Valuing the Environment Chapter 4 © 2004 Thomson Learning/South-Western.

Valuing the Environment

Chapter 4

© 2004 Thomson Learning/South-Western

2

Introduction Exxon oil spill 1989 Prince William Sound,

Exxon accepted liability for1. Cost of Clean up, and2. Compensation for damage caused to local

ecology How do we determine the appropriate

compensation? What is the value of the damage to the ecosystem?

Techniques to measure some or all of this value considered in this chapter

3

Benefit-cost analysis Reagan issued Executive Order 12291 in 1981

requiring federal agencies to do Benefit/Cost analysis for major regulatory decisions, Presidents since then have reissued

Main idea is to compare benefits of a regulation/rule to its costs

Three steps in BCA for environmental regs1. Identify those things damaged: plants, animals, human

health, aesthetics, etc.2. Determine /estimate relationship between polluant level and

damage levels3. Place monetary value on damages

Many issues in first two steps, but largely outside of economics, focus on last here

4

What is value?

First, the economic view of “value” is anthropocentric.

This means value is determined by people and not by either natural law or government.

Second, value is determined by peoples’ willingness to make trade-offs.

When an individual spends money on one good, there is less available for other goods.

5

Marginal Willingness to Pay Function

6

Consumer Surplus Consumer surplus is defined as the amount a

consumer is willing to pay beyond what they actually have to pay- thus, they receive “surplus” value from the good

Area under demand function and above price measures consumer surplus

If one has a demand function, then it is straightforward to compute consumer surplus

Example: if my demand is Q = 10 – 0.5P, or

P=20-2Q, then CS = .5(P*-P)/Q

7

Consumer Surplus Calculation

CS = .5(P*-P1)/Q1

If P1= $10, then CS=$25

If P1= $6, then CS=$49

If P1= $14, then CS=$9

P*

8

Value and Non-market Goods

Non-market goods are goods that do not typically trade in markets

Air quality, wildlife habitat, clean water, biodiversity, scenic landscapes, lack of climate change, etc.

But, time and other opportunities are sometimes sacrificed in order to obtain both market and non-market goods.

Examining these trade-offs can serve as a basis for valuing non-market goods.

9

Value and Non-market Goods

Non-market goods may have both direct use and indirect use values.

Direct use values are associated with tangible uses of environmental resources, such as use in manufacturing processes, recreational activities, or when environmental quality affects human health.

10

Value and Non-market Goods

Indirect use values are those associated with more intangible uses of the environment, such as aesthetic benefits or satisfaction derived from the existence of environmental resources.

Indirect use values are also called passive use and nonuse values.

Indirect use values include existence value, bequest value, altruistic value, option value and the value of ecological services.

11

Indirect Use Value – Bequest Value Bequest value refers to the fact that an

individual values having an environmental resource or general environmental quality available for his/her children or grandchildren to experience.

It is based on the desire to make current sacrifices to raise the well-being of one’s descendents.

12

Indirect Use Value – Existence Value Existence value refers to the fact that an

individual’s utility may be increased by the knowledge of the existence of an environmental resource even though the individual has no current or potential direct use of the resource.

An individual may never have opportunity to see a whale but places value on preserving the species.

13

Indirect Use Value – Altruistic Value Altruistic value occurs out of one individual’s

concern for another.

A person values the environment not just because that person benefits from the environmental quality but because the person values the opportunity for other people to enjoy high environmental quality.

14

Indirect Use Value – The Value of Ecological Services Ecological services include nutrient cycling,

atmospheric processes, carbon cycling, clean air, clean water, and biodiversity.

Since people do not always understand the link between ecological services and their well-being, it makes the valuation process more difficult.

15

Techniques for Measuring the Value of Non-market Goods The three major categories for measuring the

value of non-market goods include: Revealed preference techniques, which look

at decisions people make in reaction to changes in environmental quality.

Stated preference techniques, which elicit values directly through survey methods.

Benefits-transfer technique, which look at existing studies for value of analogous environmental change.

16

Revealed Preference Approaches- Travel Cost Model The travel cost model is a method for valuing

environmental resources associated with recreational activity and was first proposed by Harold Hotelling in 1947.

The basic premise is that travel cost to a site can be regarded as the price of access to the site.

Multiple observations on travel cost and quantity of visits can be used to estimate a demand curve.

