Download as pdf
Download as pdf
You are on page 1of 14
Willingness to Pay 135 Finding the demand for public schooling is no easy task, basically because of the lack of markets. But the big difference between environmental goods and ordinary public goods is the supply side. Ordinary public goods are produced at a cost. Thus citizens have a ref- erence point of the cost of supply to take into account when determining how much of the good they want at particular prices. Although strictly speaking this should not impact the value to consumption, it can make the process of valuation easier. Furthermore, peo- ple may have experience with different points on their demand curve; perhaps they have lived in different communities with different amounts of the public good or are aware of the cost of private schools, In contrast, environmental goods are often quite disconnected from the supply side. In determining how worthwhile the protection of a wilderness area might be, itis dif- ficult to anchor that value, since the wilderness area is not produced. Air quality is in fact supplied, since there is a cost associated with cleaning up the air. However, this is a ‘much more dispersed cost, much different than that associated with constructing a public school or park. ‘This too makes the job of determining demand more difficult, IL WILLINGNESS TO PAY A conventional demand curve shows quantity demanded as a function of price. Without a market there is, of course, no price. As discussed in Chapter 4, this is not really a problem—the concept of demand works perfectly well and makes perfect sense without a market. Similarly, the consumer surplus associated with consuming a certain amount of a good is defined in terms ofa market transaction. In this section we see how similar concepts are equally applicable to environmental goods in the absence of markets. The ‘key concept corresponding to price and surplus is willingness to pay. Marginal willing- ness to pay is analogous to price and total willingness to pay is analogous to consumer surplus. A. Prices and Marginal Willingness to Pay (MWTP) In the next few chapters we will be taking a closer look at how the demand for environ- mental goods is measured, We start with an example. Leaving aside how it was obtained, Figure 7.1 shows the demand (at the household level) for nitrogen oxides (NO,) air pol- lution in Boston, based on a study done in the 1970s. Figure 7.1 shows three different demand curves for NO,, corresponding to three different levels of household income. For one more unit of NO,, consumers are willing to pay a negative amount: they must be compensated to agree to consume more pollution. The demand curves are downward sloping-but since NO, is a bad, they are in the fourth quadrant (MWTP negative, NOy positive). Ifthe vertical axis were the marginal willingness to pay for pollution reductions, then the graph would be in the first quadrant, though it would be sloping upward to the right. . Although Figure 7.1 might seem a bit strange and certainly looks different than a demand curve for a market good, it is a totally appropriate representation of demand for NO, pollution. It represents the strength of preferences for households of different ( f ) UNIVERSIT UN 136 CHAPTER 7 DEMAND FOR ENVIRONMENTAL GOODS MWTP ($/pphm) 1000 2000 -3000 NO, (pphm) FIGURE 7.1 Marginal willingness to pay 46 8 10 (MWTP) for nitrogen oxides Cparts per hundred million, pphm) for Boston households at three different income levels. Figures for MWTP are annual, in 1970 US$ per pphm. (Adapted from Harrison ‘and Rutuingeld, 1978.) Income = $8500 Income = $11,500 Income = $15,000 income levels. The graph shows a negative willingness to pay for increases in pollution, or, equivalently, a positive willingness to pay for decreases in pollution, At a level of 6 pphm of NO, pollution, lower income households have a marginal willingness to pay to reduce pollution to 5 pphm of approximately $1000; higher income families would be willing to pay twice this amount. ‘There is a nuance that has to do with whether the consumer is acquiring one more nit of the good or giving up a unit. If the consumer is giving up a unit that is possessed (e.ga selling a liter of gas), the appropriate term is marginal willingness to accept compen- sation (MW'TA). In most contexts, these two concepts are equivalent. We will have more to say about differences between these two terms later in this chapter. ‘There is one issue of terminology that arises in the case of pollution. Often, we speak of the damage from pollution or the benefit of having less pollution. The marginal ben- efit of one less unit of pollution is the same as the marginal damage of one more unit of pollution, Furthermore, the marginal willingness to pay for one more unit of pollution is negative and equal to the negative of the marginal benefit of one less wnit of pollution or the negative of the marginal damage of one more unit of pollution. Confusing? Just keep: in mind that benefits of less pollution are the other side of the coin of damages from more pollution, When expressed in monetary terms, these benefits and damages are simply wwillingnesses to pay, making sure that the sign (positive or negative) is correct. B. Willingness to Pay versus Marginal Willingness to Pay With pollution, we are often concerned with the total damage from pollution or total