Normatively Acceptable Bias in Normative Economics, Prioritarianism and Sufficientarianism
This paper begins by offering a historical examination of the concept of bias in social choice theory, tracing how notions of objectivity and impartiality have informed normative economics and the axiomatic conditions imposed on social welfare functions. While it is evident, for instance, that conditions such as anonymity and unanimity serve requirements of impartiality, our aim is to uncover the more obscure connections between objectivity and impartiality and the broader set of constraints widely used in normative economics. We argue in particular that restrictions on the choice environment—such as using an ordinal framework, reliance on individual preferences rather than substantive dictums, the property of profile independence, the condition of unrestricted domain, and various independence conditions— embody a much more rigid understanding of objectivity and impartiality than is often acknowledged. Based on this discussion, we propose a multifaceted measure of bias in social choice that captures the extent to which a social welfare function meets conventional standards of objectivity and impartiality. We then rely on this multifaceted measure of bias to demonstrate that no definitive social welfare function can simultaneously satisfy all conventional safeguards against bias. This result marks an analytical milestone, justifying the introduction of “normatively acceptable bias” into social welfare analysis. In the second part of the paper, we assess whether two widely used normative theories, prioritarianism and sufficientarianism, adequately embody normatively acceptable bias. We argue that the lack of sufficient theoretical grounding in prioritarian frameworks based on extended sympathy rankings or cardinal interpersonal preferences—notably in constructing an unbiased extended sympathy ranking or in determining the appropriate weight to be assigned to the worst off—often leads prioritarian rules to yield highly counterintuitive conclusions. Sufficientarianism addresses this limitation by allowing a more acceptable form of normative bias, establishing thresholds that accord full priority to individuals whose welfare falls below a specified level. Departing from the conventional view of objectivity in normative economics, it permits differentiated treatment across profiles, thereby distinguishing decisions that address urgent needs from those aimed at promoting mere pleasure. However, a more systematic specification of this threshold is not provided; its existence is justified by appeals to the impartial spectator and compassion. It is maintained that compassion is closely related to the notion of needs, but not entirely reducible to them. Drawing on our historical analysis, we build on this position by arguing that an index of basic capabilities can provide a concrete instantiation of the sufficientarian threshold.
Area:
Keywords: normativity, bias, prioritarianism, sufficientarianism