Valuation 6: CVM, WarmGlow and Choice Modelling

1
Valuation 6 CVM, Warm-Glow and Choice Modelling

• An application to waterfowl
• Embedding and possible causes
• Warm-glow and its effect on donations to public
  goods
• Alternatives to CVM

2
Last week we looked at

• The contingent valuation method, and its many
  potential biases
• Among them are the deviation between WTP and WTAC
  suggestively contradiction micro-economic
  theory
• And the part-whole/embedding problem suggesting
  that people do not answer what it is being asked

3
Waterfowl - Survey

• Study Bill Desvouges and colleagues, 1993
• Context Exxon Valdez oil spill
• Funding Exxon Corp.
• Mall survey, developed using focus groups,
  one-on-one pretests, and two mall pretests
• Two shopping malls in Atlanta, Georgia, outside
  the Central Flyway
• 10-12 minutes
• 1205 completed questionnaires

4
Waterfowl Survey -2

• Q1 How often have you heard about waterfowl
  (select number)
• Q2 Is protecting waterfowl important to you
  (select reasons)
• Show way of Central Flyway second highest number
  of migratory waterfowl, 8.5 million a year
• Q3 How would you rate your knowledge (low, mid,
  high) of threats to the waterfowl (oil spills,
  waste oil, wetland destruction, herbicides and
  pesticides)

5
Waterfowl Survey -3

• Describe waste-oil holding ponds
• In 1989, N ducks died there
• Ponds could be covered by nets, Federal
  Government considers this, Fish and Wildlife
  Service would monitor and enforce
• Q4 Think about your income, expenses,
  alternatives. What is the most that your
  household would agree to pay each year in higher
  prices for wire-net covers to prevent about N
  migratory waterfowl from dying each year in
  waste-oil holding ponds in the Central Flyway?

6
Waterfowl Survey -4

• Q5 Is the amount greater than zero. If yes,
  select most important reason
• Q6 If no, select most important reason
• Q7 Indicate agreement to statements
• Q8 Ditto for waterfowl
• Q9 Activities
• Q10 Age Q11 Education Q12 Sex Q13 Race Q14
  Income Q15 Household size Q16 Membership

7
Waterfowl Results

• 398 answered for N2,000 408 for N20,000 399
  for N200,000
• 29 as outliers (3), protest bids (8), unlikely
  (1) and rubbish (17)
• WTP (2000) 59 ? 16 /household/year
• WTP (2000) 59 ? 10 /household/year
• WTP (2000) 71 ? 15 /household/year
• Not significantly different!

8
Waterfowl Reasons

• Desvousges et al. we find that CV yields
  estimates that fail to meet several basic
  criteria for accuracy
• Diamond and Hausman responses to CV questions
  are not consistent with the basic economic theory
  of choice
• People cannot count
• People do not listen
• People realised that 2,000 or 200,000 ducks is
  small compared to 8.5 million
• Warm glow

9
Some Puzzles

• In a large economy, no one should contribute to
  public goods like the Red Cross, the Salvation
  Army, Greenpeace yet they do
• Government support should crowd out charitable
  donations but it does not
• This suggests that people donate to public goods
  for other reasons than pure altruism social
  pressure, guilt, sympathy or warm glow may
  explain this

10
Warm Glow

• Consider
• The agent derives utility from private
  consumption, from the public good, and from
  donating
• This can be rewritten as

11
Warm Glow -2

• The donations function
• The marginal propensity to donate for altruistic
  reasons is
• The marginal propensity to donate for egoistic
  reasons is

12
Warm Glow -3

• The marginal propensity to donate is
• The marginal propensity to donate for egoistic
  reasons
• If we cut G-i by 1 and increase wi by 1 then
  the actor would spend some money on wi and some
  on G so that G would fall
• Warm glow leads to higher donations
• Warm glow leads to embedding

13
Alternatives to CVM

• (Contingent) choice modelling, or conjoint
  analysis, is similar to contingent valuation in
  that it is a stated preference technique based on
  surveys
• The main difference is that instead to asking
  who much are you willing to pay, the question
  is which situation would you prefer
• Choice modelling is choice experiments,
  contingent ranking, contingent rating and paired
  comparisons

14
Choice Experiments

• Which of the following two schemes do you favour?
• Native woodland 500 ha 700 ha
• Heather moorland 1200 ha 0 ha
• Lowland hay meadow 200 ha 300 ha
• Additional tax 25 15
• I would prefer A B Neither

15
Contingent Ranking

• Rank the alternative policy options
• Woodland 500 ha 100 ha 700 ha
• Moorland 1200 ha 600 ha 0 ha
• Meadow 200 ha 0 ha 300 ha
• Tax 25 5 15
• Your ranking 1 2 3

16
Contingent Rating

• How strongly would you prefer the following?
• Native woodland 500 ha
• Heather moorland 1200 ha
• Lowland hay meadow 200 ha
• Additional tax 25
• 1 2 3 4 5 6 7 8 9 10
• Very low preference Very
  high preference

17
Paired Comparisons

• Which of these do you prefer?
• Woodland 500 ha 700 ha
• Moorland 1200 ha 0 ha
• Meadow 200 ha 300 ha
• Tax 25 15
• 1 2 3 4 5 6 7 8 9 10
• Strongly prefer A
  Strongly prefer B

18
Choice Modelling

• Provided that do nothing is included, choice
  experiments and contingent ranking can be used to
  estimate WTP or WTAC
• If do nothing is not included, the set of
  options may be infeasible for the interviewee,
  resulting in nonsensical results
• Contingent rating does not yield WTP, as there is
  only one alternative
• Pairwise comparison is like a referendum, but
  with shades of grey that are difficult to
  interpret