benefits from reducing pollution. For instance, Table 7.1 shows an estimate of the total benefits (1970-1990) from the cleaner air provided by the 1970 U.S. Clean Air Act Amendments. The table is taken from a U.S. Government study of benefits over the period 1970-1990 and compares the situation of pollution continuing ta grow without the legislation with the levels of pollution that in fact resulted over the 20-year period. Thus ‘Table 71 does not compare existing levels of pollution with pristine conditions but rather existing levels of pollution with much dirtier conditions that might have prevailed had there been no regulation of air pollutants (a level that is very hard to forecast). The point is C f ) UNIVERSIT UNI Willingness to Pay 137 TABLE 7.1 Present Value of 1970 to 1990 Monetized Benefits* of Clean Air Act in Continental United States, by Endpoint (Billions of 1990 §, Discounted to 1990 at 5%) Present value (5) Endpoint Pollutants) Sth %ile Mean 95th sile Mortality PM10 1991 13502 30.968 Mortality Pb 125 1350 4,096 Chronic bronchitis PM-10 bas 7156 12,613 1Q (lost IQ pts. + children with 1Q.<70) Pb 29 466 656 Hypertension Pb 7 2 m0 Hospital admissions PM-10, 0, Pb, and CO is a4 33 Respiratory-related symptoms, restricted P10, 0,,NO,.$0, 3 78 9 activity and decreated productivity Seiling damage PM10 7 %6 196 Visibility Particulates 55, n 9 ‘Agriculture (net surplus) 9, u 23 33 “Total 10500-28000 40,600 “Benefits are with respec 10. hypothetical baseline af ro pollution control im 1970-1990, with resulting growth in polduion eves Source: United States Environmental Protection Agency (1997), Table 20, that Table 7.1 shows total benefits of the Clean Air Act, equivalent to the total willingness to pay for the attained levels of air quality, not marginal willingness to pay.? This illustrates the connection among damage, willingness to pay, and benefits. Damage from pollution is equal to the negative of the benefits of eliminating the pal lution, which is equal to the willingness to pay to eliminate the pollution, For example, suppose we are in an urban area with high levels of air pollution, such as Mexico City. We may think of the total willingness to pay for an individual citizen to eliminate all air pol- lution from the city. This would be the total willingness to pay to eliminate ait pollution. It is the same as the damage (in monetary units) to the individual from all of the air pol lution in Mexico City. Of course, we can also consider the willingness to pay to eliminate one unit of air pollution. This would be the marginal willingness to pay for air pollution reduction, which would be equal to the marginal damage from pollution. Figure 7.2 shows hypothetical individual demand curves for two commodities: gas- oline and air pollution. For gasoline, Figure 7.2a shows quantity demanded as a function ‘of the price of gasoline. Remember “price” can be used interchangeably with “marginal ‘willingness to pay.” For air pollution, the lower part of Figure 7.2b shows the consumer marginal willingness to pay for pollution (MW'TP) as a function of the quantity of pol- lution. The top part of Figure 7.2b shows the marginal damage from pollution, which is precisely the negative of consumer total MW'TP. Total willingness to pay to eliminate pollution involves adding up all of the marginal wwillingnesses to pay as pollution levels expand from zero to the current level. This is the area between the demand curve for pollution and the horizontal axis and is shown as the lighter shaded part of Figure 7.2b. Similarly, in Figure 7.2a, the shaded area shows the total willingness to pay for gasoline q*. This shaded area consists of payments made (the rectangle) and consumer surplus (the triangle) 138 CHAPTER 7 DEMAND FOR ENVIRONMENTAL GOODS ‘Sfliter S/unit ‘Consumer surplus Demand Total damage from pallution Payments, a iiters Consumer total WIP for pollution @ Consumer MwTP for pollution ) FIGURE 7.2 Demand curves and willingness to pay. (a) Gasoline; (b) air pollution. Note that in (b), total damage = ~consumer total WTP. Similarly, for an environmental commodity, the area under the demand curve between two different quantities of the commodity gives the willingness to pay for the environmental commodity, Since no payment is made for the pollution (as there was for gasoline), this is also consumer surplus (though note that it is negative). The lighter shaded area of Figure 7.2b is the consumer WTP for pollution. It is negative since pol- lution is bad and people must receive money to be willing to accépt the pollution. The darker shaded area of Figure 7.2b is the total damage from pollution, WTP for pollution isthe negative of total damage. lll. TYPES OF ENVIRONMENTAL GOODS HARVARD UNIVERSITY ‘The environment is a very complex place. In understanding the value of the environ- ‘ment, it helps to use a classification scheme, We can classify goods based on the nature of, the pollution: air quality is one good, water quality is another. Or we may classify goods based on the nature of the injured party: health effects of pollution, damage to agricul- ture, and damage to buildings and materials. Or we may categorize goods based on how people perceive the damages. This last categorization is a bit more subtle and has to do with whether consumers obtain utility from the environment by “using” the environment (eg., hiking