well as lung disease corresponding to a change changes in heating and cooling costs. in ozone concentrations. While this should make benefits easier The objective of the provisions of the to measure, the problem is that the ef- 1990 Amendments to the Clean Air Act fects of C02 emissions are not likely to aimed at reducing SO2 and NO2 is to be felt for decades. This implies that reduce acid rain, primarily in the Eastern valuing such damages is difficult. A dam- U.S. and Canada. Although the 10-mil- age function approach, which ignores ad- lion-ton reduction in sulfur emissions aptation possibilities, is clearly inappro- specified in the amendments is likely to priate; however, predicting technological have some health benefits, most of the possibilities for adaptation is not easy. anticipated benefits are ecological or rec- Second, the benefits of reducing greenhouse gases will not be felt until the next century. The problem here is Estimates of the effect of a given exposure usu- that, even at a discount rate of only 3 ally come from rodent bioassays, which are used to estimate a dose-response function. In addition to un- certainty regarding the parameters of the dose-re- sponse function, there is uncertainty as to how these For a dissenting view see Portney (1990). estimates should be extrapolated from rodents to 60 A useful beginning here is the work of William man. Nordhaus (1990). Cropper and Oates: Environmental Economics 727

percent, one dollar of benefits received ample, cleaner air in what was a rela- 100 years from now is worth only 5 cents tively dirty area may increase the de- today. This problem has typically been mand for residences there and drive up addressed by suggesting that benefits rents, thereby displacing low-income should be discounted at a very low rate, renters. All in all, this is a complicated if at all. An alternative approach is to issue. At any rate, Gianessi, Peskin, and make transfers to future generations to Wolff find that within urban areas the compensate them for our degradation of distribution of benefits may be slightly the environment, rather than to alter the pro-poor, but, as we shall see next, this discount rate. is likely to be offset (or more than offset) by a regressive pattern of the costs of C. The Distribution of Costs and these programs. Bene$ts We are on somewhat more solid In addition to examining the costs and ground on the distribution of the costs benefits of environmental legislation, it of environmental programs (G. B. Chris- is of interest to know who pays for pollu- tainsen and Tietenberg 1985). There ex- tion abatement and who benefits from ist data on the costs of pollution control it. Typically, studies of the distributional by industry with which one can estimate effects of environmental programs ein- how costs have influenced the prices of phasize the distribution of benefits and various classes of products and how, in costs by income class. turn, these increased prices have re- To determine how the benefits of envi- duced the real incomes of different in- ronmental programs are distributed come classes. In one early study of this across different income classes, we must kind, Gianessi, Peskin, and Wolff (1979) measure how the programs alter the examined the distributive pattern of the physical environments of different in- costs of the Clean Air Act and found that come groups. In one study of the distri- lower-income groups bear costs that con- butional effects of programs aimed at rais- stitute a larger fraction of their income ing the level of national air quality, than do higher-income classes. (See also Leonard Gianessi, Peskin, and Edward Nancy Dorfinan and Arthur Snow 1975; Wolff (1979) found striking locational dif- Gianessi and Peskin 1980.) Three inde- ferentials in benefits; not surprisingly, pendent studies of automobile pollution most of the benefits from efforts to im- control costs all reach similar findings of prove air quality are concentrated in the regressivity (Dorfman and Snow 1975; more industrialized urban areas (largely Harrison 1975; Freeman 197913). the heavily industrialized cities of the In a more recent study, Robison (1985) East) with fewer benefits accruing to ru- uses an input-output model to estimate ral residents. Even within metropolitan the distribution of costs of industrial pol- areas, air quality may differ substantially. lution abatement. Assuming that the Since the poor often live in the most pol- costs of pollution control in each industry luted parts of urban areas, they might are passed on in the form of higher be thought to be disproportionately large prices, Robison traces these price in- beneficiaries of programs that reduce air pollution-and there is evidence that this " Moreover, there is some persuasive evidence is, indeed, the case (Asch and Seneca from observed voting patterns on proposed environ- 1978; Jeffrey Zupan 1973). While this mental measures (Robert Deacon and Perry Shapiro 1975; Fischel 1979) indicating that higher income may be true, certain indirect effects can individuals are willing to pay more for a cleaner envi- follow that offset such benefits. For ex- ronment than those with lower incomes. 728 Journal of Economic Literature, Vol. XXX uune 1992)

creases through the input-output matrix grams on distributional grounds. At the to determine their impact on the pattern same time, there are opportunities to of consumer prices. Robison's model di- soften some of the more objectionable vides individuals into twenty income redistributive consequences of environ- classes. For each class, estimates are mental policies through the use of mea- available of the pattern of consumption sures like adjustment assistance for indi- among product groups. This information, viduals displaced from jobs in heavily together with predictions of price in- polluting industries and the reliance on creases for each product, is used to esti- the more progressive forms of taxation mate the increase in the prices of goods to finance public spending on pollution consumed by each income group. Robi- control programs. son finds that the incidence of control costs is quite regressive. Costs as a frac- VI. En~ironmentalEconomics and tion of income fall over the entire range Environmental Policy: Some Reflections of income classes; they vary from 0.76 percent of income for the lowest income As suggested by the lengthy (and only class to 0.16 percent of income for the partial) list of references and citations in highest income class. this survey, environmental economics It is true that these studies relate to has been a busy field over the past two existing environmental programs and do decades. Environmental economists not measure directly the potential distri- have reworked existing theory, making butional effects of a system of economic it more rigorous and clearing up a num- incentives such as effluent fees. But our ber of ambiguities; they have devised sense is that the pattern of control costs new methods for the valuation of benefits across industries would be roughly simi- from improved environmental quality; lar under existing and incentive-based and they have undertaken numerous em- programs. It is the same industries under pirical studies to measure the costs and both regimes that will have to undertake benefits of actual or proposed environ- the bulk of the abatement measures. Our mental programs and to assess the rela- conjecture thus is that the pattern of costs tive efficiency of incentive-based and for our major environmental programs is CAC policies. In short, the "intellectual likely to be distinctly regressive in its structure" of environmental economics incidence, be they of the command-and- has been both broadened and strength- control or incentive-based variety. ened since the last survey of the field While the distributional effects of envi- by Fisher and Peterson in this Journal ronmental programs may not be alto- in 1976. gether salutary, we do not wish to exag- But what about the contribution of en- gerate their importance. We emphasize vironmental economics to the design and that the primary purpose of environmen- implementation of environmental policy? tal programs is, in economic terms, an This is not an easy question to answer. efficient allocation of resources. Environ- We have seen some actual programs of mental measures, as Freeman (1972) has transferable emissions permits in the stressed, are not very well suited to the United States and some use of effluent achievement of redistributional objec- charges in Europe. And with the enact- tives. But an improved environment pro- ment of the 1990 Amendments to the vides important benefits for all income Clean Air Act, the U.S. has introduced classes-and we will be doing no groups a major program of tradable allowances a favor by opposing environmental pro- to control sulfur emissions-moving this Cropper and Oates: Environmental Economics 729

country squarely into the use of incen- nomic analysis and economic incentives tive-based approaches to regulation in at in environmental management? It is easy least one area of environmental to be pessimistic on this matter. There But, at the same time, effluent charge is still some aversion, both in the policy and marketable permit programs are few arena and across the general public, to in number and often bear only a modest the use of "market methods" for pollution resemblance to the pure programs of eco- control. While we were working on this nomic incentives supported by econo- survey, one of the leading news maga- mists. As we noted in the introduction. zines in the U.S. ran a lengthy feature certain major pieces of environmental story entitled "The Environment: Clean- legislation prohibit the use of economic ing Up Our Mess-What Works, What tests for the setting of standards for envi- Doesn't, and What We Must Do to Re- ronmental quality, while other directives claim our Air, Land, and Water" (Gregg require them! The record, in short, is a Easterbrook 1989, in Newsweek). A cen- mixed and somewhat confusing one: it tral argument in the article is that the reveals a policy environment character- attempt to place environmental policy on ized by a real ambivalence (and, in some a solid "scientific" footing has been a co- instances, an active hostility) to a central lossal error that has handcuffed efforts role for economics in environmental de- to get on with pollution control. Proceed- cision making. 63 ing "on the assumption that environmen- What is the potential and the likeli- tal protection is a social good transcend- hood of more attention to the use of eco- ing cost-benefit calculations" (p. 42), Easterbrook argues that we should not 62 under this provision, the U.S. will address the place a high priority on scientific work acid rain problem by cutbacks in sulfur emissions on the complicated issues of measuring over the next decade of 10 million tons (about a 50 percent reduction). This is to be accomplished benefits and costs and of providing care- through a system of tradable allowances under which fully designed systems of incentives, but affected power plants will be allowed to meet their should get on with enacting pollution emissions reductions by whatever means they choose--including the purchase of "excess" emissions control measures that are technologically reductions from other sources that choose to cut back feasible. In short, we should control what by more than their required quota. Also noteworthy technology enables us to control without is the U.S. procedure to implement reductions in chlorofluorocarbon emissions under the Montreal asking too many hard questions and hold- Protocol. Under this measure, EPA has effectively ing up tougher legislation until we know grandfathered the U. S. quota among existing produc- all the answers. ers and importers; from these baselines, firms are allowed to trade allowances (Hahn and McGartland Such a position has a certain pragmatic 1989). appeal. As we all know, our understand- 63 Some recent studies of actual environmental de- ing of complicated ecological systems and cision making are consistent with this "mixed" view. Magat, Krupnick, and Harrington (1986), for exam- the associated dose-response relation- ple, in a study of EPA determination of effluent stan- ships is seriously incomplete. And as our dards under the Clean Water Act Amendments of survey has indicated, our ability to place 1972, found that "simple rules based either on eco- nomic efficiency or the goal of distributional equity dollar values on improvements in envi- did not dominate the rulemaking process" (p. 154). ronmental quality is limited and impre- Their analysis did find that standards across industry cise. Nevertheless, we have some hard subcategories reflected to some extent differences in compliance costs among firms. In contrast, Cropper choices to make in the environmental et al. (1992) find that EPA decisions on pesticide arena-and whatever guidance we can regulation have, in fact, reflected a systematic balanc- obtain from a careful, if imprecise, con- ing of environmental risks and costs of control. Eco- nomic factors, it appears, have mattered in some sideration of benefits and costs should classes of decisions and not in others. not be ignored. Journal of Economic Litei-ature, Vol. XXX (June 1992)

