For the most part "welfare theory' addresses problems of social choice in situations where future population sizes are assumed to be given. So does applied welfare theory, as in models of optimal income and commodity taxation or investment project appraisal. Yet discussions of population policy often provoke fierce controversy. In a recent Discussion Paper, Programme Director Partha Dasgupta explores the normative or ethical foundations of such population policies. He considers the problems faced by economic theory in addressing issues of social choice if future population size is controllable via alterations in birth rates.

Social choice theory in general, and welfare economics in particular, has made use of the assumption of "symmetry' (otherwise called the "anonymity' axiom). This asserts that all permutations of individual levels of well-being ought to be regarded by society as equally desirable states of affairs: in other words, a person's name, or identity, ought not to matter in social choice. Dasgupta begins by exploring various philosophical arguments that can be invoked to justify the symmetry axiom, and he argues that the axiom is a compelling one. The consonance between anonymity and positive social discount rates is also discussed.

Dasgupta argues that optimum population theory has almost invariably invoked an extended version of the symmetry axiom, as in Henry Sidgwick's Classical Utilitarianism. Dasgupta suggests that this is appropriate only in what may be called the Genesis Problem, where all persons are "potential'. He demonstrates that Classical Utilitarianism in many plausible circumstances may advocate "too many people', a feature that Derek Parfit has recently dubbed the Repugnant Conclusion. It is shown that a Repugnant Conclusion is implied not so much by Classical Utilitarianism, but rather by an extension of the Pareto- principle. Therefore if the Repugnant Conclusion is to be avoided, this extension of the Pareto-principle must be rejected, which in turn is possible only by invoking explicit distributional considerations. Even at the axiomatic level, distributional issues are intrinsically related to aggregate issues , such as the Repugnant Conclusion.

Finally, Dasgupta argues that the Genesis Problem is misleading. In an "Actual Problem' there are actual people who must decide on the size of future generations, the rate of investment and the rate of resource depletion. Using a simple example, he shows that the symmetry axiom is questionable in an Actual Problem, where it is extended to the well-being of "potential persons', whose very existence is itself the subject of choice. New problems arise, however, if the symmetry axiom is rejected - in particular, a social welfare function simply does not exist without it. It is argued that social choice can be made despite this, and Dasgupta concludes by outlining how such choice can consistently be made.


The Ethical Foundations of Population Policies
Partha Dasgupta

Discussion Paper No. 35, November 1984 (ATE)