through the woods) or through more esoteric means (¢-g., thinking about elephants roaming in Africa). Because this last dichotomy is somewhat unusual, we will examine it more carefully. Digit b Original HARVARD UNIVERSITY Types of Environmental Goods 139 A. Use Value Use value is the conventional notion of value associated with the consumption of a good In the case of an environmental good, this could include current use (“I am currently vis- iting the park”), expected use (“I plan to visit the park later this year”), and possible use (“I might visit the park within the next 10 years”), Although this might be obvious, the point about possible use can be used to generate a positive value for a remote wilderness. that few currently use, particularly if the very long term is taken into account. ‘There are several primary avenues whereby environmental goods impact humans. One is through direct impact. This would include direct health effects of breathing polluted air (both mortality—death—and morbidity—sickness) as well as nonhealth effects such as bothersome odors, noise, or visual impacts. For instance, smoke from 4 power plant may result in a brawn plume intercepting a vista you have come to enjoy. Although there may be no measurable physical impaet on you from this plume, you find the view annoying and would be willing to pay some amount of money to be rid of it, should the opportunity arise. This is just as real an economic value as health effects. On the production side, pollution may enter directly into production functions, adversely affecting production processes. For instance, air pollution may make the manufacture ‘of microchips (which require superclean conditions) more costly. Environmental goods ‘may impact humans less directly, primarily through materials damage. Air pollution may ‘damage or soil buildings, necessitating additional maintenance or cleaning. ‘A second avenue for environmental goods to impact humans is through damage to ecosystems. Agriculture, forests, and fisheries are ecosystems from which humans directly derive economic benefit. Clearly pollution that degrades the performance of these ecosystems is undesirable, More subtle ecosystem effects relate to recreational use of eco- systems, Urbanization or pollution that may disrupt the ecosystem in a national park will have negative consequences for recreational users of that national park. Ecosystems are also.a source of economic benefit through mechanisms such as providing genetic material for new pharmaceuticals. ‘The value of ecosystems services has become an important and controversial area of valuation, in large part because of the difficulties in actually determining the tangi- ble consequences of many disruptions of ecosystems (for instance, how do fewer species translate into impact on man?). In an effort to bring more structure to this important but often underemphasized area of environmental goods and services, the U.S. EPA's high- level Science Advisory Board recently developed guidelines for use in valuing ecosystem services (U.S, Environmental Protection Agency, 2009). B. Nonuse Value Nonuse value is a controversial aspect of value. Nonuse value is gain in a person's utility without the person actually applying the good, applying the word “using” very broadly. ‘We may value ecosystems in remote parts of the world for reasons other than intending to visit the ecosystem or potentially obtaining something useful from the ecosystem. I may value a blue whale being able to swim freely in the open ocean, despite the fact that Ihave no intention of ever seeing a blue whale swim in the open ocean, The three basic types of nonuse value are existence value, altruistic value, and bequest value. Existence value is the value a consumer attaches to knowing something 140. CHAPTER 7 DEMAND FOR ENVIRONMENTAL GOODS exists (e.g,, the African elephant example mentioned earlier). This would be in addition to any value associated with actual or potential “use.” Altruistic value derives not from my own consumption but from the fact that I derive benefit when someone else gains utility. So when my neighbor benefits from my cleaning my front yard, if lam altruistic, 1 obtain utility from the fact that my neighbor is better-off. Bequest value is similar, though associated with the well-being of descendants, If value passing a wilderness area on to the next generation, that wilderness has a bequest value to me, even if I never use it or intend to use it. The purpose of dissecting value and placing it into these various categories is to understand the complexity whereby environmental goods confer value on consumers. Simply looking at the use value of an environmental asset may obscure much of the value, In fact some environmental assets have little use value but very significant nonuse value. We should point out that empirically it is often impossible to separately measure the components of value, for instance, how much value is alteuistic vs. bequest. This is a bit like dividing the value of an ice-cream cone into texture and flavor. The categoriza- tion simply provides an intellectual framework to help ensure completeness when valuing environmental goods IV. MEASURING DEMAND Because of the absence of markets for environmental goods, measuring demand is not straightforward, There are two basic approaches to measuring demand: revealed preference