We stress, moreover, that the role for the relevant areas under crude approxi- economic analysis in environmental pol- mations to demand curves (compensated icy making is far more important now or otherwise). In addition to measure- than in the earlier years of the "environ- ment issues, this new setting for environ- mental revolution." When we set out ini- mental policy places a much greater pre- tially to attack our major pollution prob- mium on the use of cost-effective lems, there were available a wide array regulatory devices, for the wastes associ- of fairly direct and inexpensive measures ated with the cruder forms of CAC poli- for pollution control. We were, in short, cies will be much magnified.64 operating on relatively low and flat seg- In spite of the mixed record, it is our ments of marginal abatement cost (MAC) sense that we are at a point in the evolu- curves. But things have changed. As tion of environmental policy at which the nearly all the cost studies reveal, mar- economics profession is in a very favor- ginal abatement cost functions have the able position to influence the course of typical textbook shape. They are low and policy. As we move into the 1990s, the fairly flat over some range and then begin general political and policy setting is one to rise, often quite rapidly. Both the first that is genuinely receptive to market ap- and second derivatives of these abate- proaches to solving our social problems. ment cost functions are positive-and Not only in the United States but in other rapidly increasing marginal abatement countries as well, the prevailing atmo- costs often set in with a vengeance. sphere is a conservative one with a strong We now find ourselves operating, in disposition toward the use of market in- most instances, along these rapidly rising centives, wherever possible, for the at- portions of MAC functions so that deci- tainment of our social objectives. More- sions to cut pollution yet further are be- over, as we have emphasized in this coming more costly. In such a setting, survey, we have learned a lot over the it is crucial that we have a clear sense past twenty years about the properties of the relative benefits and costs of alter- of various policy instruments and how native measures. It will be quite easy, they work (or do not work) under dif- for example, to enact new, more strin- ferent circumstances. Economists now gent regulations that impose large costs know more about environmental pol- on society, well in excess of the benefits, icy and are in a position to offer better health or otherwise, to the citizenry. As counsel on the design of measures for en- Portney (1990) has suggested, this may vironmental management. well be true of the new measures to con- This, as we have stressed, takes us trol urban air pollution and hazardous air from the abstract world of pure systems pollutants under the most recent Amend- of fees or marketable permits. Environ- ments to the Clean Air Act. Portney's mental economists must be (and, we be- admittedly rough estimates suggest that the likely range of benefits from these 64 Following our earlier discussion of the Weitzman theorem, we note its implication for the issue under new provisions falls well short of the discussion here: a preference for price over quantity likely range of their cost. instruments. So long as there is little evidence of Economic analysis can be quite helpful any dramatic threshold effects or other sources of rapid changes in marginal benefits from pollution in getting at least a rough sense of the control, the steepness of the MAC function suggests relative magnitudes at stake. This is not, that regulatory agencies can best protect against we would add, a matter of sophisticated costly error by adopting effluent fees rather than mar- ketable emission permits (Hadi Dowlatabadi and measures of "exact consumer surplus" Harrington 1989; Oates, Portney, and McGartland but simply of measuring as best we can 1989). Cropper and Oates: Environmental Economics 731

lieve, are) prepared to come to terms problems, to put it mildly. And they call with detailed, but important, matters of for an extension of existing work in the implementation: the determination of fee field to the development of an "open schedules, issues of spatial and temporal economy environmental economics" that variation in fees or allowable emissions incorporates explicitly the issues aris- under permits, the life of permits and ing in an international economy linked their treatment for tax purposes, rules by trade, financial, and environmental governing the transfer of pollution rights, flows. 65 procedures for the monitoring and en- forcement of emissions limitations, and so on. In short, economists must be ready ABDALLA,CHARLES."Measuring Economic Losses from Ground Water Contamination: An Investiga- to "get their hands dirty." tion of Household Avoidance Cost," Water Re- But the contribution to be made by sources Bulletin, June 1990, 26(3), pp. 451-63. environmental economists can be a valu- ADAMS,RICHARD M.; CROCKER,THOMAS D. AND KATZ, RICHARDW. "Assessing the Adequacy of able one. And there are encouraging Natural Science Information: A Bayesian Ap- signs in the policy arena of a growing proach," Rev. Econ. Statist., Nov. 1984, 66(4), pp. receptiveness to incentive-based ap- 568-75. ADAMS,RICHARD M.; HAMILTON,SCOTT A. AND proaches to environmental management. MCCARL,BRUCE A. "The Benefits of Pollution As we noted in the introduction, both Control: the Case of Ozone and U.S. Agriculture," in the United States and in the OECD Amer. J. Agr. Econ., Nov. 1986, 68(4), pp. 886 93. countries more generally, there have ADAR,ZVI AND GRIFFIN,JAMESM. "Uncertainty and been recent expressions of interest in the the Choice of Pollution Control Instruments," J. use of economic incentives for protection Enuiron. Econ. Manage., Oct. 1976, 3(3), pp. 178- 88. of the environment. As we were finishing ARROW,KENNETHJ. AND FISHER,ANTHONY C. "Envi- the final draft of this survey, the Council ronmental Preservation, Uncertainty, and Irrever- of the OECD issued a strong and lengthy sibility," Quart. J. Econ., May 1974, 88(2), pp. 312-19. endorsement of incentive-based ap- ASAKO,KAZUMI. "Environmental Pollution in an proaches, urging member countries to Open Economy," Econ. Rec., Dec. 1979, 55(151), "make a greater and more consistent use pp. 359-67. ASCH,PETER AND SENECA,JOSEPHJ. "MonopoIY and of economic instruments" for environ- External Cost: An Application of Second Best The- mental management (OECD 1991). ory to the Automobile Industry,"]. Environ. Econ. Finally, we note the growing aware- Manage., June 1976, 3(2), pp. 69-79. . "Some Evidence on the Distribution of Air ness and concern with global environ- Quality," Land Econ., Aug. 1978, 54(3), pp. 27% mental issues. Many pollutants display 97. a troublesome tendency to spill over na- ATKINSON,SCOTTE. AND LEWIS,DONALD H. "A Cost- Effectiveness Analysis of Alternative Air Quality tional boundaries. While this is surely Control Strategies," J. Environ. Econ. Manage., not a new issue (e.g., transnational acid Nov. 1974, 1(3), pp. 237-50. rain), the thinning of the ozone shield ATKINSON,SCOTTE. AND TIETENBERG,T.H. "The Empirical Properties of Two Classes of Designs and the prospect of global warming are for Transferable Discharge Permit Markets," J. En- pressing home in a more urgent way the uiron. Econ. Manage., June 1982, 9(2), pp. 101- need for a global perspective on the en- 21. BAILEY,MARTINJ. "Externalities, Rents, and Optimal vironment. The potential benefits and Rules." Sloan Working Paper in Urban Public Eco- costs of programs to address these issues, particularly global warming, are enor- "For a useful effort to develop a research perspec- mous-and they present a fundamental tive and agenda on the economic analysis of global policy challenge. The design and imple- change, see U.S. National Oceanic and Atmospheric Administration, National Science Foundation and mentation of workable and cost-effective National Aeronautics and Space Administration measures on a global scale are formidable (1991). Journal of Economic Litei -ature, Vol. XXX (June 1992)

nomics 1682, Economics Dept., U. of Maryland, KLING,CATHERINEL. "Estimating the Value of Wa- College Park, MA, 1982. ter Quality Improvements in a Recreational De- BARNETT,ANDYH. "The Pigouvian Tax Rule Under mand Framework," Water Resources Res., May Monopoly," Amer. Econ. Rev., Dec. 1980, 70(5), 1987, 23(5), pp. 951-60. pp. 103741. BOCKSTAEL,NANCY E. AND MCCONNELL,KENNETII BARTEL,ANNP. AND TAUBMAN,PAUL. "Health and E. "Welfare Measurement in the Household Pro- Labor Market Success: The Role of Various Dis- duction Framework," Amer. Econ. Rev., Sept. eases," Rev. Econ. Statist., Feb. 1979, 61(1), pp. 1983, 73(4), pp. 80614. 1-8. BOIIM, PETER. "Estimating Demand for Public BARTIK,TIMOTIIYJ. "Evaluating the Benefits of No11- Goods: An Experiment," European Econ. Rev., marginal Reductions in Pollution Using Informa- June 1972, 3(2), pp. 111-30. tion on Defensive Expenditures," J. Enuiron. . Deposit-refund systems: Theory and applica- Econ. Manage., March 1988a, 15(1), pp. 111-27. tions to environmental, conservation, and con- -. "Measuring the Benefits of Amenity Im- sunzer policy. Washington, DC: Johns Hopkins U. provements in Hedonic Price Models, Land Press for Resources for the Future, 1981. Econ., May 198813, 64(2), pp. 172-83. BOIIM,PETERAND RUSSELL,CLIFFORD F. "Compara- . "The Effects of Environmental Regulation tive Analysis of Alternative Policy Instruments," on Business Location in the United States," in Handbook of natural resource and energy eco- Growth Change, Summer 1988c, 19(3), pp. 22- nomics. Volume 1. Eds.: ALLENV. KNEESEAND 44. JAMESL. SWEENEY.Amsterdam: North-Holland, BAUMOL,WILLIAM J. "On Taxation and the Control 1985, pp. 395-460. of Externalities," Ainer. Econ. Rev., June 1972, BOLAND,JAMES"Ecollomic Instruments for Environ- 62(3), pp. 307-22. mental Protection in the United States," ENVI BAUMOL,WILLIAMJ. AND BRADFORD,DAVIDF. EC0186.14. Paris, France: Organisation for Eco- "Detrimental Externalities and Non-Convexity of nomic Cooperation and Development. Sept. 11, the Production Set," Economica, May 1972, 1986. 39(154), pp. 16C76. BOWER,BLAIRET AL. Incentives in water quality man- BAUMOL.WILLIAM1. AND OATES.WALLACE E. "The agement: France and the Ruhr area. Washington, Use of Standardiand Prices for Protection of the DC: Resources for the Future, 1981. Environment," Swedish J. Econ., March 1971, BOWKER,JAMES AND STOLL,JOHNR. "Use of Dichoto- 73(1), pp. 42-54. mous Choice Nonmarket Methods to Value the -. Economics, enuironinental policy, and qual- Whooping Crane Resource," Amer. J. Agr. Econ., ity of life. NY: Prentice-Hall, 1979. May 1988, 70(2), pp. 372-81. . The theory of environmental policy, Second BOYLE,KEVIN J. AND BISIIOP,RICIIARD C. "Valuing Edition. Cambridge, England: Cambridge U. Wildlife in Benefit-Cost Analyses: A Case Study Press, 1988. Involvillg Endangered Species," Water Resources BECKER,GARYS. "Crime and Punishment: A11 Eco- Res., May 1987, 23(5), pp. 943-50. nomic Approach," J. Polit. Econ., Mar.1Apr. 1968, BRADEN,JOIINB. AND KOLSTAD,CI~ARLES D., EDS. 76(2), pp. 169-217. Measuring the demand for environmental quality. BERGER,MARKC. ET AL."Valuing Changes in Health Amsterdam, North Holland, 1991. Risks: A Comparison of Alternative Measures," BRAIVIIIALL,D. F. AND MILLS,EDWIN S. "A Note Southern Econ. J., Apr. 1987, 53(4), pp. 967-84. on the Asymmetry between Fees and Pay- BINKLEY,CLARK S. AND HANEMANN,W. MICIIAEL. ments," Water Resources Res., 1966, 2(3), pp. 615 The recreation benejits of water quality improve- 16. ment: Analysis of day trips in an urban setting. BRESSERS,HANS."The Effectiveness of Dutch Water EPA-60015-78-010. Washington, DC: U.S. Envi- Quality Policy." Mimeo. The Netherlands: Twente ronmental Protection Agency, 1978. University of Technology, 1983. BISIIOP,RICIIARDC. AND HEBERLEIN,TIIOMAS A. BROMLEY,DANIEL W. "Markets and Externalities," "Measuring Values of Extramarket Goods: Are In- in Natural resource economics: Policy problems direct Measures Biased?'Amer. J. Agr. Econ., and contemporary analysis. Ed.: DANIELW. Dec. 1979, 61(5), pp. 92630. BROMLEY.Hingham, MA: Kluwer Nijhoff, 1986, BISIIOP,RICIIARD C.; HEBERLEIN,TIIOMAS A. AND pp. 37-68. KEALY,MARYJo. "Contingent Valuation of Envi- BROOKSIIIRE,DAVIDD. AND COURSEY,DONL. "Mea- ronmental Assets: Comparisons with a Simulated suring the Value of a Public Good: An Empirical Market," Natural Res. J., 1983, 23(3), pp. 619- Comparison of Elicitation Procedures," Amer. 34. Econ. Rev., Sept. 1987, 77(4), pp. 554-66. BLOMQUIST,GLENN C. "Value of Life Saving: Implica- BROOKSIIIRE,D. S. ET AL.Methods development for tions of Consumptioll Activity," J. Polit. Econ., assessing air pollution control benejits. Vol. 2: Ex- June 1979, 87(3), pp. 54C-58. periments in valuing non-market goods: A case BLOMQUIST,GLENNC.; BERGER,MARK C. AND study of alternative benefit measures of air pollu-

