Stated Preference

1
Stated Preference Experimental Markets

• Methods for measuring non-market benefits

2
Todays Menu

• Constructed market approaches
• Designing a CVM survey
• Conducting an experiment

3
Recall revealed preference

• Remember, our first choice is always to use
  revealed preference approaches to value
  environmental assets.
• What if value (or part of value) cant be
  captured within an existing market?
• Answer use constructed market methods

4
What are constructed market methods?

• Stated preference methods
• More generally known as Contingent Valuation
  Methods (CVM)
• Ask people How much is the environmental good
  worth to you?
• Experimental markets
• Typical goal learn how people think about
  environmental goods
• Two types of experiments
• Field
• Laboratory

5
Some examples

• The value of the Channel Islands marine reserve?
  
• Do people value the organisms or the reserve?
• Should DDT plume off of Rancho Palos Verde be
  cleaned up with state funds?
• How large are the wildlife damages from Exxon
  Valdez spill?
• The value of restoring steelhead to Ventura
  River?
• The environmental damage from drilling in ANWR?

6
How elicit willingness to pay?

• No market exists within which to measure value
• Use survey
• Many challenges to do a credible job
• Need to design survey to minimize opportunities
  for bias of results.

7
Value of solar-generated electricity

• Introduction
• Explain purpose of survey and how long it will
  take, importance of truthful answers
• Background attitudes
• Solar as green electricity
• Compare to brown electricity
• Importance of air quality, current opinion about
  air quality
• Have you participated in green markets?
• Eliciting value of green electricity
• Concept of Renewable Energy Credit (REC)
• Would your purchase a REC for X? (vary X)
• Socioeconomic characteristics
• Gender, Age, Race, Income, Greenness

8
Example Seoul CVM Survey Instrument

• Introduction explain purpose of survey and time
  to complete (30 min)
• Request background attitudes and info
• What is your opinion about current tap water in
  Seoul? (very good, good, etc)
• In the last five years have you or anyone in your
  household taken any of these steps to obtain
  cleaner water? (install water filter, etc)
• Value of Water Quality Improvement
• Explain major pollution accident in 1991 that
  contaminated drinking water of Seoul
• If the government takes no action, how many times
  in the next five years do you think we will have
  a similar accident?
• Give one minute description of continuous
  monitoring system that will warm your household
  of contamination
• What is the most your household would pay in
  taxes for such monitoring?
• Background information
• Age, education, household size, income

9
Cookbook procedure 1 of 2

• Define the market scenario or payment vehicle
• How much are you willing to pay for vs.
• Bond issue has been proposed to pay for
• Determine elicitation method
• Direct question, discrete choice, bidding game,
  payment card, etc
• Design elicitation scheme
• Mail, telephone, email, web, in person

10
Cookbook procedure 2 of 2

• Determine sample design
• Population, randomization of respondents,
  randomization of questions, etc.
• Determine experimental design
• How obtain demand curve and conduct further
  analysis to answer question
• Estimate demand function
• Conduct analysis to answer question

11
Problems with CVM

• WTA or WTP?get different answers.
• Hypothetical bias
• How real is the budget constraint?
• Ambiguity about what is being valued
• Warm Glowpurchase of moral satisfaction?
• Open-ended is unfamiliar market context
• Embedding problem
• Value of one park? Value of suite of parks?
• Existence value is a problematic concept since it
  is unobserved

12
NOAA guidelines for CVM

• Minimize non-response
• Personal interviews
• Pretest for interviewer effects etc.
• WTP not WTA
• Referendum format
• Provide adequate background info.
• Remind of substitute commodities
• Include explain non-response option (not 0)

13
Types of payment vehicle

• Payment affects all parties
• National tax, local tax, fee or charge for use,
  price increase for use.
• Note respondent may disagree with agency
  responsible for managing resource.
• Voluntary payment
• Donation to trust fund
• Note remember free-riding problem, also may get
  strategic bias.

14
Open-ended elicitation

• What is the maximum amount you would be prepared
  to pay every year vehicle to XXX?
• For
• Straightforward, no implied value cues/anchoring
  bias, gives max WTP
• Against
• Large non-response/protest, unrealistically large
  bids, unreliable, unlike normal market transaction

15
Bidding game elicitation

• Would you pay X every year via vehicletax,
  utility bill? keep increasing bid until
  answer is No or decrease until Yes
• For
• Forces respondent to consider preferences.
• Against
• Anchor bias, yea-saying, cannot be used in mail
  surveys.

16
Payment card elicitation

• Which of the amounts listed below describes your
  maximum WTP every year using vehicle Y to
  improve X? list of values
• For
• Avoids starting point bias, values can be actual
  tax or household benchmarks
• Against
• Range of numbers can induce bias, cannot be used
  on telephone.