17

Revealed Preference Approaches- Travel Cost Model Once a demand curve has been estimated, the

value of the site to an individual can be estimated by computing consumer surplus for each individual in the survey, calculating an average consumer surplus and then multiplying by the estimated number of recreationists.

The travel cost demand curve is often expanded to include other explanatory variables such as age, income, family size, education level, and other socioeconomic variables.

18

EXAMPLE

Discuss Iowa Lakes Project and study here

19

Revealed Preference Approaches Hedonic pricing techniques are based on the

theory of consumer behavior that suggests that people value a good because they value the characteristics of that good rather than the good itself.

An examination of how the price of the good varies with change in the levels of these characteristics can reveal the prices (value) of the characteristics.

20

Revealed Preference Approaches- Hedonic Pricing Techniques Assume that all the characteristics of houses

and neighborhoods are the same throughout the city.

Houses with higher air quality would have higher prices.

This positive relationship can be represented by the following equation: H = a + bQ, where H is housing price, Q is air quality and “b”

tells the researcher how many units H will increase with each unit of air quality.

21

Revealed Preference Approaches- Hedonic Pricing Techniques In reality prices of houses are dependent on many

different characteristics, all of which need to be reflected in the calculation of the impact of air quality on the price of housing.

The regression analysis can be expanded to include many right-hand side variables and can be estimated in a fashion that represents non-linear relationships among the variables.

Two types of variables have traditionally been included as explanatory variables in the hedonic housing price functions. Characteristics of the house itself Characteristics of the neighborhood

22

Revealed Preference Approaches- Hedonic Price Studies

23

Revealed Preference Approaches- Hedonic Wage Studies The hedonic wage approach is based on the

idea that an individual will choose the city in which he or she resides in order to maximize his/her utility.

The individual will consider wages and a host of other positive (educational or recreation opportunities) and negative (crime, pollution) factors.

Wages adjust to compensate people for different city characteristics.

24

Revealed Preference Approaches- Hedonic Wage Studies Suppose a person has two job offers, one in a cold

weather city and the other in a warm weather city. Suppose each job offers the same salary. If the person chooses the warm weather job, and

others do too, the labor pool will increase in the warm weather city and wages will fall.

The reverse happens in the cold weather city. The difference between the wages in the warm

weather city and the cold weather city compensates people for the disutility of living in the cold weather.

25

Revealed Preference Approaches- Hedonic Wage Studies This compensating differential can be used to look

at value placed on environmental amenities or risk. Value of Statistical Life studies: how much are

people willing to trade off income (wages) to accept riskier jobs?

Thaler and Rosen: collected information on probability of death on job and wages; deduced that people were willing to accept $390/year in lower wages to reduce risk to life by .001

Their estimate of Value of Statistical Life is $390,000 per year ($390 x 1000)

26

Revealed Preference Approaches- Value of Statistical Life Does this apply to everyone in population?

People who like risk likelier to take riskier jobs so those of us who don’t like risk would probably need higher compensation

Trade offs for occupational risk may not be same as tradeoffs for risk from environment or elsewhere

Other studies routinely show that people need higher compensation for risk they have no control of relative to risk they can control

Is it crazy to talk about the value of (statistical) life? Do we make tradeoffs like this in every day life? In policy decisions?

27

Stated Preference Techniques Stated preference techniques solicit values

directly by asking individuals hypothetical questions.

Contingent valuation is the most widely used stated preference valuation technique.

This method ascertains value by asking people their willingness to pay for a change in environmental quality.

28

Stated Preference Techniques-Contingent Valuation The questions used in contingent valuation can take

both open-ended and close-ended form. In open-ended questions, respondents are asked to

state their maximum willingness to pay. In close-ended questions, respondents are asked to

say whether or not they would be willing to pay a particular amount.

The questions must also specify the mechanism by which payment will be made.