HOEIIN,TOIIN P. "New Estimates of Quality- . of Life tion in the south coast air basin of Southern Cali- in urban Areas," Amer. Econ. Reu., Mar. 1988, fornia. Prepared for the U.S. Environmental Pro-

7811)., .&A DD. 89-107. tection Agency, Washington, DC, 1979. BOCKSTAEL,NANCYE.;HANEMANN, W. MICIIAELAND BROWN,GARDNER, JR. AND BRESSERS,HANS."Evi- Cropper and Oates: Environmental Economics

dence Supporting Effluent Charges." Mimeo. The Runs: A System of Rental Emission Permits," Can. Netherlands: Twente University of Technology, J. Econ., May 1982, 15(2), pp. 347-54. Sept. 1986. COOPER,BARBARAS. AND RICE, DOROTHYP. "The BHO'IVN,GAHDNEK, Jn. AND JOIINSON,RALPIIW. "Pol- Economic Cost of Illness Revisited," Soc. Sec. lution Control by Effluent Charges: It Works in Bull., Feb. 1976, 39(2), pp. 21-36. the Federal Republic of Germany, Why Not in COKNES,RICHARDAND SANDLEH,TODD.The theory the U.S.," Natural Res. J., Oct. 1984, 24(4), pp. of externalities, public goods, and club goods. 929-66. Cambridge, Eng.: Cambridge U. Press, 1986. BROWN,GAKDNEH,Jn. AND MENDELSOIIN,ROBERT. COSTANZA,ROBERT AND PERRINGS,CHARLES. "A Flex- "The Hedonic Travel Cost Method," Rev. Econ. ible Assurance Bonding Svstem for Improved En- Statist., Aug. 1984, 66(3), pp. 427-33. vironmental ~ana~ement;"Ecological ^Economics, BROWN,JAMESN. AND ROSEN,HARVEY S. ''On the Feb. 1990. 2, DP. 57-75. & Estimation of Structural Hedonic Price Models," COURANT,PAUL N. AND PORTER,RICHARD. "Averting Econometrica, May 1982, 50(3), pp. 765-68. Expenditure and the Cost of Pollution," J. Enui- BROWN,TIIOMASC. AND SLOVIC,PAUL. "Effects of ron. Econ. Manage., Dec. 1981, 8(4), pp. 321- Context on Economic Measures of Value," in 29. Amenity resource valuation: Integrating economics COUKSEY,DONALD L.; HOVIS,JOHN L. AND SCHULZE, with other disciplines. Eds.: G. PETERSON,B. WILLIAMD. "The Disparity Between Willingness DRIVER,AND R. GREGORY.State College, PA: Ven- to Ac:ept and Willingness to Pay Measures of ture Publishing, Inc., 1988. Value, Quart. J. Econ., Aug. 1987, 102(3), pp. BUCIIANAN,JAMES M. "External Diseconomies, Cor- 679-90. rective Taxes, and Market Structure," Amer. Econ. CROPPER,MAUREENL. AND ARHIAGA-SALINAS,AMA- Rev., Mar. 1969, 59(1), pp. 174-77. LIA."Inter-citv Waee Differentials and the Value Bunnows, PAUL.The economic theory of pollution of Air ~ualiti,"J. 5rban Econ., 1980, 8(2), pp. control. Oxford: Martin Robertson, 1979. --.28654.-. Bunnows, PAUL."Controlling the Monopolistic Pol- CHOPPER,MAUREENL. ET AL. "The Determinants luter: Nihilism or Eclecticism?" J. Environ. Econ. of Pesticide Regulation: A Statistical Analysis of Manage., Dec. 1981, 8(4), pp. 372-80. EPA Decisionmaking," J. Polit. Econ., Feb. 1992, BUKKO'IVS,PAUL. "Nonconvexity Induced by External 100(1), pp. 175-97. Costs on Production: Theoretical Curio or Policy CROPPER,MAUREENL. AND KHUPNICK,ALAN J. "The Dilemma?"J. Environ. Econ. Manage., June 1986, Social Costs of Chronic Heart and Lung Disease." 13(2), pp. 101-28. Quality of the Environment Division Discussion BUTLER,RICIIAHDV. AND MAIIEK,MICIIAEL D. "The Paper QE89-16. Washington, DC: Resources for Control of Externalities in a Growing Urban Econ- the Future, 1989. omy," Econ. Inquiry, Jan. 1982, 20(1), pp. 155 CHOPPER,MAUREENL. AND POHTNEY,PAUL R. "Dis- 63. counting and the Evaluation of Lifesaving Pro- . "The Control of Externalities: Abatement grams," J. Risk Uncertainty, Dec. 1990, 3(4), pp. vs. Damage Prevention," Southern Econ. J., Apr. 369-79. 1986, 52(4), pp. 108&1102. CHOPPER,MAUREENL. AND SUSSMAN,FRANCES G. CAMERON,TRUDYANNAND JAMES,MICIIELLE D. "Ef- "Valuing Future Risks to Life," J. Environ. Econ. ficient Estimation Methods for 'Closed-Ended' Manage., Sept. 1990, 19(2), pp. 160-74. Contingent Valuation Surveys," Rev. Econ. Sta- CUMBERLAND,JOIINH. "Efficiency and Equity in In- tist., May 1987, 69(2), pp. 269-76. terregional Environmental Management," Rev. CARSON,RICIIAKDT. AND MITCIIELL,ROBERT CAM- Reg. Stud., Fall 1981, 10(2), pp. 1-9. ERON. "The Value of Clean Water: The Public's DALES,JOHN HAKKNESS. Pollution, property, and Willinglless to Pay for Boatable, Fishable, and prices. Toronto, Ont.: U. of Toronto Press, Swimmable Quality Water." Discussion Paper 88- 1968. 13. La Jolla, CA: U. of California at San Diego, DARDIS,RACIIEL. "The Value of a Life. New Evi- 1988. dence from the Marketplace," Amer. Econ. Rev., CIIESTNUT,LAURAINEAND RO'IVE,ROBERTD. "Eco- Dec. 1980, 70(5), pp. 1077-82. nomic Valuation of Changes in Visibility: A State D'ARGE,RALPHC. AND KNEESE,ALLENV. "Interna- of the Science Assessment for NAPAP" in Methods tional Trade, Domestic Income, and Environmen- for valuing acidic deposition and air pollution ef- tal Controls: Some Empirical Estimates," in Man- fects. Report 27, 1989, pp. 5-1-5-44. aging the environment: International economic CIIKISTAINSEN,G. B. AND TIETENBERG,TIIOMAS H. cooperation for pollution control. Eds.: ALLENV. "Distributional and Macroecollomic Aspects of En- KNEESE,SIDNEY E. ROLFE, AND JOSEPHW. vironmental Policy, in Handbook of natural re- HAKNED.NY: Praeger, 1971, pp. 289-315. source and energy economics. Vol. 1. Eds.: ALLEN DEACON,ROBERTT. AND SHAPIKO,PERKY."Private V. KNEESEAND JAMESL. SWEENEY.Amsterdam: Preference for Collective Goods Revealed Through North-Holland, 1985, pp. 34593. Voting on Referenda," Amer. Econ. Rev., Dec. COASE,RONALD H. "The Problem of Social Cost," 1975, 65(5), pp. 943-55. J. Law Econ., Oct. 1960, 3, pp. 1-44. DECISIONFOCUSINCOKPOKATED. Deuelop~nent and COLLINGE,ROBERTA. AND OATES,WALLACE E. "Effi- design of a contingent value survey for measuring ciency in Pollution Colltrol in the Short and Long the public's value for visibility improvements at 734 Journal of Economic Literature, Vol. XXX (June 1992)

the Grand Canyon National Park. Revised Draft HAINE. "The Value of Reducing Risks of Death: A Report, Los Altos, CA, Sept. 1990. Note on New Evidence," J. Policy Anal. Manage., DEWEES,DONALDN. "Instrument Choice in Envi- Winter 1989, 8(1), pp. 8&100. ronmental Policy," Econ. Inquiry, Jan. 1983,21(1), FISIIER,ANTIIONY C. Resource and environmental pp. 53-71. economics. Cambridge, Eng.: Cambridge U. DEWEES,DONALDN. AND SIMS,W. A. "The Symme- Press, 1981. try of Effluent Charges and Subsidies for Pollution FISHER,ANTHONYC. AND PETERSON,FREDERICK M. Control," Can.J. Econ., May 1976, 9(2), pp. 323- "The Environment in Economics: A Survey," J. 31. Econ. Lit., Mar. 1976, 14(1), pp. 1-33. DICKENS,WILLIAMT. "Differences Between Risk FREEMAN,A. MYRICK,111. "Distribution of Environ- Premiums in Union and Nonunion Wages and the mental Quality," in Environinental quality analy- Case for Occupational Safety Regulation," Amer. sis: Theory and method in the social sciences. Eds.: Econ. Rev., May 1984, 74(2), pp. 320-23. ALLENV. KNEESEAND BLAIRBOWER. Baltimore, DICKIE,MARK T. ET AL. "Reconciling Averting Be- MD: Johns Hopkins University Press for Resources havior and Contingent Valuation Benefit Estimates for the Future, 1972, pp. 243-78. of Reducing Symptoms of Ozone Exposure (draft)," . "On Estimating Air Pollution Control Bene- in Improving accuracy and reducing costs of envi- fits from Land Value Studies," J. Environ. Econ. ronmental benefit assessments. Washington, DC: Manage., May 1974, 1(1), pp. 74-83. U.S. Environmental Protection Agency, 1987. , The benefits of environmental improvement. DICKIE,MARK; FISHER, ANN AND GERKING,SIIELBY. Baltimore, MD: Johns Hopkins U. Press for Re- "Market Transactions and Hypothetical Demand sources for the Future, Inc., 1979a, p. 63. Data: A Comparative Study," J. Amer. Stat. As- -. "The Incidence of the Cost of Controlling soc., Mar. 1987, 82(397), pp. 69-75. Automobile Air Pollution," in The distribution of DICKIE,MARK AND GERKING,SHELBY."Willingness econoinic well-being. Ed.: F. TIIOMASJUSTER. to Pay for Ozone Control: Inferences From the Cambridge, MA: Ballinger, 1979b. Demand for Medical Care," J. Urban Econ., forth- -. Air and water pollution control: A benefit- coming. cost assessment. NY: John Wiley, 1982. DORFMAN,NANCY S., ASSISTED BY SNOW,ARTIIUR. GARCIA,PHILIPET AL. "Measuring the Benefits of "Who Will Pay for Pollution Control?-The Distri- Environmental Change Using a Duality Approach: bution by Income of the Burden of the National The Case of Ozone and Illinois Cash Grain Farms," Environmental Protection Program," Nat. Tax J., J. Environ. Econ. Manage., Mar. 1986, 13(1), pp. Mar. 1975, 28(1), pp. 101-15. 69-80. DOWLATABADI,HADIAND HARRINGTON,WINSTON. GEGAX,DOUGLAS;GERKING, SIIELBY AND SCIIULZE, "The Effects of Uncertainty on Policy Instruments: WILLIAM."Perceived Risk and the Marginal Value The Case of Electricity Supply and Environmental of Safety." Working paper prepared for the U.S. Regulations." Quality of the Environment Division Environmental Protection Agency, Aug. 1985. Discussion Paper QE89-20. Washington, DC: Re- GERKING,SIIELBY;DEHAAN, MENNO AND SCIIULZE, sources for the Future, 1989. WILLIAM."The Marginal Value of Job Safety: A DOWNING,PAUL B. AND WATSON,WILLIAM D., JR. Contingent Valuation Study," J. Risk Uncertainty, "The Economics of Enforcing Air Pollution Con- June 1988, 1(2), pp. 18599. trols," J. Enuiron. Econ. Manage., Nov. 1974, 1(3), GERKING,SHELBY AND STANLEY,LINDA R. "An Eco- pp. 219-36. nomic Analysis of Air Pollution and Health: The EASTERBROOK,GREGG."Cleaning Up," Newsweek, Case of St. Louis," Rev. Econ. Statist., Feb. 1986, July 24, 1989, 114, pp. 2M2. 68(1), pp. 11521. EPPLE,DENNIS. "Hedonic Prices and Implicit Mar- GIANESSI,LEONARDP.; PESKIN,HENRY M. AND kets: Estimating Demand and Supply Functions WOLFF,EDWARD N. "The Distributional Effects for Differentiated Products," J. Polit. Econ., Feb. of Uniform Air Pollution Policy in the United 1987, 95(1), pp. 59-80. States," Quart.J. Econ., May 1979, 93(2), pp. 281- on3 FEENBERG,DANIEL AND MILLS,EDWIN S. Measuring JUl. the benefits of water pollution abatement. NY: Aca- GIANESSI,LEONARD P. AND PESKIN,HENRY M. "The demic Press. 1980. Distribution of the Costs of Federal Water Pollu- FISCHEL,WILLIAM A. "Fiscal and Environmental tion Control Policy," Land Econ., Feb. 1980, 56(1), Considerations in the Location of Firms in Subur- pp. 85-102. ban Communities," in Fiscal zoning and land use GOULD,J. R. "Total Conditions in the Analysis of controls. Eds.: EDWINS. MILLSAND WALLACEE. External Effects," Econ. J., Sept. 1977, 87(347), OATES.Lexington, MA: Heath, 1975, pp. 119-73. pp. 558-64. . "Determinants of Voting on Environmental GRIGALUNAS,TIIOMAS A. AND OPALUCEI,JAMES J. "AS- Quality: A Study of a New Hampshire Pulp Mill sessing Liability for Damages Under CERCLA: A Referendum," J. Environ. Econ. Manage., June New Approach for Providing Incentives for Pollu- 1979, 6(2), pp. 107-18. tion Avoidance?" Natural Res. J., Summer 1988, FISHELSON,GIDEON. "Emission Control Policies U11- 28(3), pp. 509-33. der Uncertainty," J. Environ. Econ. Manage., Oct. HAIIN,ROBERTW. "Market Power and Transferable 1976, 3, pp. 189-97. Property Rights," Quart. J. Econ., Nov. 1984, FISHER,ANN; VIOLETTE, DAN AND CIIESTNIJT,LAU- 99(4), pp. 753-65. Cropper and Oates: Environmental Economics 735