and stated preference. In revealed preference, we observe a real choice’ in some market and cleverly infer information on the trade-off between money and the environ- mental good. For instance, we may notice that two communities are identical except that one has high housing prices and clean air and the other has lower housing prices and dirty air. We may infer that the difference in housing prices reflects the value people place on clean air. The second approach, stated preferences, basically involves asking people how they would trade off money for the environmental good if they had the opportunily to do so. Opinion polls and surveys are used to derive this information. This approach is controversial because of the absence of real choices; only real choices invalve actual trades between money and environmental goods. Hypothetical choices lack a realism that may be necessary to obtain accurate information. There is a good deal of division within the economics profession on this point. There is also a gray region between revealed preferences and stated preferences, for example, experimental markets in which subjects (usually students) are given some money and asked to make choices (pethaps somewhat unusual choices) using that money. The decisions are reals the context is contrived. Within the category of revealed preference, there are two basic approaches to mea~ suring demand, hedonics and household production. In the hedonic approach, the goal is to see how the price of a conventional good (e.g.,a house) variesas the amount of aclosely related environmental good changes (e.g., the air quality in the vicinity of the house). This relationship is then used to infer value, The household production approach starts with the assumption that consumers will combine private goods with the environmental good to“produce” another good, which isthe real source of utility. For instance, soundproofing may be combined with noise that impinges on the outside of a house to obtain particular Measuring Demand 141 indoor noise levels. Indoor noise is what is assumed to matter. Or a national park may be combined with expenditures to visit the park to generate a park visit, which is what really gives utility. In either case, by observing expenditures om the complementary pri- vate good, we obtain a lower bound on the value of the environmental good or bad. The outside noise must cause at least as much damage as the expenditures on soundproofing. ‘The national park visit must be worth at least the travel expenditures. In the arena of stated preferences, the dominant approach is termed contingent val- uation. Contingent valuation relies on direct revelation of demand from consumers. The name literally means “value contingent on there being a market”—if there were a mar- ket, how much would you pay for the environmental good? These values are obtained by directly questioning a sample of potential consumers of the environmental good. Contingent valuation is a type of constructed market. In a constructed market, a researcher will take a situation in which no market exists and generate a market. Constructed markets can be hypothetical or real. For instance, in a study of duck hunt- ing, Bishop et al. (1983), constructed a market for buying and selling duck hunting per- mits (the permits were normally distributed through a lottery). In this manner they were able to generate a demand curve for duck hunting permits. Laboratory experiments are also widely used to generate information on how people trade money for environmental goods. In a laboratory study of the value of risk reduction, Shogren et al. (1994) con- structed an experiment (see box on pages 146-147 “Experiments with Food Safety” later in this chapter) in which students were given the option of eating food prepared subject to conventional safety standards (which entails some risk, though small) with food that had been prepared with even stricter standards. The researchers were interested in the behavior of demand for these risk reductions. Such laboratory experiments are useful because they enable the investigation of questions for which market data just are not available. These studies also involve real resource decisions (“real” money is on the line), Which is an advantage over contingent valuation. Unfortunately, they are expensive to conduct, often involve non-randomily selected samples, typically students in economics or related fields, and are difficult to generalize. Another type of constructed market is the official referendum. Some environmental good is ta be provided using a tax on resi dents of a community. The proposal is put to the voters of a community. Although this is a very real choice situation, its usefulness is obviously limited to the environmental goods offered to communities through referenda, A somewhat different approach to valuation has come to be known as “benefits transfer In a sense this is a poor man’s approach to benefits estimation. If you do not have the resources to conduct a new study for the particular issue in question, survey the literature and find similar analyses and “transfer” them to your application. Valuing Beach Recreation with Benefits Transfer On Feb. 7, 1990, off of Southern California, an oil tanker named The American Trader spilled a significant amount of crude oil onto the beaches made famous in song by the Beach