The American approach to the new welfare economics
What shall be called here “the American approach” is associated with the position of Abram Bergson, from the Massachusetts Institute of Technology (MIT), and Paul Samuelson, from Harvard University, that is, both coming from Cambridge, Massachusetts in the United States.
Bergson formalized the concept of social welfare in 1938 (Burk [Bergson] 1938). He defines it as a function of all the elements relevant for welfare: all products, consumer’s goods, the amount of work of each type, non-labour factors, characteristics of the environment, and so on. Through the application of the Pareto criterion, the function may emphasize the “fundamental value of individual preference.” The social welfare function, as eventually formulated by Samuelson (1947), is defined as a function of the individual utility functions that each individual derives from the social state. The shape of these functions captures some value judgements that are explicitly formulated.How can we legitimately decide which would be the right social welfare function? What does a “social preference” even mean? The question was notably asked by the logician Olaf Helmer to Kenneth Arrow when both were working at the Rand Corporation in 1948. Consistently, this function should rely on the individuals’ views, yet without resorting to interpersonal comparisons of utility. Arrow (1963) provides a first answer in 1950. He shows that, under certain conditions, it is impossible to aggregate the preferences of at least three rational individuals in a single collective preference, which would itself be rational (that is, represented by a complete and transitive relation over social states). These conditions are the following: we must not exclude any combination of individual preferences (no restriction domain); we do not wish to resort to dictatorial decision (nondictatorship); the collective decision should not contradict the unanimous preferences (Pareto principle). Arrow also imposes an independence to non-relevant alternatives condition, which he interpreted as a ban on interpersonal comparisons of utility. This impossibility is at the very least annoying: we cannot derive a collective judgement on the basis of individual preferences unless it is dictatorial. It is hence questionable whether the notion of collective welfare would at all make sense. For this reason, the new welfare economics seemed bound to a failure again. Fortunately, this prediction did not materialize.