. "Economic Prescriptions for Environmental entissions standards. Cambridge, MA: Ballinger, Problems: How the Patient Followed the Doctor's 1975. Orders," J. Econ. Perspectiues, Spring 1989, 3(2), HARRISON,DAVIDH. "The Regulation of Aircraft pp. 95-114. Noise," in Incentioes for enuironmental protection. . "The Political Economv of Environmental Ed.: TIIOMASC. SCIIELLING.Cambridge, MA: MIT Regulation: Towards a unifying Framework," Pub- Press, 1983, pp. 41-143. lic Choice, Apr. 1990, 65(1), pp. 21-47. HAZILLA,MICIIAELAND KOPP, RAYMONDJ. "Social HAHN,ROBERTW. AND HESTER,GORDONL. "Mar- Cost of Environmental Quality Regulations: A ketable Permits: Lessons for Theory and Practice," General Equilibrium Analysis," J. Polit. Econ., Ecology Law Quarterly, 1989a, 16(2),pp. 361-406. Aug. 1990, 98(4),pp. 853-73. -. "Where Did All the Markets Go? An Analysis HOEIIN,JOIINP. AND RANDALL,ALAN."TOO Many of EPA's Emissions Trading Program," YaleJ. Re- Proposals Pass the Benefit Cost Test,"Amer. Econ. gul., Winter 1989b, 6(1), pp. 109-53. Rec., 1989, 789(3),pp. 544-51. HAIIN,ROBERT W. AND MCGARTLAND,ALBERT M. HOTELLING,HAROLD."The Economics of Exhaust- "The Political Economy of Instrument Choice: An ible Resources," J. Polit. Econ., Apr. 1931, 39(2), Examination of the U. S. Role in Implementing the pp. 137-75. Montreal Protocol," Northwestern U. Law Rec., JOIINSON,RALPIIW. AND BHO\\~N,GARDNER M., JR. Spring 1989, 83(3),pp. 592-611. Cleaning up Europe's waters. NY: Praeger, 1976. HAIIN,ROBERTW. AND NOLL,ROGER G. "Designing JONES-LEE,MICIIAELW. "The Value of Changes in a Market for Tradable Emission Permits," in Re- the Probability of Death or I~ljury,"]. Polit. Econ., form of environmental regulation. Ed.: WESLEY JulyIAug. 1974, 82(4),pp. 835-49. MAGAT.Cambridge, MA: Ballinger, 1982,pp. 119- JONES-LEE,MICIIAELW.; HAIIMERTON,M. AND 46. PIIILIPS,P. R. "The Value of Safety: Results of a HAMILTON,JONATIIAN H.; SIIESIIINSKI,EYTAN AND National Sample Survey," Econ. J., Mar. 1985, SLUTSKY,STEVEN M. "Production Externalities and 95(377),pp. 49-72. Long-Run Equilibria: Bargaining and Pigovian Tax- JORGENSON,DALE W. AND WILCOXEN,PETER J. "En- ation," Econ. Inquiry, July 1989, 27(3), pp. 453- vironmental Regulation and U. S. Economic 7 -11. Growth," Rand J. Econ., Summer 1990a, 21(2), HAMILTON,SCOT A,; MCCARL,BRUCE A. AND AD- pp. 31440. AMS, RICIIARDM. "The Effect of Aggregate Re- -. "Reducing U.S. Carbon Dioxide Emissions: sponse Assumptions on E~lvironmental Impact The Cost of Different Goals." Harvard U., 1990b. Analyses," Amer. J. Agr. Econ., May 1985, 67(2), KAIINEMAN,DANIEL; KNETSCII, JACK L. AND TIIALER, pp. 407-13. RICIIARDH. "Experimental Tests of the Endow- HAMMACK,JUDD AND BROWN,GARDNER M., JR.Wa- ment Effect and the Coase Theorem," J. Polit. terfowl and wetlands: Toward bioeconolnic analy- Econ., Dec. 1990, 98(6),pp. 1325-48. sis. Baltimore, MD: Johns Hopkins U. Press for KAIINEMAN,DANIELAND TVERSKY,AMOS."Prospect Resources for the Future, 1974. Theory: An Analysis of Decisions Under Risk," HANEMANN,W. MICIIAEL."Di~creteIC~~~ti~~u~us Econometrica, Mar. 1979, 47(2), pp. 263-91. Models of Consumer Demand," Econometrica, KAMIEN,MORTON I.; SCIIWARTZ,NANCY L. AND DOL- May 1984, 52(3),pp. 541-61. BEAR,F. T. "Asymmetry between Bribes and HANEMANN,W. MICEIAEL."Willingness to Pay and Charges," Water Resources Res., 1966, 2(1), pp. Willingness to Accept: How Much Can They Dif- 147-57. fer?" Amer. Econ. Reu., June 1991, 81 (3),pp. 635- KANE,SALLY; REILLY, JOIINAND TOBEY,JAMES."An 47 Empirical Study of the Economic Effects of Cli- HARFORD,JOND. "Firm Behavior Under Imperfectly mate Change on World Agriculture." Mirneo Enforceable Pollution Standards and Taxes," J. En- USDA Report, Jan. 1991. viron. Econ. Manage., Mar. 1978, 5(l), pp. 26- KANEMOTO,YOSIIITSUGU. "Hedonic Prices and the 43. Benefits of Public Prices," Econon~etrica, July HARRINGTON,WINSTON."Enforceme~lt Leverage 1988, 56(4),pp. 981-90. When Penalties Are Restricted," J. Public Econ., KEALY,MARYJo; DOVIDIO,JACK AND ROCKEL,MARK Oct. 1988, 37(1),pp. 2!3-53. L. "Willingness to Pay to Prevent Additional Dam- HARRINGTON,WINSTON; KRUPNICK, ALAN J. AND PES- ages to the Adirondacks from Acid Rain," Reg. Sci. KIN,HENRY M. "Policies for Nonpoint-Source Wa- Reo., 1987, 15, pp. 11M1. ter Pollution Contro1,"J. Soil and Water Conserca- KELMAN,STEVENJ. What price incentices? Econo- tion, Jan.-Feb. 1985, 40, pp. 27-32. lnists and the encironn~ent.Boston, MA: Auburn HARRINGTON,WINSTON;KRUPNICK, ALAN J. AND House, 1981. SPOFFORD,WALTER O., JR. "The Economic Losses KNEESE,ALLENV.; AYRES,ROBERT V. AND D'ARGE, of a Waterborne Disease Outbreak," J. Urban RALPIIC. Econontics and the environment: A mate- Econ., Jan. 1989, 25(1),pp. 11637. rials balance approach. Washington, DC: Re- HAHRINGTON,WINSTON AND PORTNEY,PAUL R. "Valu- sources for the Future, 1970. ing the Benefits of Health and Safety Regulations," KNEESE,ALLENV. AND BOWEH,BLAIRT. Managing J. Urban Econ., July 1987, 22(1), pp. 101-12. water qztality: Economics, technology, institutions. HARRISON,DAVID,JR. Who pays for clean air: The Baltimore, MD: Johns Hopkins University Press cost and benefit distribution of federal auton~obile for Resources for the Future, 1968. 736 Journal of Economic Literature, Vol. XXX uune 2992)