Boys, beaches which also happened to be a major source of recreation for the 15 million residents of the Las Angeles area. In the ensuing litigation to recover damage from the oil spill ‘a major issue was the damage to environmental resources and recreation. Digi Original HARVARD UNIVERSITY HARVARD UNIVERSITY 142 CHAPTER 7 DEMAND FOR ENVIRONMENTAL GOODS In the case of beach recreation, the question faced by both sides of the litigation was twofold: what was the loss in beach use and what was the willingness to pay by recreators to avoid that loss of use? The loss in beach WTP is where benefits transfer enters, and both sides of the case used benefits transfer to estimate the damages. One of these benefits transfers is documented in Deacon and Kolstad (2000) and involves the use of studies of beach recreation in Florida, Massachusetts, Hawaii, Rhode Island, and several Southern California beaches. Although there is significant variation in the results of the studies and the applicability to Southern California, all studies placed the value of a lost person-day of beach recreation between $1 and $12 (in 1990$) with some narrowing of the range possible based on characteristics of the individual stu Although an approach such as this does not give a definitive figure for the damage to the particular beaches in question, itis far cheaper than undertaking a new study; depending on what the policy need is, narrowing the range of damage may provide sufficient precision. V. REVEALED PREFERENCE AND RESTRICTED DEMAND Revealed preference measures are the gold standard for economic estimates of the demand for environmental goods and services. The concept is simple: based on actual choices people make, we infer how they must be making trade-offs between market goods and environmental goods. This allows us to derive a demand curve for environmental goods, For instance, we observe the demand for houses changes when air quality changes. People pay more for identical houses when the air is cleaner. That behavior in a market should allow us to infer how people are valuing air quality. The standard approach to estimating demand for goods is to obtain a great deal of transaction data involving quantity demanded and price. Using statistical methods, we then tease out a relationship between price and quantity demanded. Demand for ga will depend on the price of gasoline (of course) as well as the price of public transporta- tion, the price of alternative fuels, the price of housing close to work, the price of vehicles and a host of other prices (and income). A major problem with using this approach with environmental goods is that there are no prices for environmental goods. It turns out that we can often get around this problem using the concept of restricted demand. We measure (statistically, using observations) how demand for a market good changes with the price of the market good and the quan- tity of the environmental good, We then use some sleight of hand to deduce the MWTP for the environmental good as a function of the quantity of the environmental good and the price of the market goods. ‘To see this, consider the simplest setup with one market good (x) and one environ- ‘mental good (q). The price of the market good is p. For instance, houses (of a certain type) may be the market good and air quality the environmental good, in which case p, is the price of a house and q is air quality, Using data on house sales and air quality in many Digi /ARD UNIVERS| Original RD UN Revealed Preference and Restricted Demand 143 imate the following relationship ‘of housing demanded (x): different locations in an urban area, we are able to between the price of houses, air quality and the quant x= hp,a) eo) Normally a demand curve has prices as arguments on the right-hand side; Eq, (71) has one price and one quantity on the right-hand side. We refer to this as a restricted demand curve in the sense that the arguments of the demand curve are both prices and quanti- ties. For a conventional demand curve, the arguments on the right-hand side would all be prices (perhaps with other variables such as income). ‘A relationship such as Eq. (7.1) certainly helps us understand how housing markets work, but how could it help us understand the demand for the environmental good? The answer depends on finding the MWTP for q. ‘To find MWTP, we find the total WTP for the market good and observe haw that changes when we change qa littl. The ratio of the change in WTP to the change in q gives us the MWTP, This is illustrated in Figure 7,3. Shown in the figure is the restricted demand for the market good (housing in our example). Infact, two such demand curves are shown, ‘one for q and one for q + Ag, with the assumption being that Aq is small positive number. ‘The figure is consistent with our intuition in our example—when air quality goes up in an area, demand for housing in the area also increases. The change in the willingness to pay is the change in consumer surplus for the two demand curves. This change is the shaded area between the two demand curves, as shown in Figure 73. If we know Ag and can compute AWTP, and then let the Ag become infinitesimally small, we can derive the M’ AWTP / Aq = MWTP = /ip.a) (72) This is exactly what we are after—a demand curve for the environmental good, relat- ing the marginal willingness to pay for the environmental good