Warm glow in biodiversity valuation and policy
by Paulo A.L.D. Nunes
Institutions for Providing Global Environmental Goods Louvain-la-Neuve, 15|16 June 2006
Warm glow in biodiversity valuation and policy by Paulo A.L.D. - - PowerPoint PPT Presentation
Warm glow in biodiversity valuation and policy by Paulo A.L.D. Nunes Institutions for Providing Global Environmental Goods Louvain-la-Neuve, 15|16 June 2006 Outline Motivation and background Research proposal: assessment of the profile
Institutions for Providing Global Environmental Goods Louvain-la-Neuve, 15|16 June 2006
2
Motivation and background Research proposal: assessment of the profile of a warm glower Discussion of the estimation results (two studies) Policy implications and conclusions
3
We illustrate the embedding problem with the following experiment: Group 1: willingness to pay X for the protection plan A Group 2: willingness to pay Y for the protection plan B Group 3: willingness to pay Z for the protection plans A and B Embedding is present when X+Y>Z, => the rejection of the adding-up valuation hypothesis.
4
5
Verbal protocols allow researchers to register how respondents think while they answer the elicitation
mental models that are able to account for the presence of a part-whole interpretation, i.e., the respondent values much more than the researcher intended to value: – In practice: when the respondent is asked the value of each public good independently, e.g. X and Y, such an individual corrects the “foolish” question asked by a “silly” researcher and provides a value for the total provision of the public good, Z.
6
View CV responses as a way to signal warm glow considerations, interpreted as the pleasure derived from giving to good causes or simply from concern over the environment. If this warmglow or moral satisfaction presents rapidly diminishing marginal utility with respect to the size of the gift, then we may expect that respondents will bid about the same for obtaining moral satisfaction. – Bearing on our illustration, we would have X=X’+C; Y=Y’+C and Z=Z’+C where C is the value attached to the warmglow
– Then the sum of the individual protection plans will be Z’+2C, which is greater than Z=Z’+C, the value of the two protection plans jointly.
7
Jerry Hausman, Peter Diamond, William Desvousges and Paul Milgrom express their doubts with respect to the suitability of CV results for inclusion in benefit- cost analyses - “Is a number better than no number?” The skepticism is particularly strong when the CV application is addressed to measure nonuse values. For nonuse values, as measured by survey methods, to be consistent with economic theory “it would be necessary for respondent’s individual existence values to reflect only their own personal economic motives and not altruistic motives, or a sense of duty,
8
Tendency for CV respondents to signal noneconomic considerations, indicating that respondents biding captures inter alia consumer welfare associated with the personal satisfaction provided by the act of giving – Kahneman and Knetsch, 1992, JEEM moral satisfaction – Nunes, 2002, EJOR, consumer motivationings and warm glow motivation function – Nunes and Schokkaert, 2003, JEEM, latent valuation function, warm glow valuation mechanism and dry WTP
9
This idea is inspired by the work of Andreoni on impure altruism: (1989, Journal of Political Economy;1990, Economic Journal) The underlying idea: individual consumer contributes to the provision of a public good for two reasons. First, because she simply wants more of the public good and, secondly, because she derives some private benefit from contributing to its provision.
Ui (Yi, G, xi) with G = xi +xj Ui (Yi, G, xi) with G = xi +xj
10
11
Our family admires the individuals who, on voluntarily basis, participate in collecting donations for national programs for social aid and solidarity (8WG.) I am happy with myself whenever I give a financial contribution to national fund raising campaigns (WG20.) Our family admires the individuals who, on voluntarily basis, participate in collecting donations for national programs for social aid and solidarity (8WG.) I am happy with myself whenever I give a financial contribution to national fund raising campaigns (WG20.)