KNEESE,ALLENV. AND SCHULZE.WILL.IAA~ D. "Eth- LOEIIMAN.EDNAT. ET AL. "Distributional Analysis ics and Environmental Economics," in Handbook of Regional Benefits and Cost of Air Quality Con- of natural resource and energy economics. Vol. trol,"J. Environ. Econ. Manage., Sept. 1979, 6(3), 1.Eds.: ALLENV. KNEESEAND JAMESL. SWEENEY. pp. 222-43. Amsterdam: North-Holland, 1985, pp. 191-220. LOEIIMAN,EDNA; BOLL>T, D. AND CIIAIKIN,K. "Mea- KNETSCII,JACKL. AND SINDEN,J. A. "Willingness suring the Benefits of Air Quality Improvements to Pay and Compensation Demanded: Experimen- in the San Francisco Bay Area." Prepared for the tal Evidence of an Unexpected Disparity in Mea- U.S. Environmental Protection Agency by SRI In- sures of Value," Quatt J. Econ., Aug. 1984, 99(3), ternational, Menlo Park, CA, 1981. pp. 507-21. LYON,RANDOLPHM. "Auctions and Alternative Pro- KOHN,ROBERT E. "A General Equilibrium Analysis cedures for Allocating Pollution Rights," Land of the Optimal Number of Firms in a Polluting Econ., Feb. 1982, 58(l), pp. 1632. Industry," Can. J. Econ., May 1985, 18(2), pp. MACFARLAND,K. K. ET AL. "An Examination of 347-54. Methodologies and Social Indicators for Assessing KOLSTAD,CIIARLES D. "Empirical Properties of Eco- the Value of Visibility," in Managing air quality nomic Incentives and Command-and-Control Reg- and scenic resources at national parks and wilder- ulations for Air Pollt~tionControl:" Land Econ., ness areas. Eds.: ROBERTD. ROWEAND LAURAINE Aug. 1986, 62(3), pp. 25C68. G. CIIESTNUT.Boulder, CO: Westview Press, . "Uniformity versus Differentiation in Regu- 1983. lating Externalities," ,I. Environ. Econ. Manage., MAGAT,WESLEYA. "Pollution Control and Techno- Dec. 1987, 14(4), pp. 38699. logical Advance: A Dynamic Model of the Firm," Koo, ANT~IONYY. C. "Environmental Repercussions J. Enciron. Econ. Manage., Mar. 1978, 5(l), pp. and Trade Theory," Rev. Econ. Statist., May 1974, 1-25. 56(2), pp. 235-44. MAGAT,WESLEYA,; KRUPNICK,ALAN J. AND HAR- KoP,, RAYMONDJ. AND KRIJPNICK,ALAN J. "Agricul- RINGTON,WINSTON. Rules in the making: A statisti- tural Policy and the Benefits of Ozone Control," cal analysis of regulatory agency behavior. Wash- Amer. J. Agr. Econ., Dec. 1987, 69(5), pp. 956 ington, DC: Resources for the Future, 1986. 62. MLLER,KARL-G~RAN. Encironmental economics: A KOPP, RAYMONDJ. ET AL. "Implications of Environ- theoretical inquiry. Baltimore, MD: Johns Hop- mental Policy for U.S. .4griculture: The Case of kins U. Press for Resources for the Future, 1974. .4rnbient Ozone Standards," J. Enciron. Manag., -. "Welfare Economics and the Environment." June 1985, 20(4). pp. 321-31. in Handbook of natural resource and energy eco- KRUPNICK,ALAN J. "Costs of Alternative Policies for nomics. Vol. 1. Eds.: ALLENKNEESE AND JAMES the Control of NO2 in the Baltimore Region." Un- SWEENEY.Amsterdam: North-Holland, 1985, pp. pub. working paper. Washington, DC: Resources 3-60. for the Future, 1983. MALIK,ARUN."Markets for Pollution Control When KRUPNICK,ALANJ.; OATES,WALLACE E. AND VAN Firms Are Noncompliant," J. Environ. Econ. Man- DE VERG,ERIC. "On Marketable Air-Pollution Per- age., Mar. 1990, 18(2, Part l), pp. 97-106. mits: The Case for a System of Pollution Offsets," MAUSKOPF,JOSEPIIINEA. "Projections of Cancer J. Environ. Econ. !Manage., Sept. 1983, 10(3), pp. Risks Attributable to Future Exposure to Asbes- 233-47. tos," Risk Analysis, 1987, 7(4), pp. 477-86. KRUPNICK,ALAN J. AND PORTNEY,PAUL R. "Control- MCCONNELL,VIRGINIA D. AND SCIIWAB,ROBERT M. ling Urban Air Pollution: A Benefit-Cost Assess- "The Impact of Environmental Regulation in In- ment," Science, Apr. 26, 1991, 252, pp. 522-28. dustry Location Decisions: The Motor Vehicle In- KRUTILLA,JOIIN\'. "Conservation Reconsidered," dustry," Land Econ., Feb. 1990, 66, pp. 67-81, Anzer. Econ. Rec., Sept. 1967, 57(4), pp. 777-86. ~~cCUUBINS,MATTIIEWD.: XOLL. ROGERG. AND KWEREL,E\.ANR. "To Tell the Truth: Imperfect In- WEINGAST,BARRY R. "Structure and Process, Polit- formation and Optimal Pollution Control," Rev. ics and Policy: Administrative Arrangements and Econ. Stud., Oct. 1977, 44(3), pp. 595601. the Political Control of Agencies," Virginia Law LEE, DLVIGIITR. "Efficiency of Pollution Taxation Rev., Mar. 1989, 75(2), pp. 431-82. and Market Structure," J. Enciron. Econ. Man- MCGAIITLAND,ALBERTM. Marketable periilit sys- age., Sept. 1975, 2(1), pp. 69-72. tem~for air pollution control: An einpirical study. LEE, DWIGIITR. AND MISIOLEK,WALTER S. "Substi- Ph.D. Dissertation, U. of Maryland, College Park, tuting Pollution Taxation for General Taxation: ISID,1984. Some Implications for Efficiency in Pollution Taxa- -. "A Comparison of Two Marketable Dis- tion," J. Enoiron. Econ. Manage., Dec. 1986, charge Permits Systems," J. Environ. Econ. Man- 13(4), pp. 338-47. age., Mar. 1988, 15(1), pp. 35-44. LEONARD,H. JEFFREY.Pollzction and the struggle MCGAHTLAND,ALBERT hi. 4ND OATES,WALLACE E. for the world product. Cambridge, Eng. Cam- "Marketable Permits for the Prevention of Envi- bridge U. Press, 1988. ronmental Deterioration," J. Environ. Econ. Man- LICIITENBERG,ERIK AND ZILBEKA~AN,DAVID. "The age., Sept. 1985, 12(3), pp. 207-28. Welfare Economics of Price Supports in U. S. Agri- MCGUIRE,MARTIN. "Regulation, Factor Rewards, culture," Anler. Econ. Rev., Dec. 1986, 76'(5), pp. and International Trade," J. Public Econ., Apr. 1135-41. 1982, 17(3), pp. 335-54. Cropper and Oates: Environmental Economics

MENDELSOHN,ROBERT. "Estimating the Structural NORDHAUS,WILLIAM D. "To Slow or Not to Slow: Equations of Implicit Markets and Household Pro- The Economics of the Greenhouse Effect." Cowles duction Functions," Rev. Econ. Statist., Nov. Foundation Discussion Paper, 1990. 1984, 66(4), pp. 673-77. OATES,WALLACEE. "The Regulation of Externalities: -. -. "Regulating Heterogeneous Emissions," J. Efficient Behavior by Sources and Victims," Public Environ. Econ. Manage., Dec. 1986, 13(4), pp. Finance, 1983, 38(3), pp. 362-75. 301-12. OATES, WALLACEE.; PORTNEY,PAUL R. AND -. -. "Modeling the Demand for Outdoor Recre- MCGARTLAND,ALBERT M. "The Net Benefits of ation," Water Resources Res., May 1987, 23(5), Incentive-based Regulation: A Case Study of Envi- pp. 961-67. ronmental Standard Setting," Amer. Econ. Rev., MERRIFIELD,JOHN D. "The Impact of Selected Dec. 1989, 79(5), pp. 123-2. Abatement Strategies on Transnational Pollution, OATES,WALLACEE. AND SCHWAB,ROBERT M. "Eco- the Terms of Trade, and Factor Rewards: A Gen- nomic Competition Among Jurisdictions: Effi- eral Equilibrium Approach," J. Environ. Econ. ciency Enhancing or Distortion Inducting?"]. Pub- Manage., Sept. 1988, 15(3), pp. 259-84. lic Econ., Apr. 1988a. 35(3), pp. 33S54. MESTELMAN,STUART. "Production Externalities and OATES,WALLACEE. ANDSCHWAB, ROBERT M. "The Corrective Subsidies: A General Equilibrium Anal- Theory of Regulatory Federalism: The Case of En- ysis," ]. Environ. Econ. Manage., June 1982, 9(2), vironmental Management." Working Paper No. pp. 18693. 88-26, Dept. of Economics. U. of Maryland, MILLIMAN,SCOTTR. AND PRINCE,RAYMOND. "Firm 1988b. Incentives to Promote Technological Change in OATES,WALLACE E. AND STRASSMANN., DIANAL. Pollution Control," J. Environ. Econ. Manage., "Effluent Fees and Market Structure," J. Public Nov. 1989, 17(3), pp. 247-65. Econ., June 1984, 24(1), pp. 29-46. MINTZ,JACKM. AND TULKENS,HENRY. "Commodity OHSFELDT,ROBERTL. "Assessing the Accuracy of Tax Competition between Member States of a Fed- Structural Parameter Estimates in Analyses of Im- eration: Equilibrium and Efficiency," ]. Public plicit Markets," Land Econ., May 1988, 64(2), pp. Econ., Mar. 1986, 29(2), pp. 133-72. 135-46. MISIOLEK,WALTERS. "Effluent Taxation in Monop- OHSFELDT,ROBERTL. AND SMITH,BARTON A. Esti- oly Markets," J. Environ. Econ. Manage., June mating the Demand for Heterogeneous Goods," 1980, 7(2), pp. 103-07. Rev. Ecm. Statist., Feb. 1985, 67(1), pp. 165-71. MITCHELL,ROBERTC. AND CARSON,RICHARDT. O'NEIL,WILLIAM B. "The Regulation of Water Pollu- "Valuing Drinking Water Risk Reductions Using tion Permit Trading Under Conditions of Varying the Contingent Valuation Methods: A Methodolog- Streamflow and Temperature," in Buying a better ical Study of Risks from THM and Giardia." Paper environment: Cost-effective regulation through prepared for Resources for the Future, Washing- permit trading. Eds. ERHARDF. JOERESAND MAR- ton, DC, 1986. TIN H. DAVID.Madison, WI: U. of Wisconsin . Using surveys to value public goods: The Press, 1983, pp. 219-31. contingent valuation inethod. Washington, DC: O'NEIL, WILLIAMET AL. "Transferable Discharge Resources for the Future, 1989. Permits and Economic Efficiency: The Fox River," MJELDE,JAMES W. ET AL. "Using Farmers' Actions J. Environ. Econ. Manage., Dec. 1983, 10(94), pp. to Measure Crop Loss Due to Air Pollution," J. 34655. Air Pollution Control Association, Apr. 1984, 34(4), OPALUCH,JAMESJ. AND GRIGALUNAS,THOMAS A. pp. 360-64. "Controlling Stochastic Pollution Events through MONTGOMERY,W. DAVID."Markets in Licenses and Liability Rules: Some Evidence from OCS Leas- Efficient Pollution Control Programs," J. Econ. ing," Rand ]. Econ., Spring 1984, 15(1), pp. 142- Theory, Dec. 1972, 5(3), pp. 395-418. 51. MOREY,EDWARD R. "The Choice of Ski Areas: A ORCAN~ZATIONFOR ECONOMICCOOPERATION AND DE- Generalized CES Preference Ordering with Char- VELOPMENT. The application of econoniic instru- acteristics," Rev. Econ. Statist., Nov. 1984, 66(4), inents for environn~ental protection. Paris: pp. 584-90. O.E.C.D., 1989. MULLEN,JOHNK. AND MENZ,FREDRIC C. "The Ef- . Reconi~nendationof the council on the use fect of Acidification Damages on the Economic of econoniic instruments in environmental policy. Value of the Adirondack Fishery to New York An- Paris: O.E.C.D, Jan. 1991. glers," Ainer. ]. Agr. Econ., Feb. 1985, 67(1), pp. PALMER,ADELER. ET AL. "Economic Implications 112-19. of Regulating Chlorofluorocarbon Emissions from MUMY,GENEE. "Long-Run Efficiency and Property Nonaerosol Applications." Report R-2524-EPA. Rights Sharing for Pollution Control," Public Santa Monica, CA: Rand Corp., 1980. Choice, 1980, 35(1), pp. 59-74. PALIIQUIST,RAYMONDB. "Estimating the Demand NICHOLS,ALBERTL. Targeting econoinic incentives for the Characteristics of Housing," Rev. Econ. for environniental protection. Cambridge, MA: Statist., Aug. 1984, 66(3), pp. 394-404. MIT Press, 1984. . "Pollution Subsidies and Multiple Local Op- NICHOLS,ALBERTL. AND ZECKHAUSER,RICHARD J. tima," Land Econ., Nov. 1990, 66, pp. 394 "The Perils of Prudence," Regulation, Nov./Dec. 401. 1986, 10(2), pp. 13-24. PARSONS,GEORGE R. AND KEALY,MARY Jo. "Measur- 738 Journal of Economic Literature, Vol. XXX (June lYY.2)

