Applying Rawlsian Principles of Justice to Decide how to Rank Disaggregated General Measures of Social Welfare and Poverty

There is a standard set of information problems associated with comparing levels of social welfare and levels of poverty—even when it is assumed that households derive their well-being from only a limited range of relevant variables. A number of comparatively general results set out, or are alluded to, in the relevant literature indicate how these information problems may be circumvented with varying degree success. To motivate the present discussion extended forms of some, but by no means all, of these results are derived. These derivations raise two central issues. First, some strong assumptions need to be applied in the process of manipulating household information. This matter is considered comparatively briefly. The more important issue, however, is that of determining which of these results employs the most socially acceptable system of interpersonal comparisons. In investigating this issue it is argued that (an interpretation of) the central components of the Rawlsian theory of distributive justice—in particular Rawls’s first principle of justice in distribution and the priority rule—should be applied in order to decide this matter. Relevant empirical evidence is employed in support of this line of argument. It turns out that only one of the previously-derived results provides a satisfactory criterion for the social ranking of levels of poverty. (It also turns out that, given reasonable assumptions, the rank order of levels of social welfare must be the same as that for the ranking of levels of poverty.) The discussion is then extended to allow for practical considerations associated with assessing poverty-alleviation policies. This discussion draws on Rawls’s second principle of justice and the requirement that individuals are treated as being anonymous.