12
Our family admires the individuals who, on voluntarily basis, participate in collecting donations for national programs for social aid and solidarity (8WG.) I am happy with myself whenever I give a financial contribution to national fund raising campaigns (WG20.) Our family admires the individuals who, on voluntarily basis, participate in collecting donations for national programs for social aid and solidarity (8WG.) I am happy with myself whenever I give a financial contribution to national fund raising campaigns (WG20.) RESPONSE CATEGORIES WORDING (CODING): I COMPLETELY AGREE (5) I AGREE (4) SOMETIMES I AGREE, SOMETIMES I DISAGREE (3) I DISAGREE (2) I COMPLETELY DISAGREE (1) RESPONSE CATEGORIES WORDING (CODING): I COMPLETELY AGREE (5) I AGREE (4) SOMETIMES I AGREE, SOMETIMES I DISAGREE (3) I DISAGREE (2) I COMPLETELY DISAGREE (1)
13
Consumer motivation latent factors (see Nunes, 2002, EJOR)
14
Estimate Survey motivational question 0.56 Our family admires the individuals who, on voluntary basis, participate in collecting donations for national programs for social aid and solidarity. 0.60 There are some funding campaigns to which my family and I feel very close and therefore we do not hesitate to contribute with a donation. 0.47 It is difficult for me to decline my help to other individuals who, either in the streets or at my door, beg for charity. 0.57 I am happy with myself whenever I give a financial contribution to national fund raising campaigns. 0.58 My family and I like to contribute to good causes such as the protection of the environment, and whenever we can afford it, we do not decline our help to such fund raising campaigns.
Source: Nunes (2002)
15
Integrating latent motivational factors in the valuation function (see Nunes and Schokkaert, 2003, JEEM)
Estimate individual warm glow motivation scores and plug it into the valuation function
ij r = β
j kj ki k
hi h
BAU Latent structure
16
Integrating latent motivational factors in the valuation function (see Nunes and Schokkaert, 2003, JEEM)
Estimate individual warm glow motivation scores and plug it into the valuation function
ij r = β
j kj ki k
hi h
BAU Latent structure
17
Explore in more detail de warm glow transmission mechanism, focusing on, inter alia,
consideration (and its provision levels),
individuals, or not
18
Two valuation studies: CV: one valuation object and explore individual warm glow motivational profile ABM: two valuation objects and warm glow motivational profile per stakeholder
19
glower’? And if yes, what personal, socio-economic, ideological and ethical variables characterize such an individual?
altruism, where individuals’ contribution to the public good enters into her utility function twice: firstly as a contribution to the public good provision and secondly as a private good.
valuation transmission mechanism. In which type of contributors is this valuation transmission mechanism “stronger”?
j warmglow j k k k j warmglow
, , ,
20
21
The profile of warm glowers and respective inspiration factors
“Ego Driven” Warm Glowers “Socially Oriented” Warm Glowers
Inspiration factors Inspiration factors
Derive a personal advantage or satisfy a personal interest. Derive a feeling of usefulness or goodness towards society and himself. Strengthen personal reputation in the local community. Feel more socially integrated; feel more cohesion with society as a whole, feel less social emargination.
22
Variables Parameter estimate Standard error p value Intercept
0.28322406 0.465 ‘No-No’ WTP answers
0.07426440 0.812 WTP (with WTP > 0) 0.012984 0.01164136 0.265 WTP2 (with WTP > 0)
0.00173268 0.206** Personal features Age
0.02604717 0.178** Gender = male
0.06018154 0.002* Recreational consumption features Beach visitors Surfers Bikers Trekkers Camping Hunters
0.014737 0.030597 0.153364
0.005262 0.13928771 0.06083655 0.07277344 0.17202065 0.06038607 0.06148617 0.889 0.808 0.674 0.372 0.646 0.931 Ideological features Catholics
0.03000387 0.002* Information on public policies issues
