The news of the death of welfare economics is greatly exaggerated
Résumé
The paper reexamines the controversy about Bergson-Samuelson social welfare functions (BSFs) that took place between welfare economists and social choice theorists as a consequence of Arrow's (1951) impossibility theorem. The 1970's witnessed a new version of the theorem that was meant to establish that BSFs "make interpersonal comparisons of utility or are dictatorial". Against this, Samuelson reasserted the existence of well-behaved "ordinalist" BSFs and generally denied the relevance of Arrovian impossibilities to welfare economics. The paper formalizes and reassesses each camp's arguments. While being also critical of Samuelson's, it eventually endorses his conclusion that welfare economics was left untouched by the controversy. It draws some connections of BSFs with contemporary normative economics.