29

Question formats

1. Open ended CV: directly ask maximum wtp for environmental improvement (or to avoid a loss)

2. Iterative bidding technique: ask whether wtp $X, if say yes, then ask $(X+y), if yes, then $(X+2y), etc. until say no

3. Payment card: Provide a list of possible payment amounts and ask respondent to circle one that is closest to their max wtp

4. Dichotomous choice (discrete choice or closed-ended): Ask respondents are you wtp $X, yes or no and have information on this bound only

30

Open Ended Questions

Example: Sutherland and Walsh LAND Econ (1985)“Assume that the only way to protect water quality in Flathead Lake and River is for all people to pay into a special fund to be used exclusively for this purpose. What is the maximum amount of money your household would be willing to pay annually to protect water quality in Flathead River and Lake?” N=171, mean; Mean response = $64/year

Problems with open ended questions Difficult to answer, not typical of purchasing decisions Especially when considering a good like environmental

quality High non-response observed in samples

31

Iterative bidding technique Farmington Experiment: Schulze, d’Arge and Brookshire (late 70s);

used iterative bidding to value visibility at Four Corners Region of New Mexico and AZ Starting point bias: tested whether ending bids depended on

starting values, found if raise starting bid by $1, get about $0.60 higher ending bid, replicated elsewhere, major bias

Strategic bias: after collected bid, told respondent they would not have to pay actual bid, but mean bid then asked if respondent wanted to revise bid; strategic behavior would raise bid if wtp>mean, lower if

wtp<mean Found only 1 person who revised bid: occupation?

Although easier to answer, iterative bidding has strong starting point bias, not used much

32

Payment Card Duffield and Patterson, joined forces with TNC to

simultaneously do actual elicitation with CV study

Used payment card; just like standard charity form to send back with check

Resource=increased stream flows in 2 Montana streams, currently badly dewatered by good potential spawning grounds for 2 endangered fish: Arctic Grayling and Yellowstone Cut Throat trout- neither game fish, so existence value

Contributions elicited for Montana Water Leasing Trust Fund to purchase water rights from farmers to increase stream flow

33

Duffield and Patterson Three treatments

1. Cash-TNC, real elicitation for $ , TNC letterhead2. Hypothetical-TNC, just like treatment #1, but

hypothetical, TNC letterhead3. Hypothetical-UM, also hypothetical, but a little

more vague, referred to a generic “trust fund,” UM letter head, etc

One hypothesis was that 1 and 2 would provide similar answers and 3 quite different

Sampled from Montana resident & nonresident fishing license holders, 200-400 responses for each treatment (sent more out for #1, fewer for #2, least for #3)

34

Duffield and Patterson findingsResidents Delivered Returned # Willing to

contributeAverage contributionPer Per

Contribution Returned

TNC-Cash 2278 205

(9%)

26

(1% delivered)

(13% returned)

$17.69 $ 2.24

TNC-Hypo 1013 193

(19%)

64

(6% delivered)

(33% returned)

$14.92 $4.64

UM 524 388

(74%)

77

(14% delivered)

(20% returned)

$15.26 $3.03

35

D & P’s conclusions

1. Contributions from among those who said they would contribute quite similar

2. Quite different behavior when look at per delivered group because much higher participation rate among 2 hypothetical treatments

3. Bottom line?

36

Discrete choice Closed-ended

Referendum style= popular, ask “Would you vote yes on a referendum that would raise your property tax by $X and use the money to preserve your local prairie…?”

Advantage of discrete choice questions Familiar social context, take-it-or-leave it is common

buying environment Simple decision problem, less non-response problems

Key disadvantage, do not obtain respondent’s wtp, only lower or upper bound

Must ask many respondents discrete choice questions using different “bid” values and then do some econometrics to estimate wtp

37

How accurate are these “discrete-choice” studies? Test by asking people to really pay for a good in an experimental

setting and compare to hypothetical group (Cummings, Elliott, Harrison and Murphy, JPE 1997)

Good purchased: booklet describing areas of contaminated groundwater in a low-income area of Albuquerque that would advise residents how to have their water tested and what to do it tests are positive, booklets cost $5 to be produced and distributed.

Described this good to subjects and showed them the booklets then said (roughly): If everyone in this room were to contribute $10, this money would be sufficient to cover the cost to produce and distribute these booklets to X households. We will have a secrete ballot to decide whether or not we will do this.

Had a vote, actually collected money in real case, asked whether you would vote yes on such a referendum in hypothetical case.