ing Water Quality Benefits Using a Random Utility the Future for Pacific Northwest Laboratory, Oct. Model of Lake Recreation in Wisconsin." Working 1990. Paper No. 90-14, Dept. of Economics, U. of Dela- Rowe, ROBERTD.; D'ARGE,~LPII C. AND BROOK- ware, 1990. SIIIRE,DAVID S. "An Experiment on Economic PELEMAN,SAMAND TIDEMAN,T. NICOLAUS."Local Value of Visibility," J. Environ. Econ. Manage., versus National Pollution Control: Note," An~er. Mar. 1980, 7(1), pp. 1-9. Econ. Rev., Dec. 1972, 62(5), pp. 95!3-63. RUSSELL,CLIFFORDS. "Monitoring and Enforce- PETIIIG,RUDIGER. "Pollution, Welfare, and Environ- ment," in Public policiesfor encironmental protec- mental Policy in the Theory of Comparative Advan- tion. Ed.: PAULPORTNEY. Washington, DC: Re- tage,"J. Environ. Econ. Manage., Feb. 1976, 2(3), sources for the Future, 1990, pp. 24S74. pp. 160-69. RUSSELL,CLIFFORDS.; HARRINGTON:WINSTON AND PETIIIG,RUDIGER ET AL. Trade and Environment: A VAUGIIAN,WILLIAM J. Enforcing pollution control Theoretical Inquiry. Amsterdam: Elsevier, 1980. laws. Washington DC: Resources for the Future, PEZZEY,JOHN. "Changes versus Subsidies versus 1986 Marketable Permits as Efficient and Acceptable SAMPLES,KAHLC.; DIXON,JOIIN A. AND GOWEN, Methods of Effluent Controls: A Property Rights MARCIAM. "Information Disclosure and Endan- Analysis." Unpub. paper, 1990. gered Species Valuation," Land Econ., Aug. 1986, PORTES,RICHARD D. "The Search for Efficiency in 62(3), pp. 306-12. the Presence of Externalities," in Unfashionable SCIIERR,BRUCEA. AND BABB,EMERSON M. "Pricing economics: Essays in honor of Lord Balogh. Ed.: Public Goods: An Experiment with Two Proposed PAULSTREETEN. London: Weidenfeld and Nichol- Pricing Systems," Public Choice, Fall 1975, 23(3), son, 1970, pp. 348-61. pp. 35-48. PORTNEY,PAULR. "Policy Watch: Economics and SCIIULZE,WILLIAMD. AND BROOKSIIIRE,DAVIDS. the Clean Air Act," J. Econ. Perspectives, Fall "The Economic Benefits of Preserving Visibility 1990, 4(4), pp. 173-81. in the National Parklands of the Southwest," Natu- h~,DOUGLAS A. "Benefits of Visual Air Quality in ral Res. J., Jan. 1983, 23(1), pp. 149-73. Cincinnati: Results of a Contingent Ranking Sur- SCIIULZE,WILLIAMD. AND D'ARGE,RALPH C. "The vey." RP-1742, final report prepared by Charles Coase Proposition, Information Constraints, and River Associates for Electric Power Research Insti- Long-Run Equilibrium," Anier. Econ. Rec., Sept. tute, Palo Alto, CA, 1984.. 1974, 64(4), pp. 76S72. RANDALL,ALAN; IVES, BERRY AND EASTMAN,CLYDE. SEGERSON,KATHLEEN. "Uncertainty and Incentives "Bidding Games for Valuation of Aesthetic Envi- for Nonpoint Pollution Control," J. Enuiron. Econ. ronmental Improvements," J. Environ. Econ. Manage., Mar. 1988, 15(1), pp. 87-98. Manage., Aug. 1974, 1(2), pp. 132-49. -"Institutional 'Markets': The Role of Liability ~UFER,ROGERK. AND FELDIIAN,STEPHENL. Acid in Allocating Environ~nentalResources," in Pro- rain and einissions trading. Totowa, NJ: Rowman ceedings of AERE workshop on natural resource & Littlefield, 1987. market meckanisn~s.Association of Environmental RIDKER,RONALD G. AND HENNING,JOIINA. "The and Resource Economists, June 1990. Determinants of Residential Propert2 Values with SESKIN,EUGENEP.; ANDERSON,ROBERT J., JR. AND Special Reference to Air Pollution, Rev. Econ. REID,ROBERT 0. "An Empirical Analysis of Eco- Statist., May 1967, 49(2), pp. 24657 nomic Strategies for Controlline: Air Pollution." I. KOBACK,JENNIFER."Wages, Rents, and the Quality Environ. con. Manage., June 1983, 10(2), pi. of Life," J. Polit. Econ., Dec. 1982, 90(6), pp. 112-24. 1257-78. SHAVELL,STEVEN."A Model of the Optimal Use of ROBERTS,MARCJ. AND SPENCE,MICEIAEL. "Effluent Liability and Safety Regulations," Rand J. Econ., Charges and Licenses Under Uncertainty," J. Pub- Summer 1984a, 15(2), pp. 271-80. lic Econ., Apr./May 1976, 5(3)(4), pp. 19S208. -"Liability for Harm versus Regulation of ROBISON,H. DAVID."Who Pays for Industrial Pollu- Safety," J. Legal Stud., 1984b, 13(2), pp. 357- tion Abatement?" Rev. Econ. Statist., Nov. 1985, 74. 67(4), pp. 702-06. SIIIBATA,HIROFUMI AND WINRICII,J. STEVEN."Con- ROSEN,SHERWIN."Hedonic Prices and Implicit Mar- trol of Pollution When the Offended Defend Them- kets: Product Differentiation in Pure Competi- selves," Econoniica, Nov. 1983, 50(200), pp. 425- tion," J. Polit. Econ., Jan./Feb. 1974, 82(1), pp. 37. 34-55. SIEBERTHORST."Environmental Protection and In- ROSEN,SHEHWIN. "Wage-Based Indexes of Urban ternational Specialization," Weltwirtsck. Arch., Quality of Life," in Current issues in urban eco- 1974, 110(3), pp. 494508. nomics. Eds.: PETERMIESZKOWSKI AND MAIILON SLOVIC,PAUL; FISCIILOFF, BARUCII AND LICIITEN- STRASZHEIM.Baltimore, MD: Johns Hopkins U. STEIN, SARAII."Facts Versus Fears: Understanding Press for Resources for the Future, 1979, pp. 74 Perceived Risk," in Social risk assess~nent:Hots 104. safe is enough. Ed. WALTERA. ALBERS.NY: Ple- ROSENBERG,NORMAN J. ET AL. "Processes for Identi- num, 1980. fying Regional Influences of and Responses to In- . "Response Mode, Framing, and Informa- creasing Atmospheric C02and Climate Change- tion-Processing Effects in Risk Assessment," in The MINK Project." Prepared by Resources for Question franiing and response consistency. Ed.: Cropper and Oates: Environmental Economics 739

ROBINM. HOGARTII.San Francisco, CA: Jossey- "Indivisible Toxic Torts: The Economics of Bass, 1982. Joint and Several Liability, Land Econ., Nov. SIIITEI,V. KERRY."The Role of Site and Job Charac- 1989, 65, pp. 305-19. teristics in Hedonic Wage Models," J. Urban "Economic Iilstruments for Environmental Econ., May 1983, 13(3), pp. 296-321. P,egulation," Oxford Rev. Econ. Policy, Mar. 1990, SMITH,V. KERRYAND DESVOUSGES,WILLIAM H. "The 6.. Generalized Travel Cost Model and Water Quality TOBEY,JAMESA. "The Impact of Domestic Environ- Benefits: A Reconsideration," Southern Econ. J., mental Policies on International Trade." Ph.D. Oct. 1985, 52(2), pp. 371-81. Dissertation, Dept. of Economics, U. of Maryland, Measuring water quality benefits. Norwell, College Park, 1989. MA: Kluwer-Nijhoff, 1986a. "TheEffects of Domestic Environmental Poli- -"Averting Behavior: Does It Exist?" Econom- cies on Patterns of World Trade: An Empirical ics Letters, 1986b, 20(3), pp. 291-96. Test," Kyklos, 1990, Fasc. 2. SMITII,V. KERRY;DES\~OUSGES, WILLIAM H. AND TOLLEY,GEORGEET AL. "Establishing and Valuing MCGIVNEY,MATTIIEW P. "Estimating Water Qual- the Effects of Improved Visibility in Eastern ity Benefits: An Econometric Analysis," Land United States." Prepared for the U.S. Environ- Econ., Aug. 1983, 59(3), pp. 259-78. ~nental Protection Agency, Washington, DC, SMITII,V. KERRY;PALMQUIST, RAYMOND B. AND JA- 1986a. KUS,PAUL. "Combining Farrell Frontier and He- -. Valuation of reductions in hunian health donic Travel Cost Models for Valuing Estuarine symptoms and risks. Final Report for the U.S. En- Quality." Nov. 15, 1990, second revision. vironmental Protection Agency. U. of Chicago, SIIITII,VERNON L. "The Principle of Unanimity and 1986b. Voluntary Consent in Social Choice," J. Polit. TRAVIS,CURTISAND Don, CAROLYN."Superfund: A Econ., Dec. 1977, 85(6), pp. 1125-39. Program Without Priorities," Enuironinental Sci- -"Incentive Compatible Experimental Pro- ence and Technology, 1989, 23(11), pp. 1333-34. cesses for the Provision of Public Goods," in Re- TRIJONIS,J. ET AL. "Air Quality Benefits for Los Ange- search in experiniental econoniics. Ed.: VERNON les and San Francisco Based on Housing Values L. SMITI~.Greenwich, CT: JAI Press, 1979. and Visibility." Final Report for California Air Re- SPOFFORD,WALTER O., JR."Efficiency Properties of sources Board, Sacramento, CA, 1984. Alternative Control Policies for Meeting A~nbient TURVEY,RALPII."On Divergences between Social Air Quality Standards: An Empirical Application Cost and Private Cost," Econoinica, Aug. 1963, to the Lower Delaware Valley." Discussion Paper 30(119), pp. 30S13. D-118. Washington, DC: Resources for the Fu- U.S. EN\JIHONMENTALPHOECTION AGENCY. "EPA's ture, Feb. 1984. Use of Benefit-Cost Analysis: 1981-1986." Aug. SPULBER,DANIEL F. "Effluent Regulation and Long- 1987, EPA Report 23C05-87-028. Run Optimality," J. Environ. Econ. Manage., June "Environmental Investments: The Cost of a 1985, 12(2), pp. 10S16. Clean Environment." Dec. 1990, EPA Report 23C STARRETT,DAVIDA. "Funde~nantalNonconvexities 12-9CO84. in the Theory of Externalities," J. Econ. Theory, U.S. GENERALACCOUNTING OFFICE. Energy issues. Apr. 1972, 4(2), pp. 180-99. Washington, DC, Nov. 1988. STAVINS,ROBERTN., ed. Project 88-Harnessing U.S. NATIONALOCEANIC AND ATIIOSPIIERICADMINIS- inarket forces to protect oztr enuironment: Initia- TRATION, NATIONALSCIENCE FOUNDATION, AND NA- tives for the new president. A Public Policy Study TIONAL AERONAUTICSAND SPACEADMINISTRATION. Sponsored by Senator Timothy E. Wirth, Colo- Economic and global change, New Hacen Work- rado, and Senator John Heinz, Pennsylvania. shop, May 1990. Washington, DC, Jan. 1991. Washington, DC, Dec. 1988. VAUGIIAN,WILLIAMJ. AND RUSSELL,CLIFFORDS. TERKLA,DAVID."The Efficiency Value of Effluent "Valuing a Fishing Day: An Applicatioil of a Sys- Tax Revenues," J. Environ. Econ. Manage., June tematic Varying Parameter Model," Land Econ., 1984, 11(2), pp. 107-23. Nov. 1982, 58(4), pp. 45C63. TIIALER,RICI~ARD AND ROSEN,SIIERWIN. "The Value VISCUSI,W. KIP; MAGAT,WESLEY A. AND HUBERT, of Life Savings," in Household production and con- JOEL."Pricing Environ~nentalHealth Risks: Sur- .sumption. Ed.: NESTERTERLECKYJ. NY: Columbia vey Assessmeilts of Risk-Risk and Risk-Dollar Tra- U. Press, 1976. deoffs," in AERE work.shop proceedings: Estimat- TIETENBERG,TITOMAS H. "Spatially Differentiated Air ing and valuing morbidity in a policy context, 1988. Pollutant Emission Charges: A11 Economic and Le- WALTEH,INGO."Pollutioil and Protection: U. S. Envi- gal Analysis," Land Econ., Aug. 1978, 5493), pp. ron~nentalControls as Competitive Distortions," 265-77. Weltwirtsch. Arch., 1974, 110(1), pp. 10413. "Transferable Discharge Permits and the "Trade, Environment and Comparative Ad- Control of Stationary Source Air Pollution: A Sur- vantage," in International economic.^ of pollution. vey and Synthesis," Land Econ., Nov. 1980, 56(4), Ed.: INGOWALTER. NY: Wiley, 1975, pp. 77-93. pp. 391-416. "Environinentally Induced Industrial Reloca- -Emissions trading: An exercise in reforining tion in Developing Countries," in Enuirontnent pollution policy. Washington, DC: Resources for and trade. Eds.: SEYMOURJ. RUBIN AND TIIOMAS the Future, 1985. R. GRAIIA~I.Totowa, NJ: Allanheld, Osmun, and Jou,rnal of Economic Literature, Vol. XXX (June 1992)