0.06654169 0.584 Opinion with respect to the conservation of nature
0.08663370 0.214** Medical assistance and social security 0.074389 0.04509218 0.099* Environmental protection 0.036043 0.03508173 0.304 Public security 0.019357 0.03328786 0.561 Unemployment
0.04503069 0.390 Quality of public education system
0.03157677 0.008* *(**) Significant at 10% (25%)
+ + +
23
Variables Parameter estimate Standard error p value Charitable behavior Financial donation (via organizations) 0.381388 0.08082874 0.001* Charity (on the streets) 0.490620 0.08211214 0.001* Blood donation 0.008321 0.06652234 0.900 Material wealth features Net income
0.02767285 0.974 Luxury lodging 0.408643 0.26555291 0.124** Low rank lodging 0.076524 0.07268706 0.292 Sociological features Household dimension
0.02800090 0.111** Rural areas 0.157144 0.12514521 0.209** Urban areas
0.06435315 0.280 Executives or directories Unskilled workers Retired
0.158039 0.015915 0.15359910 0.15513062 0.10738206 0.024* 0.308 0.882 Education level
0.02481894 0.003* Local inhabitants 0.396502 0.13085086 0.002* R2 0.196 *(**) Significant at 10% (25%)
+ + + + + +
24
Variables Parameter estimate Standard error p value Charitable behavior Financial donation (via organizations) 0.381388 0.08082874 0.001* Charity (on the streets) 0.490620 0.08211214 0.001* Blood donation 0.008321 0.06652234 0.900 Material wealth features Net income
0.02767285 0.974 Luxury lodging 0.408643 0.26555291 0.124** Low rank lodging 0.076524 0.07268706 0.292 Sociological features Household dimension
0.02800090 0.111** Rural areas 0.157144 0.12514521 0.209** Urban areas
0.06435315 0.280 Executives or directories Unskilled workers Retired
0.158039 0.015915 0.15359910 0.15513062 0.10738206 0.024* 0.308 0.882 Education level
0.02481894 0.003* Local inhabitants 0.396502 0.13085086 0.002* R2 0.196 *(**) Significant at 10% (25%)
+ + + + + +
Red: socially oriented warm glower (a la Andrioni)
Derive moral satisfaction or warm glow because they will feel directly related and responsible for the provision of the public and, in this way, experience a tighter sense or sentiment of social- cohesion together and ultimately, seek a social return. We call this category of ‘social-oriented’ warm-glowers.
Red: socially oriented warm glower (a la Andrioni)
Derive moral satisfaction or warm glow because they will feel directly related and responsible for the provision of the public and, in this way, experience a tighter sense or sentiment of social- cohesion together and ultimately, seek a social return. We call this category of ‘social-oriented’ warm-glowers.
25
Variables Parameter estimate Standard error p value Charitable behavior Financial donation (via organizations) 0.381388 0.08082874 0.001* Charity (on the streets) 0.490620 0.08211214 0.001* Blood donation 0.008321 0.06652234 0.900 Material wealth features Net income
0.02767285 0.974 Luxury lodging 0.408643 0.26555291 0.124** Low rank lodging 0.076524 0.07268706 0.292 Sociological features Household dimension
0.02800090 0.111** Rural areas 0.157144 0.12514521 0.209** Urban areas
0.06435315 0.280 Executives or directories Unskilled workers Retired
0.158039 0.015915 0.15359910 0.15513062 0.10738206 0.024* 0.308 0.882 Education level
0.02481894 0.003* Local inhabitants 0.396502 0.13085086 0.002* R2 0.196 *(**) Significant at 10% (25%)
+ + + + + +
Red: socially oriented warm glower (a la Andrioni)
Derive moral satisfaction or warm glow because they will feel directly related and responsible for the provision of the public and, in this way, experience a tighter sense or sentiment of social- cohesion together and ultimately, seek a social return. We call this category of ‘social-oriented’ warm-glowers.
Red: socially oriented warm glower (a la Andrioni)
Derive moral satisfaction or warm glow because they will feel directly related and responsible for the provision of the public and, in this way, experience a tighter sense or sentiment of social- cohesion together and ultimately, seek a social return. We call this category of ‘social-oriented’ warm-glowers.