to the quantity of the environmental good.” ‘Although this seems to answer the question of how to obtain demand for an envi- ronmental good from demand for market goods, it turns out that this is really a lower bound. All of the value from the environmental good was expressed through demand for FIGURE 7.3 Marginal willingness to pay for an environmental good. ,. rice of x; h, demand for x: p*,, market price of x; Ag, change in environmental good, g. Pe awTP Demand for x after increase in 4p q+ dq) Dna Digi Original ‘MRD UNIVERSITY HARVARD UN 144 CHAPTER 7 DEMAND FOR ENVIRONMENTAL GOODS the market good. If in fact the only purpose of the environmental good is to enhance the experience of consuming the market good, then the question has been resolved. However, that is rarely the case. Air pollution, for instance, certainly affects enjoyment of one’s ‘dwelling. But it also influences utility in ways unrelated to housing—in fact, whenever one is traveling, working, or recreating in a reduced air-quality environment—or even. indirectly, through ecosystems that simultaneously provide services to people but are harmed by adverse environmental quality. This is not to diminish the value of determining the demand for air quality through the demand for housing. It is only to caution that the demand so found may in fact be a lower bound on demand for the environmental good. VI. DISCRETE CHOICE ‘Most empirical methods of measuring demand for environmental goods involve observ- ing consumer choices, A consumer may consider many beaches for recreation but visit only one, or homeowners consider many houses but buy only one. Making real choices is common form of revealed preference, and observing choices is a common way in which we infer demand for environmental goods. The important thing to keep in mind is that we learn from what consumers choose but we also learn from what consumers reject. Because so many empirical approaches to estimating demand involve discrete choice, itis, appropriate to consider the general approach to such problems. We will consider here what has been called the “random utility model," but the method can be used in a variety of contexts. Consider Jose, who is buying a house. Jose observes many houses and chooses to buy one. A house has characteristics, 2, and a price, p: (2.p). Let there be N diffe cent houses in Jose’s town. Index all the houses by iso that the houses can be listed as (€pP)s-+-(2,P)o---s(2ypP,). From these, Jose picks house k. That means he likes house k better than all the other houses in his town. ‘Think of Jose spending his money on a composite consumption good, x, and his house. The units for the consumption good are defined so that its price is 1. Jose’s income is y. The big question is, what can we learn from observing his choice of house? One approach is to assume Jose has an indirect utility function, V(z,p.y:a) where a is a vector of parameters that we hope to estimate econometrically.* V says how much utility Jose gets when he chooses a house with characteristics z, spending pon the house, and the rest of his income, y, on consumption goods, But like many things in economics, V is an approximation of Jose's utility—Jose’s true utility is V plus an error term, e. This means that the actual utility Jose gets from choosing house i is V(z,p,93a) + €, Now consider Jose’s choice between house k (which he chooses) and house mt (which he rejects). This must mean house k gives higher utility than house m; that is, Choosing house k over house m © V(z,.p,yia) + €, > V(z_p_yia) + €,. (7.3) But since ¢ is an error term, with noise and randomness to view the choice probabilistical » it is really more appropriate s ie. Eq, (7.3) is equivalent to saying Probability [choosing k over m] = Prob [e,~€, > VUz,p,y30)- Veypyyna)] (74) ( f ) UNIVERSIT UNI WIP versus WTA 145 ‘Since Jose did in fact choose house k over m, we can find the value of the parameter vector a which maximizes this probability, | But if we stopped there, we have not used all the information at our disposal, Jose actually chose house k over all the other houses. Using Eq. (7.4), we can write the slightly more complicated expression for the probability of choosing house k over all the other houses in town. The econometric problem is to find a value of a which maximizes that probability since we know that Jose did, in fact, choose house k over all the others.’ ‘This is a random utility model. We specify a functional form for indirect utility, complete with parameters to estimate, and append an error term, Rather than observe utility directly (which would make estimation very straightforward), we only observe choices. In this case, we find the parameter vector that makes the calculated choices most consistent with the choices we observe. The end product is an estimated indirect utility function. Fresh Water Recreation in North Carolina Dan Phaneuf (2002) asks a simple question: what would be the willingness of recreators to pay for improving water quality at a number water-based recreation sites in North Carolina Cin the Southeastern United States)? The author has a database on the water quality in various recreation sites in the state (58 different watersheds in the state). He also has information on over 1,000 day trips to these recreation sites—where people come from, which site they choose, and their characteristics. This