Co., 1982, pp. 67-101. WILDASIN,DAVID E. "Interjurisdictional Capital Mo- WALTER,INGOAND UGELOW,JUDITH."Environrnen- bility: Fiscal Externality and a Corrective Sub- tal Policies in Developing Countries," Ambio, sidy,"]. Urban Econ., Mar. 1989, 25(2), pp. 193- 1979, 8(2,3), pp. 102-09. 212. WEITZMAN,MARTIN L. "Prices vs. Quantities," Reu. WILLIG,ROBERT D. "Consurner's Surplus Without Econ. Stud., Oct. 1974, 41(4), pp. 477-91. Apology," Amer. Econ. Reo., Sept. 1976, 66(4), -. -. "Optimal Rewards for Econornic Regula- pp. 589-97. tion," Amer. Econ. Reo., Sept. 1978, 68(4), pp. WILSON,JOHN D. "A Theory of Interregional Tax 683-91. Competition," J. Urban Econ., May 1086, 19(3), WENDERS,JOHNT. "Methods of Pollution Control pp. 296-315. and the Rate of Change in Pollution Abatement ZODROW,GEORGER. AND MIESZKOWSKI,PETER. "Pi- Technology," Water Resources Res., 1975, 11(3), gou, Tiebout, Property Taxation and the Under- pp. 393-96. Provision of Local Public Goods," J. Urban Econ., WHITEHEAD,JOHNC. AND BLOMQUIST,GLENN C. May 1986, 19(3), pp. 356-70. "Measuring Contingent Values for Wetlands: Ef- ZUPAN,JEFFREY.The distribution of air quality in fects of Information About Related Environmental the Neu; York region. Washington, DC: Resources Goods," Water Resources Research, Oct. 1991, for the Future, 1973. 27(10). pp. 2523-31. http://www.jstor.org

LINKED CITATIONS - Page 1 of 10 -

You have printed the following article: Environmental Economics: A Survey Maureen L. Cropper; Wallace E. Oates Journal of Economic Literature, Vol. 30, No. 2. (Jun., 1992), pp. 675-740. Stable URL: http://links.jstor.org/sici?sici=0022-0515%28199206%2930%3A2%3C675%3AEEAS%3E2.0.CO%3B2-D

This article references the following linked citations. If you are trying to access articles from an off-campus location, you may be required to first logon via your library web site to access JSTOR. Please visit your library's website or contact a librarian to learn about options for remote access to JSTOR.

[Footnotes]

2 On Taxation and the Control of Externalities William J. Baumol The American Economic Review, Vol. 62, No. 3. (Jun., 1972), pp. 307-322. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28197206%2962%3A3%3C307%3AOTATCO%3E2.0.CO%3B2-V

2 Conservation Reconsidered John V. Krutilla The American Economic Review, Vol. 57, No. 4. (Sep., 1967), pp. 777-786. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28196709%2957%3A4%3C777%3ACR%3E2.0.CO%3B2-V

2 The Environment in Economics: A Survey Anthony C. Fisher; Frederick M. Peterson Journal of Economic Literature, Vol. 14, No. 1. (Mar., 1976), pp. 1-33. Stable URL: http://links.jstor.org/sici?sici=0022-0515%28197603%2914%3A1%3C1%3ATEIEAS%3E2.0.CO%3B2-8

13 Total Conditions in the Analysis of External Effects J. R. Gould The Economic Journal, Vol. 87, No. 347. (Sep., 1977), pp. 558-564. Stable URL: http://links.jstor.org/sici?sici=0013-0133%28197709%2987%3A347%3C558%3ATCITAO%3E2.0.CO%3B2-Y

NOTE: The reference numbering from the original has been maintained in this citation list. http://www.jstor.org

LINKED CITATIONS - Page 2 of 10 -

14 Control of Pollution When the Offended Defend Themselves Hirofumi Shibata; J. Steven Winrich Economica, New Series, Vol. 50, No. 200. (Nov., 1983), pp. 425-437. Stable URL: http://links.jstor.org/sici?sici=0013-0427%28198311%292%3A50%3A200%3C425%3ACOPWTO%3E2.0.CO%3B2-H

40 Discrete/Continuous Models of Consumer Demand W. Michael Hanemann Econometrica, Vol. 52, No. 3. (May, 1984), pp. 541-561. Stable URL: http://links.jstor.org/sici?sici=0012-9682%28198405%2952%3A3%3C541%3ADMOCD%3E2.0.CO%3B2-4

44 Prospect Theory: An Analysis of Decision under Risk Daniel Kahneman; Amos Tversky Econometrica, Vol. 47, No. 2. (Mar., 1979), pp. 263-292. Stable URL: http://links.jstor.org/sici?sici=0012-9682%28197903%2947%3A2%3C263%3APTAAOD%3E2.0.CO%3B2-3

45 Experimental Tests of the Endowment Effect and the Coase Theorem Daniel Kahneman; Jack L. Knetsch; Richard H. Thaler The Journal of Political Economy, Vol. 98, No. 6. (Dec., 1990), pp. 1325-1348. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28199012%2998%3A6%3C1325%3AETOTEE%3E2.0.CO%3B2-W

53 The Welfare Economics of Price Supports in U.S. Agriculture Erik Lichtenberg; David Zilberman The American Economic Review, Vol. 76, No. 5. (Dec., 1986), pp. 1135-1141. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28198612%2976%3A5%3C1135%3ATWEOPS%3E2.0.CO%3B2-G

61 Private Preference for Collective Goods Revealed Through Voting on Referenda Robert Deacon; Perry Shapiro The American Economic Review, Vol. 65, No. 5. (Dec., 1975), pp. 943-955. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28197512%2965%3A5%3C943%3APPFCGR%3E2.0.CO%3B2-L

NOTE: The reference numbering from the original has been maintained in this citation list. http://www.jstor.org

LINKED CITATIONS - Page 3 of 10 -

63 The Determinants of Pesticide Regulation: A Statistical Analysis of EPA Decision Making Maureen L. Cropper; William N. Evans; Stephen J. Berardi; Maria M. Ducla-Soares; Paul R. Portney The Journal of Political Economy, Vol. 100, No. 1. (Feb., 1992), pp. 175-197. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28199202%29100%3A1%3C175%3ATDOPRA%3E2.0.CO%3B2-9

64 The Net Benefits of Incentive-Based Regulation: A Case Study of Environmental Standard Setting Wallace E. Oates; Paul R. Portney; Albert M. McGartland The American Economic Review, Vol. 79, No. 5. (Dec., 1989), pp. 1233-1242. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28198912%2979%3A5%3C1233%3ATNBOIR%3E2.0.CO%3B2-8

References

The Pigouvian Tax Rule Under Monopoly A. H. Barnett The American Economic Review, Vol. 70, No. 5. (Dec., 1980), pp. 1037-1041. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28198012%2970%3A5%3C1037%3ATPTRUM%3E2.0.CO%3B2-C

On Taxation and the Control of Externalities William J. Baumol The American Economic Review, Vol. 62, No. 3. (Jun., 1972), pp. 307-322. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28197206%2962%3A3%3C307%3AOTATCO%3E2.0.CO%3B2-V

Detrimental Externalities and Non-Convexity of the Production Set W. J. Baumol; David F. Bradford Economica, New Series, Vol. 39, No. 154. (May, 1972), pp. 160-176. Stable URL: http://links.jstor.org/sici?sici=0013-0427%28197205%292%3A39%3A154%3C160%3ADEANOT%3E2.0.CO%3B2-9

NOTE: The reference numbering from the original has been maintained in this citation list. http://www.jstor.org

LINKED CITATIONS - Page 4 of 10 -

Crime and Punishment: An Economic Approach Gary S. Becker The Journal of Political Economy, Vol. 76, No. 2. (Mar. - Apr., 1968), pp. 169-217. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28196803%2F04%2976%3A2%3C169%3ACAPAEA%3E2.0.CO%3B2-I

Value of Life Saving: Implications of Consumption Activity Glenn Blomquist The Journal of Political Economy, Vol. 87, No. 3. (Jun., 1979), pp. 540-558. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28197906%2987%3A3%3C540%3AVOLSIO%3E2.0.CO%3B2-B

New Estimates of Quality of Life in Urban Areas Glenn C. Blomquist; Mark C. Berger; John P. Hoehn The American Economic Review, Vol. 78, No. 1. (Mar., 1988), pp. 89-107. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28198803%2978%3A1%3C89%3ANEOQOL%3E2.0.CO%3B2-%23