Blue: ego driven warm glower
Derive moral satisfaction or warm glow because they want to obtain a personal advantage or a personal sense of pride from their contribution, because they will obtain a social status, or because they feel that they will obtain a social status, from the act of contributing itself. We call this category as ‘ego driven’ warm-glowers
Blue: ego driven warm glower
Derive moral satisfaction or warm glow because they want to obtain a personal advantage or a personal sense of pride from their contribution, because they will obtain a social status, or because they feel that they will obtain a social status, from the act of contributing itself. We call this category as ‘ego driven’ warm-glowers
26
Stated choice valuation method Choice between the current management plan and a set of alternative management plans Difference attributes (including biodiversity status)
27
Cockle fishery in the Wadden Sea:
birds -- Identified areas where fishing is not allowed
sediment structure of the Wadden Sea in an irreversible way (abiotic)
Sea (biotic)
28
(The Dutch) Wadden Sea
29 Current situation (A) Policy proposal (B ) Policy measures Surface area where it is allowed to fish
Allowed number of cockles harvest Rotation or fixed areas where it is allowed to fish on cockles Current area Current level No rotation Half of the current level Lower level No rotation Environmental quality: Number of the local bird population Current level More than in current situation Costs: Costs per household 0 euro 50,- euro
30
31
Dutch citizens Tourists Local residents Policy- makers Natural scientists Constant term 0.56*** (7.94) 0.24* (1.93) 0.711*** (6.19) 0.371 (0.944)
(-1.043) Level of birds Less birds
(-8.57)
(-6.17)
(-3.01)
(-1.514)
(-0.580) More Birds 0.684*** (15.13) 0.80*** (8.35) 0.50*** (6.13) 0.0053 (0.020) 0.012 (0.044) Much more birds 0.697*** (13.86) 0.98*** (9.06) 0.362*** (3.99) 0.870*** (2.930) 0.558* (1.768) Price
(-33.17)
(-16.19)
(-11.83)
(-3.085)
(-3.789) Log likelihood
12981 2932 3578 293 218 Adjusted ρ2 0.17 0.15 0.13 0.08 0.11 The significance of the estimates is indicated by ***, ** and *, referring to the 1%, 5% and 10% level, respectively; t-values are between brackets.
32
Dutch citizens Tourists Local residents Policy- makers Natural scientists Constant term 41*** 24* 101*** 65
Level of birds Less birds
Current level of birds
30 More Birds 50*** 80*** 71*** 1 2 Much more birds 51*** 97*** 51*** 122*** 78*
The significance of the estimates is indicated by ***, ** and *, referring to the 1%, 5% and 10% level, respectively; t-values are between brackets.
33
Estimation model with individual warm glow motivational factor
Dutch citizens Tourists Local residents Constant term 0.48*** (6.43) 0.17 (1.33) 0.68*** (5.72) Level of birds Less birds
(-9.22)
(-6.12)
(-3.17) More Birds 0.73*** (15.30) 0.75*** (7.35) 0.56*** (6.46) Much more birds 0.76*** (14.29) 0.90*** (7.86) 0.37*** (3.80) Price
(-31.20)
(-13.98)
(-11.07) F3*more birds 0.23*** (8.58) 0.07 (1.42) 0.09** (2.15) F3* much more birds 0.13*** (4.48) 0.16*** (2.99) 0.02 (0.51) F3*price 0.002*** (4.04) 0.001** (2.08)
(-1.37) Log likelihood
12981 2932 3578 Adjusted ρ2 0.27 0.22 0.20
34
motivational profile on the choice of the specific environmental quality program, captured by ‘F3*more birds’ and ‘F3*much more birds’.
project, in this case to a specific environmental quality protection program. For this reason let us name this effect as ‘project specific warm glow’.
35
significant for the ‘Dutch citizens’ sample. In fact, respondents belonging to this stakeholder type who are, ceteris paribus, relatively sensitive to ‘project specific warm glow’ and thus revealing a higher choice propensity for environmental quality.
36
measured in terms of the ‘project specific valuation mechanism’, is **different** across the two environmental quality protection programs under consideration, ‘more birds’ and ‘much more birds’.
37
with ‘much more birds’ than in the scenario with ‘more birds’.
38
test statistic for the restriction of equal warm glow effects is 4.503, well above the 95% critical level of the chi-square distribution with one degree of freedom.
39
‘project specific warm glow’ motivation on individual choices is **not** the same for the different environmental quality protection programs under consideration.
40
mechanism, i.e., ‘project specific warm glow’ effect increases with the proposed environmental quality protection program, however it increases at a decreasing rate.
41
effect of the warm glow motivational profile. Such an effect is described by the impact of this psychological motivation on the cost and thus likelihood for contributing, which is captured by ‘F3*price’.
42
10.The respective parameter estimates capture the general feeling of warm glow of contributing, independently of the environmental quality protection program in question. For this reason let us name this effect as ‘global warm glow’ or ‘A la Andrioni’.