allows him to construct a cost for visiting any site, based on distance traveled—this is the “price” of visiting a site. Thus for each of the over 1,000 trips, he knows the “price” of visiting each possible site, the characteristics of each possible site, the characteristics of the individual recreator and the actual site the recreator chooses from among the 58 choices. The problem he develops and estimates is almost identical to the example in the text with a recreation site replacing a house. The result of the estimation is an indirect utility function that indicates the utility a specific individual obtains from visting a site with a particular cost of visiting and site characteristics. He can then ask what would be the Person's increased utility from and thus willingness to pay for an increase in site quality. He then examines several important recreation areas in state and asks what the WTP would be for a 10% and a 25% improvement in.a number of water-quality parameters. For a 10% improvement in water-quality indicators, he finds a willingness to pay ranging from $2.67 to $6.29 per recreation trip. Vil. WTP VERSUS WTA ‘An issue that has arisen for environmental goods and services is the distinction between ‘what a person would pay for more of a particular environmental good versus what a person would accept as compensation for a little less of an environmental good. Both al UNIVE! Digitiz HARVARD UNIVERSITY SITY 146 CHAPTER 7 DEMAND FOR ENVIRONMENTAL GOODS concepts could be used in making decisions on environmental programs, and if the answers are very different, we should understand why. Furthermore, which measure should we use in evaluating an environmental action? Recall the example of Finch Ellsworth and his demand curve for gasoline—Finch demands 50 liters of gasoline when the price is €1 per liter. If Finch is voluntarily choos- ing to consume 50 liters at that price, what would he be willing to pay for an extra liter? Alternatively, what would he be willing to accept as compensation to reduce consump- tion to 49? Intuition, and economic reasoning, tells us that the answer in both cases is the same or very nearly the same and approximately equal to £1. Finch values a marginal unit of gasoline at £1 per liter, which means one more or one less liter is worth about the same—£1. Finch’s marginal willingness to pay (MWTP) for one more liter of fuel is approximately equal to his marginal willingness to. accept (MWTA) compensation for ne less liter of fuel. In this example, the difference between these two concepts (MWTP and MTWA) is of little consequence. But now think about the air pollution in the city or area in which you live. How much would you be willing to pay to clean up the pollution by one unit (MWTP)? Now suppose a manufacturing facility is expanding and will result in a decrease in air pol- lution of one unit. How much would the manufacturer need to pay you for you to agree to the one unit decrease in pollution (MWTA)? Many people would view MWTA >> MWTP. Experimental evidence is consistent with this example, even for market goods (such as Finch’s gasoline in the earlier paragraph)” For instance, Kahneman et al. (1990) set up an experimental situation where students buy and sell Cornell University coffee mugs. Some students are given the mugs and the opportunity to sell them to other students. All students are given cash to use in the transactions. The researchers find the sellers express 4 WTA for their newly endowed coffee mugs that exceeds the WTP for those students who are in a position to purchase one. Other experimenters have found similar results (see box below). In the context of environmental goods and services, the question is often a non- marginal one. Is a person's willingness to pay to avoid the flooding of a scenic river (due to a dam) equal to the person's willingness to accept compensation for that flooding? Willingness to pay is limited by the income or wealth of an individual whereas willing- ness to accept is not. For this reason alone, itis easy to see that WTA and WTP may not be the same for many environmental goods. The reason this is an important question is that in generating demand curves for ‘environmental goods and services, we need to know whether we should be dealing with WTA or WTP (MWTA or MWTP), Which is correct? Experiments with Food Safety To test theories about the difference between willingness to pay and willingness to accept, Jay Shogren and colleagues at lowa State University set up a series of laboratory experiments involving students at lowa State (Shogren et al., 1994). Shogren was interested in whether there was a gap between WTP and WTA, and also whether there was a difference between market goods and nonmarket goods. Digitized by al from HARVARD UNIVERSITY HARVARD UNIVERSITY WTP versus WTA 147 The experiment involved thirty students at a time, divided into two groups—buyers and sellers. In one experiment (involving a market good), the good in question was an easily purchased candy bar. In the second ‘experiment, each student was given a sandwich purchased at the local grocery store—safe, but as with all commercial food, carrying some small chance of food poisoning. An alternative good in the second experiment ‘was a sandwich that had been subjected to stringent tests to assure that it would have much lower health risks. This choice of