38

Results Table 1 : Unconditional Experiment Responses

Hypothetical Real Total

No 102 (55%) 73(73%) 175

Yes 84 (45%) 27(27%) 111

Total 186 100

Conclude the hypothetical say yes more often than real

(Statistically reject null that there is no difference between responses for real and hypothetical (Chi-sq test))

39

Cheap Talk

Cummings and Taylor (“Unbiased Value Estimates for Environmental Goods: A Cheap Talk Design for CVM”)

Follow up to experimental studies that “often, but not always” find significant differences between hypothetical valuation and real payments

Some have sought to find calibration factors, here attempt to induce subjects to provide accurate responses

40

Approach: Cheap Talk

Tell subjects explicitly about idea of hypothetical bias and ask them to try to avoid it

Present results from 16 different experiments: 8 baseline experiment with 4 public goods, 2 sets of referenda for each

1. A real referendum where subjects vote on proposition that requires real payments

2. Hypothetical referenda, CV Using these data, test whether there is significant

difference in yes answers, when reject equivalence, conduct follow-up with cheap talk design

41

Four goods

1. Donation to Georgia Nature Conservancy to buy additional TNC lands in Georgia

2. Albuquerque good=contributions to a program to produce a booklet about contaminated water (same as Cumming et al paper)

3. Nature Conservancy’s “Adopt an Acre” program to purchase rain forest land in Costa Rica

4. Path Foundation, greenways in Atlanta, money used to construct greenway along highways

Students from Georgia State recruited, paid $10

42

Cheap Talk script

Some excerpts

43

Findings

Referenda # Participants Percent yes Reject null ?

NC-H

NC-R

NC-CT

115

71

74

43

25

27

Yes

Yes

No

ABQ-H

ABQ-R

ABQ-CT

211

182

84

46

30

35

Yes

Yes

No

44

Findings

Referenda # Participants Percent yes Reject null ?

RF-H

RF-R

RF-CT

63

90

33

68

49

49

Yes

No

PF-H

PF-R

97

49

22

25

No

45

List: Sportscard trading experiments Follow-up to Cummings and Taylor Uses “field experiments” = sportscards

trading, actual market place, some experienced and some inexperienced traders

Three treatments1. Hypothetical2. Hypothetical with cheap talk3. Actual Vickery 2nd price auction – field

46

Experimental procedure

1. Subject approaches administrator’s table and asks about sale, administrator asks them if they want to be in auction market,

2. Administrator gives participant an instruction and bidding sheet, (about 1/3 hypo, 1/3 hypo with cheap talk, 1/3 real – varied by hour)

3. Bid placed

4. Participant told that if he won, he would be contacted by phone in 3 days if real auction

47

Results

Table 1: dealers have higher bids than nondealers For dealers, no difference (statistically) between

hypo and hypo with cheap talk, but significant difference between actual and both hypo and hypo with cheap talk

For nondealers, no difference between actual and hypo with cheap talk, but significant difference between hypo and cheap talk as well as hypo and actual

48

Conclusions

Cheap talk works for nondealers, but not dealers

why? What do we learn about the validity of this type of work?

49

Stated Preference Techniques-Contingent Valuation Although the contingent valuation method

has been widely used for two decades, there is controversy over whether it adequately measures people’s willingness to pay.

These arguments are based on informational issues and on the fact that people may be indicating their value of something other than the particular environmental issue.

50

Stated Preference Techniques-Contingent Valuation The informational issue revolves around the fact that people do

not have practice valuing environmental issues so they are not certain what they are willing to pay.

A second issue revolves around the fact that the expressed willingness to pay may be biased because the respondent wants to feel good, be thought of as a good person, or signal importance of the issue but not actual willingness to pay.

A different problem is the potential link the respondent makes between the question and other issues, for example reduced pollution and reduced health risks. This may result in double counting.

Finally, some researchers believe there is a fundamental difference between hypothetical decisions and actual decisions.

51

Stated Preference Techniques-Benefit Transfer Approaches The process of estimating values using revealed

preferences or stated preferences approaches can be quite expensive.

Benefit transfer refers to the process of taking values from studies that were previously completed in other areas, and applying them to the area where the new decision must be made.

Although benefit transfer is not a “scientific methodology” it is associated with hypothesis testing and formal confidence intervals and there is a set of alternative methodologies that can be employed to maximize the quality of information.

52

Stated Preference Techniques-Benefit Transfer Approaches It is important to use a reference study that is

congruous. If many reference studies are available, the

process becomes much easier. The appropriate reference study can be

chosen, or a weighted average of the values can be employed, where weights are chosen according to similarity between the reference study and the problem at hand.

53

Summary

Many environmental regulations require benefit/cost analysis

Ideally, comparing what people are willing to give up for environmental quality to what it costs to provide it gives decision makers important information about how much environmental quality should be provided

In practice, much is unknown Still, revealed preference approaches can help

measure direct use values. Stated preference approaches can help estimate

indirect use values.