43
11.The empirical magnitude of this warm glow mechanism is significant for both ‘Dutch citizens’ and ‘Tourists’ sub-
stakeholders, respondents who show a relatively low responsiveness to the financial cost by choosing independently of whether the environmental program refers to ‘more birds’ or ‘much more birds’.
44
12.In other words, for the same financial cost (price), these respondents are more likely to choose for environmental protection rather than for the current situation, independently
45
13.Furthermore, the statistical magnitude of this warm glow mechanism is comparable for the two stakeholders ‘Dutch citizens’ and ‘Tourists’. 14.This result is confirmed by formal testing of the ‘F3*price’ parameter estimate across the two sub-samples. According to estimation results ‘F3*price’ is not statistically significant across the two sub-samples and for this reason may be polled to a single estimate.
46
monetary estimates expressed for a 90% level of confidence.
47
0,00283 0,00276 0,00239 0,00190 0,00142 0,00163 0,00098 0,00009 0,00087 0,00000 0,00050 0,00100 0,00150 0,00200 0,00250 0,00300 Dutch Tourists Pooled model
Estimate for F3*pice
As we can see, Figure 1 shows a large overlap between the confidence intervals regarding the two interval estimates This confirms that the magnitude of the warm glow mechanism is comparable for the two stakeholders under consideration, i.e., ‘Dutch citizens’ and ‘Tourists’.
48
Once the warm glow effect is targeted and the warm- glower is identified, what should a policy maker do? Use original WTP results or the WTP estimates corrected for warm glow effect? In our opinion, such analysis might be useful for critically tackling CV results in a public policy and implementation perspective. By comparing reported WTP estimates and WTP estimates corrected for warm glow effect two scenarios are open:
49
Once the warm glow effect is targeted and the warm- glower is identified, what should a policy maker do? Use original WTP results or the WTP estimates corrected for warm glow effect? In our opinion, such analysis might be useful for critically tackling CV results in a public policy and implementation perspective. By comparing reported WTP estimates and WTP estimates corrected for warm glow effect two scenarios are open: Reported WTP > costs of provision Dry WTP > costs of provision The CV answers reflect real economic preferences and should, therefore, be used in cost benefit analysis. In this case, cost-benefit analysis suggests to provide the good. Reported WTP > costs of provision Dry WTP > costs of provision The CV answers reflect real economic preferences and should, therefore, be used in cost benefit analysis. In this case, cost-benefit analysis suggests to provide the good.
50
Once the warm glow effect is targeted and the warm- glower is identified, what should a policy maker do? Use original WTP results or the WTP estimates corrected for warm glow effect? In our opinion, such analysis might be useful for critically tackling CV results in a public policy and implementation perspective. By comparing reported WTP estimates and WTP estimates corrected for warm glow effect two scenarios are open: Reported WTP > costs of provision Dry WTP > costs of provision The CV answers reflect real economic preferences and should, therefore, be used in cost benefit analysis. In this case, cost-benefit analysis suggests to provide the good. Reported WTP > costs of provision Dry WTP > costs of provision The CV answers reflect real economic preferences and should, therefore, be used in cost benefit analysis. In this case, cost-benefit analysis suggests to provide the good. Reported WTP > costs of provision Dry WTP < costs of provision Therefore, providing the public good creates a situation where its costs are higher its directly benefits. Reported WTP > costs of provision Dry WTP < costs of provision Therefore, providing the public good creates a situation where its costs are higher its directly benefits.
51
a) The policy maker does not take into account the warm glow effect and does not provide the public good. – A policy maker that is inspired by efficiency-oriented principles, will not pay for the public good, since the its public benefits do not cover the public costs of the provision. – In this situation, one might infer that the decision-maker does not respect consumer’s preferences (who have expressed a positive willingness to pay for the contribution of a public good). The “warm glow” motivational element can be interpreted by an efficiency-oriented decision-maker as an irrelevant valuation effect, only dry WTP matters.
52
b) The policy maker takes into account the warm glow effect and provides the public good. – A policy maker decides to take into account the “warm-glow” valuation mechanism and provides the good – In this situation, one might infer that the decision-maker does respect consumer’s preferences
The identification of the “warm-glower” can create a support in pursuing and targeting embedded public policies, in addition to the provision itself of the public good.