17
Single-bounded dichotomous choice elicitation
(referendum)

• Would you pay X per year using vehicle to
  improve Y? randomly vary X
• For
• Simplifies choice (similar to market), minimizes
  non-response, straightforward
• Against
• May get inflated values, some yea-saying, less
  informative, starting point bias.

18
Double-bounded dichotomous choice elicitation

• Would you pay X every year using vehicle to
  improve Y? If yes And would you pay Z (gtX)?,
  If no And would you pay W (ltX)?.
• For
• More efficient than referendum (because know
  bounds.
• Against
• Same as referendum

19
Major sources of error in CVM

• Scenario misspecification
• Divergence between what respondent is answering
  and what researcher is asking
• Implied value cues
• Unfamiliar choice problem, respondent wants to
  give right answer.
• Strategic bias
• Low bid (think taxes will inc. but good wont)
• High bid (think wont have to pay for it)

20
Reliability of CVM estimates

• Compare results with comparable revealed
  preference results
• Construct market compare results
• Use CVM to measure demand for market good
• Test method (same sample over time)
• Surveys of purchase intentions and actual
  purchases (market research)

21
1. Green Power Demand

• Objective derive a demand curve for renewable
  energy credits in Davis, CA
• Advise REC company on pricing and capacity
• Bren GP, 2005
• 400 survey responses from Davis residents
• Socioeconomic characteristics, political,
  knowledge of green power, etc.
• Method contingent valuation survey distributed
  in person to Davis residents
• Regression of Yes/No on characteristics, price,
  etc

22
Regression results
23
Predicted demand
24
2. Green building demand

• 2006 Bren GP
• Demand for houses constructed of alternative
  green materials
• Primarily compressed straw block
• Key issues
• Reliability, energy savings, durability,
  earthquake, appearance, etc.
• Method contingent valuation survey by Knowledge
  Networks 1000 responses from across US.
• Regression of Yes/No on price, socioeconomics,
  location, rainfall, temp, etc.
• Second survey to industry asked to predict
  consumer demandconsistent with consumer
  responses?

25
Consumer vs. Industry
26
Spatial density of demand
27
3. Goose Hunting in Wisconsin

• Duck licenses limited in number and issued by
  lottery
• Compare three techniques
• Travel Cost
• CVM
• Experimental Methods
• CVM WTP and WTA bracketed true value

28
4 Ecotourism in Kenya

• Ecotourism captures some of WTP for preserving
  wildlife
• Lake Nakuru Natl Park in Kenya (360 species of
  birds)
• Farming has reduced water quality and subsequent
  wildlife pops.
• CVM estimates value park (value to users only) at
  7.5 million.

29
5 Economic value of noncommercial fish (US)

• Rivers in Four Corners region
• 2465 miles of river habitat for 9 endangered
  fish.
• Protection requires fish bypasses, passageways,
  habitat improvements.
• What is the economic value of preserving the
  habitat?

30
The application (Four Corners fish)

• Respondents given
• Maps of critical habitat
• Told that officials though habitat protection too
  costly, were going to eliminate critical habitat
  designation
• Asked if they would contribute to Four Corners
  Endangered Fish Trust Fund
• Also told all taxpayers would contribute

31
Voting for Four Corners fish

• If majority votes in favor of contribution to
  fund
• Rivers managed for optimum fish protection. Fish
  will be saved removed from ESA over next 20
  years.
• If majority votes against
• Rivers managed for maximum hydroelectic output. 4
  of 9 would likely become extinct.

32
Actual survey instrument

• Suppose a proposal to establish a Four Corners
  Region Threatened and Endangered Fish Trust Fund
  was on the ballot in the next nationwide
  election. How would you vote on this proposal?
  Remember, by law, the funds could only be used to
  improve habitat for fish. If the Four Corners
  Region Threatened and Endangered Fish Trust Fund
  was the only issue on the next ballot and it
  would cost your household ______ every year,
  would you vote in favor of it? Circle Yes or No.
• Dollar value randomly chosen from 1 - 350.

33
Results of Four Corners fish

• Survey sent to random sample of 800 Four Corners
  households. Additional 800 to households in rest
  of US.
• Average WTP 195.
• Extrapolated to rest of population.
• Benefits far exceeded costs.

34
Experimental markets

• Main criticism of CVM is that it is hypothetical.
• Develop an experimental market, with actual
  participants to elicit values.
• Researcher constructs market, including the
  good(s) and money
• Observe behavior of subjects (may be field or
  laboratory)