Envy and efficiency

Joseph Heath1 The Pareto principle states that if a proposed change in the condition of society makes at least one person better off, and does not make anyone else worse off, then that change should be regarded as an improvement. This principle forms the conceptual core of modern welfare economics, and exercises enormous influence in contemporary discussions of justice and equality. It does, however, have an Achilles’ heel. When an individual experiences envy, it means that improvements in the condition of others may worsen the condition of that individual. As a result, envy has the potential to block a vast range of changes that we might intuitively be inclined to regard as Pareto improvements. (Or more precisely, envy results in too many states getting classified as Pareto-optimal, not because, intuitively, they cannot be improved upon, but because no one’s condition can be improved upon without making someone else envious.) For example, a market exchange between two people might not wind up being classified as a Pareto improvement if the benefits produced for the two parties generated envy in some otherwise uninvolved third. Some theorists take this as sufficient reason to dismiss the Pareto standard entirely. These views are not my concern here. The more common response, among both economists and philosophers, has been to treat preferences that involve envy as illegitimate. This is not to deny that agents have such preferences. The claim is simply that these preferences should carry no weight, from the normative point of view. The most memorable formulation of this idea was provided by Robert Goodin (1986), with his image of “laundering preferences.” The claim, roughly, is that a theory of justice will take preferences as “input,” then apply some set of principles, or a social welfare function, in order to generate a normative ranking of possible states as “output.” Such a theory may contain both an input and an output filter. The former will filter out illegitimate (e.g.