53
b) The policy maker takes into account the warm glow effect and provides the public good. – A policy maker decides to take into account the “warm-glow” valuation mechanism and provides the good – In this situation, one might infer that the decision-maker does respect consumer’s preferences
The identification of the “warm-glower” can create a support in pursuing and targeting embedded public policies, in addition to the provision itself of the public good.
54
The decision-maker, for instance, might be willing to provide the public good because, by taking into account “warm-glowers of the social oriented type”, she can ALSO promote social cohesion, support minorities or integration of more emarginated social groups of population. By contributing to the public good, the warm glowers might feel more socially integrated. [In the Portuguese reality, for instance, the feeling of satisfaction provided by the act of giving is stronger among unskilled workers or low educational categories].
55
56
By pursuing an warm glow interest, an individual somehow, improves society. If we reason in Pareto principle terms, we can say that a person in need is better off after being helped (for instance because she receives food or clothes that could not afford) and the donor or charity giver is better off because she feels better her/himself. Therefore the identification of a warm glower and its consideration when designing a public policy, is useful in pursuing public objectives, others than the simple provision of the public good.
57 Campo S. Maria Formosa 30122 Venezia - Italy tel +39 | 041 | 27 11 400 fax +39 | 041 | 27 11 461 web http://www.feem.it
58
respondents on the 26 attitudinal items as presented in the instrument survey;
sample respondents on the retained motivations;
between the answers on the 26 items and the respondents’ factor scores;
59
State that the underlying consumer factors scores are normalized or standardized variables; therefore they have mean zero and variances one. Moreover, consumer motivation factors scores are uncorrelated with each other
State that the underlying consumer factors scores are normalized or standardized variables; therefore they have mean zero and variances one. Moreover, consumer motivation factors scores are uncorrelated with each other
cov(Ξ Ω ) = =
ω ω
1 2 26
f Ξ =
Assert that the residual part is assumed to have zero mean and have specific variance - given by the main diagonal of – declaring also that the residual parts are uncorrelated with each
Assigns that the residual component and the constructed motivation factors scores are uncorrelated with each other. Assert that the residual part is assumed to have zero mean and have specific variance - given by the main diagonal of – declaring also that the residual parts are uncorrelated with each
Assigns that the residual component and the constructed motivation factors scores are uncorrelated with each other.
60
61
The communalities, however, are not known, since they are the elements of information based on the unobserved factor loadings. The main objective is then to estimate the communalities such that the underlying motivational structure is able to reproduce the sample correlation matrix, in place of cov(av) i.e. We propose to estimate the communalities using the squared multiple correlation. This way we are able to neglect and factor S into
62
Finally, and as to find a frame of reference where the retained factors are more interpretable, the estimated matrix will be submitted to a rotation and * = T (where T is orthogonal) will be obtained. We propose the varimax method as the selected
characterized by seeking rotated loadings that maximize the variance of the squared loadings in each column of *.
Λ ˆ Λ ˆ Λ ˆ Λ ˆ
63
Finally, and as to find a frame of reference where the retained factors are more interpretable, the estimated matrix will be submitted to a rotation and * = T (where T is orthogonal) will be obtained. We propose the varimax method as the selected
characterized by seeking rotated loadings that maximize the variance of the squared loadings in each column of *.
Λ ˆ Λ ˆ Λ ˆ Λ ˆ
Intuition: Varimax attempts to make the factor loadings, i.e., the estimates that show the correlations between the respondents answers on the 26 items and the respondents’ latent/unobserved motivations, either large (in absolute value close to one) or small (close to zero) and is this way assist in the interpretation of the motivational structure. Intuition: Varimax attempts to make the factor loadings, i.e., the estimates that show the correlations between the respondents answers on the 26 items and the respondents’ latent/unobserved motivations, either large (in absolute value close to one) or small (close to zero) and is this way assist in the interpretation of the motivational structure.
64
Integrating latent motivational factors in the valuation function (see Nunes and Schokkaert, 2003, JEEM)
Estimate individual warm glow motivation scores and plug it into the valuation function
ij r = β
j kj ki k
hi h
BAU Latent structure