Welfare Measurement in the Household Production Framework Nancy E. Bockstael; Kenneth E. McConnell The American Economic Review, Vol. 73, No. 4. (Sep., 1983), pp. 806-814. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28198309%2973%3A4%3C806%3AWMITHP%3E2.0.CO%3B2-6

Measuring the Value of a Public Good: An Empirical Comparison of Elicitation Procedures David S. Brookshire; Don L. Coursey The American Economic Review, Vol. 77, No. 4. (Sep., 1987), pp. 554-566. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28198709%2977%3A4%3C554%3AMTVOAP%3E2.0.CO%3B2-F

On the Estimation of Structural Hedonic Price Models James N. Brown; Harvey S. Rosen Econometrica, Vol. 50, No. 3. (May, 1982), pp. 765-768. Stable URL: http://links.jstor.org/sici?sici=0012-9682%28198205%2950%3A3%3C765%3AOTEOSH%3E2.0.CO%3B2-Q

NOTE: The reference numbering from the original has been maintained in this citation list. http://www.jstor.org

LINKED CITATIONS - Page 5 of 10 -

External Diseconomies, Corrective Taxes, and Market Structure James M. Buchanan The American Economic Review, Vol. 59, No. 1. (1969), pp. 174-177. Stable URL: http://links.jstor.org/sici?sici=0002-8282%281969%2959%3A1%3C174%3AEDCTAM%3E2.0.CO%3B2-1

The Determinants of Pesticide Regulation: A Statistical Analysis of EPA Decision Making Maureen L. Cropper; William N. Evans; Stephen J. Berardi; Maria M. Ducla-Soares; Paul R. Portney The Journal of Political Economy, Vol. 100, No. 1. (Feb., 1992), pp. 175-197. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28199202%29100%3A1%3C175%3ATDOPRA%3E2.0.CO%3B2-9

The Value of a Life: New Evidence from the Marketplace Rachel Dardis The American Economic Review, Vol. 70, No. 5. (Dec., 1980), pp. 1077-1082. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28198012%2970%3A5%3C1077%3ATVOALN%3E2.0.CO%3B2-7

Private Preference for Collective Goods Revealed Through Voting on Referenda Robert Deacon; Perry Shapiro The American Economic Review, Vol. 65, No. 5. (Dec., 1975), pp. 943-955. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28197512%2965%3A5%3C943%3APPFCGR%3E2.0.CO%3B2-L

Differences Between Risk Premiums in Union and Nonunion Wages and the Case for Occupational Safety Regulation William T. Dickens The American Economic Review, Vol. 74, No. 2, Papers and Proceedings of the Ninety-Sixth Annual Meeting of the American Economic Association. (May, 1984), pp. 320-323. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28198405%2974%3A2%3C320%3ADBRPIU%3E2.0.CO%3B2-N

NOTE: The reference numbering from the original has been maintained in this citation list. http://www.jstor.org

LINKED CITATIONS - Page 6 of 10 -

Hedonic Prices and Implicit Markets: Estimating Demand and Supply Functions for Differentiated Products Dennis Epple The Journal of Political Economy, Vol. 95, No. 1. (Feb., 1987), pp. 59-80. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28198702%2995%3A1%3C59%3AHPAIME%3E2.0.CO%3B2-M

The Environment in Economics: A Survey Anthony C. Fisher; Frederick M. Peterson Journal of Economic Literature, Vol. 14, No. 1. (Mar., 1976), pp. 1-33. Stable URL: http://links.jstor.org/sici?sici=0022-0515%28197603%2914%3A1%3C1%3ATEIEAS%3E2.0.CO%3B2-8

Total Conditions in the Analysis of External Effects J. R. Gould The Economic Journal, Vol. 87, No. 347. (Sep., 1977), pp. 558-564. Stable URL: http://links.jstor.org/sici?sici=0013-0133%28197709%2987%3A347%3C558%3ATCITAO%3E2.0.CO%3B2-Y

Discrete/Continuous Models of Consumer Demand W. Michael Hanemann Econometrica, Vol. 52, No. 3. (May, 1984), pp. 541-561. Stable URL: http://links.jstor.org/sici?sici=0012-9682%28198405%2952%3A3%3C541%3ADMOCD%3E2.0.CO%3B2-4

Willingness to Pay and Willingness to Accept: How Much Can They Differ? W. Michael Hanemann The American Economic Review, Vol. 81, No. 3. (Jun., 1991), pp. 635-647. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28199106%2981%3A3%3C635%3AWTPAWT%3E2.0.CO%3B2-G

Social Cost of Environmental Quality Regulations: A General Equilibrium Analysis Michael Hazilla; Raymond J. Kopp The Journal of Political Economy, Vol. 98, No. 4. (Aug., 1990), pp. 853-873. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28199008%2998%3A4%3C853%3ASCOEQR%3E2.0.CO%3B2-3

NOTE: The reference numbering from the original has been maintained in this citation list. http://www.jstor.org

LINKED CITATIONS - Page 7 of 10 -

Too Many Proposals Pass the Benefit Cost Test John P. Hoehn; Alan Randall The American Economic Review, Vol. 79, No. 3. (Jun., 1989), pp. 544-551. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28198906%2979%3A3%3C544%3ATMPPTB%3E2.0.CO%3B2-N

The Economics of Exhaustible Resources Harold Hotelling The Journal of Political Economy, Vol. 39, No. 2. (Apr., 1931), pp. 137-175. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28193104%2939%3A2%3C137%3ATEOER%3E2.0.CO%3B2-G

The Value of Changes in the Probability of Death or Injury Michael Jones-Lee The Journal of Political Economy, Vol. 82, No. 4. (Jul. - Aug., 1974), pp. 835-849. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28197407%2F08%2982%3A4%3C835%3ATVOCIT%3E2.0.CO%3B2-A

The Value of Safety: Results of a National Sample Survey M. W. Jones-Lee; M. Hammerton; P. R. Philips The Economic Journal, Vol. 95, No. 377. (Mar., 1985), pp. 49-72. Stable URL: http://links.jstor.org/sici?sici=0013-0133%28198503%2995%3A377%3C49%3ATVOSRO%3E2.0.CO%3B2-W

Experimental Tests of the Endowment Effect and the Coase Theorem Daniel Kahneman; Jack L. Knetsch; Richard H. Thaler The Journal of Political Economy, Vol. 98, No. 6. (Dec., 1990), pp. 1325-1348. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28199012%2998%3A6%3C1325%3AETOTEE%3E2.0.CO%3B2-W

Prospect Theory: An Analysis of Decision under Risk Daniel Kahneman; Amos Tversky Econometrica, Vol. 47, No. 2. (Mar., 1979), pp. 263-292. Stable URL: http://links.jstor.org/sici?sici=0012-9682%28197903%2947%3A2%3C263%3APTAAOD%3E2.0.CO%3B2-3

NOTE: The reference numbering from the original has been maintained in this citation list. http://www.jstor.org

LINKED CITATIONS - Page 8 of 10 -

Hedonic Prices and the Benefits of Public Projects Yoshitsugu Kanemoto Econometrica, Vol. 56, No. 4. (Jul., 1988), pp. 981-989. Stable URL: http://links.jstor.org/sici?sici=0012-9682%28198807%2956%3A4%3C981%3AHPATBO%3E2.0.CO%3B2-W

Controlling Urban Air Pollution: A Benefit-Cost Assessment Alan J. Krupnick; Paul R. Portney Science, New Series, Vol. 252, No. 5005. (Apr. 26, 1991), pp. 522-528. Stable URL: http://links.jstor.org/sici?sici=0036-8075%2819910426%293%3A252%3A5005%3C522%3ACUAPAB%3E2.0.CO%3B2-3

Conservation Reconsidered John V. Krutilla The American Economic Review, Vol. 57, No. 4. (Sep., 1967), pp. 777-786. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28196709%2957%3A4%3C777%3ACR%3E2.0.CO%3B2-V

The Welfare Economics of Price Supports in U.S. Agriculture Erik Lichtenberg; David Zilberman The American Economic Review, Vol. 76, No. 5. (Dec., 1986), pp. 1135-1141. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28198612%2976%3A5%3C1135%3ATWEOPS%3E2.0.CO%3B2-G

The Net Benefits of Incentive-Based Regulation: A Case Study of Environmental Standard Setting Wallace E. Oates; Paul R. Portney; Albert M. McGartland The American Economic Review, Vol. 79, No. 5. (Dec., 1989), pp. 1233-1242. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28198912%2979%3A5%3C1233%3ATNBOIR%3E2.0.CO%3B2-8

Local versus National Pollution Control: Note Sam Peltzman; T. Nicolaus Tideman The American Economic Review, Vol. 62, No. 5. (Dec., 1972), pp. 959-963. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28197212%2962%3A5%3C959%3ALVNPCN%3E2.0.CO%3B2-G

NOTE: The reference numbering from the original has been maintained in this citation list. http://www.jstor.org

LINKED CITATIONS - Page 9 of 10 -

Wages, Rents, and the Quality of Life Jennifer Roback The Journal of Political Economy, Vol. 90, No. 6. (Dec., 1982), pp. 1257-1278. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28198212%2990%3A6%3C1257%3AWRATQO%3E2.0.CO%3B2-7

Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition Sherwin Rosen The Journal of Political Economy, Vol. 82, No. 1. (Jan. - Feb., 1974), pp. 34-55. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28197401%2F02%2982%3A1%3C34%3AHPAIMP%3E2.0.CO%3B2-U

The Coase Proposition, Information Constraints, and Long-Run Equilibrium William Schulze; Ralph C. D'Arge The American Economic Review, Vol. 64, No. 4. (Sep., 1974), pp. 763-772. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28197409%2964%3A4%3C763%3ATCPICA%3E2.0.CO%3B2-R

Control of Pollution When the Offended Defend Themselves Hirofumi Shibata; J. Steven Winrich Economica, New Series, Vol. 50, No. 200. (Nov., 1983), pp. 425-437. Stable URL: http://links.jstor.org/sici?sici=0013-0427%28198311%292%3A50%3A200%3C425%3ACOPWTO%3E2.0.CO%3B2-H

The Principle of Unanimity and Voluntary Consent in Social Choice Vernon L. Smith The Journal of Political Economy, Vol. 85, No. 6. (Dec., 1977), pp. 1125-1139. Stable URL: http://links.jstor.org/sici?sici=0022-3808%28197712%2985%3A6%3C1125%3ATPOUAV%3E2.0.CO%3B2-R

On Divergences between Social Cost and Private Cost Ralph Turvey Economica, New Series, Vol. 30, No. 119. (Aug., 1963), pp. 309-313. Stable URL: http://links.jstor.org/sici?sici=0013-0427%28196308%292%3A30%3A119%3C309%3AODBSCA%3E2.0.CO%3B2-T

NOTE: The reference numbering from the original has been maintained in this citation list. http://www.jstor.org

LINKED CITATIONS - Page 10 of 10 -

Optimal Rewards for Economic Regulation Martin L. Weitzman The American Economic Review, Vol. 68, No. 4. (Sep., 1978), pp. 683-691. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28197809%2968%3A4%3C683%3AORFER%3E2.0.CO%3B2-%23

Consumer's Surplus Without Apology Robert D. Willig The American Economic Review, Vol. 66, No. 4. (Sep., 1976), pp. 589-597. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28197609%2966%3A4%3C589%3ACSWA%3E2.0.CO%3B2-0

NOTE: The reference numbering from the original has been maintained in this citation list.