Estimating Social Welfare Preferences Helen Scarborough Deakin University

Distributional issues n Increasing awareness of need to incorporate distribution in policy analysis n Disparity in distributional effects of environmental policy both environmental quality and financial

Examples of distributional issues n n n Benefits and costs of reduction of greenhouse emissions Disparity between private costs of revegetation and social benefits Location of environmental “bads” between areas of higher and lower income

CM to estimate distributional preferences n n Interested in development of CM as a method of estimating distributional preferences Emphasis on estimating social welfare function and social welfare preferences. q q q Social Welfare Function (SWF) ranks social states Function of utility for individuals (or groups) in society Each SWF represents one person’s view of allocation of utility across individuals, or groups, in society

Choice experiment n n n Welfare maximisation Three hypothetical environmental policies Attributes in terms of utility to different generations q q q Described in dollar terms Respondents reminded that dollars represent total well-being Described in terms of change in utility to person representing particular group

Attributes and levels n Attributes: q q q n Utility change Person Aged 50 Utility change Person Aged 25 Utility change Newborn Levels: Five q -$1, 000, -$500, +$1, 000 +$1500 n (focus groups suggested these amounts considered realistic and large enough to influence choice)

Research instrument n Reference key for choice set

Example of choice set

Social Marginal Rates of Substitution Aged 25/ Aged 50 Newborn/ Aged 25 MNL Model 1. 63 (0. 94, 3. 76) 2. 23 (1. 25, 5. 26) 1. 37 (0. 89, 2. 14) ML Model 1. 75 (0. 59, 4. 71) 2. 34 (1. 00, 5. 83) 1. 36 (0. 77, 2. 67)

Five areas for discussion n n n 1. Strategies for conveying to respondents the distinction between social welfare maximisation and utility maximisation 2. Determining attributes and levels to describe utility changes 3. Exploring sensitivity to choice of numéraire in estimating marginal utility 4. Analysing decision strategies and how these advance social welfare literature 5. Incorporating efficiency cost? 6. Need to split distributional weight?

1. Welfare max not utility max n Can we expect respondents to adopt social welfare max as decision making criteria? n Application of veil of ignorance? n “Cheap talk” strategies?

2. Attributes and levels describing utility changes n n n Index of well-being? Using dollars to describe utility changes? Difficulties with common metric q q Over time Over geographical boundaries

3. Choice of numéraire n Exploring difference between utility changes expressed in terms of dollars or for example, access to environmental quality n Which marginal utility?

4. Analysing welfare max strategies n Linking decision-making strategies to welfare literature q q q n Utilitarian? Rawlsian? Egalitarian? CM provides potential to analyse heuristics.

5. Need to split distributional weights n n Two componentsq Marginal utility of income and q Marginal social welfare Is it possible to estimate each component?

Conclusion n Distribution important aspect of policy analysis Significant unresolved issues. Many further research areas