a level of safety is not ‘one normally faced in a market. Students were divided into two groups. buyers and sellers. In the case of the candy bar, all students were given $3; the members of the seller group were also given a candy bar. A market price was determined by each student indicating: for sellers, what they would accept ($) to sell the candy bar (WTA); for buyers, what they would pay ($) for the candy bar (WTP). This bidding process was repeated five times. Interestingly, in the first trial, the average WTA was 50% higher than the average WTP, and statistically different. However, with repetition, the WTA declined «the WTP stayed relatively constant) so that by the third trial WTA and WTP were virtually identical, The same sort of experiment was conducted for safe sandwiches except in this case each student was endowed with $15 and a sandwich. The buying students were endowed with the grocery-store-quality sandwich and were given an opportunity to exchange it for a super-safe sandwich. The selling students were endowed with a super-safe sandwich and were given an opportunity to exchange it for a grocery-store quality sandwich. Students were asked WTP and WTA questions as in the candy bar ‘experiment. At the end, trades occurred and each stucient was required to eat his or her sandwich. The results show that for these nonmarket goods (sandwich safety), the average WTA started much higher than the average WTP and that this difference did not disappear with repeated trials. These results lend credence to the theory that substitutability with market goods reduces the difference between WTP and WTA. It also Suggests that we may expect large differences between these measures for many environmental goods. But with any experiments it is important to keep in mind that results that are consistent with theory are not the same thing as a confirmation of theory. Experiments can disprove a theory, but in order to provide support for a theory, it would be important to test (and reject) other theories that have also been put forward as to why WTP and WTA may differ. ‘To understand the problem more clearly, we can simplify to a problem with two goods—an environmental good (q) and a numeraire good representing all other con- sumption of market goods (x). We adjust the units of the numeraire good so its price (as is usual) is 1. Our starting point involves consumption of (x;,4,)s yielding utility U,. Consider an alternative with increased consumption of the environmental good: con- sumption of (x,,4,), yielding utility U, (>U,). We are interested in two numbers: (1) start- ing at q,, the WTP to move to q,; and (2) starting at q,, the WTA to move to q,,"" Digitiz Original HARVARD UNIV 148 CHAPTER 7 DEMAND FOR ENVIRONMENTAL GOODS FIGURE 7.4 Willingness to accept vs. willingness to pay for a change in quantity of an environmental good «¢) ‘This situation is illustrated in Figure 7.4. Shown in the figure are indifference curves between the environmental good and the numeraire good. WTP and WTA can be measured in units of x, since the price of xis 1. WTP is the distance BC, since going from , to q, increases utility to U,. our individual would be willing to give up as much of x as will keep him at the original level of utility, U,, WTA, on the other hand, is the amount our individual would be willing to take in liew of moving from q, to q,. Since 4, Yields utility U,, our individual would require enough x to keep utility at U, which is the distance AD. Although Figure 7.4 has been drawn so that WTA > WTP, simply drawing a fig- ure does not make it so. Suppose for instance that the indifference curves were close to linear (straight parallel lines). It is easy to see that in that case WTA = WTP, There are two situations that would result in such linearity. One is if q and x were perfect substitutes. [fone more unit of q could be offset (keeping utility constant) by a fixed amount of x, then we would have perfect substitutes. In terms of the demand curve for 4g, if it is perfectly elastic (and the price of x is fixed), then this condition would hold. Another case that would result in nearly linear indifference curves is if the change in q from q, to q, is “very small.” If the change in g and the corresponding change in U is small, then we are dealing with a much smaller region of the graph in Figure 7.4, where indifference curves are locally nearly linear. The easiest way to see that WTA and WTP are similar in this situation is to mentally move q, to q, in the figure, simul- taneously (in your head) moving U, toward U,, The difference between WTA and WTP shrinks to zero. ‘What conditions might lead to a significant divergence between WTA and WTP? One condition is if the change in the environmental good is large while x is not a perfect substitute for q. There are a number of conditions that yield this outcome. One necessary condition is that the indifference curves be curved (nonlinear), as drawn in Figure 7.4. Phis will occur with a downward sloping demand curve for q (which would be the normal situation). But it is also necessary that the change in q be substantial (nonmarginal). This ‘would occur if the change in the environmental good is really big or, as is more likely the case, the environmental good is unique and losses in it are not easily compensated for by extra consumption of market goods.

You might also like