Interpersonal Comparisons of Utility
Interpersonal comparisons of utility (that is, of preferred positions on an individual’s preference scale) are known to be scientifically impossible in economics. LSE economist Lionel Robbins, who published the second edition of his famous Essay on the Nature and Significance of Economic Science in 1935, argued that interpersonal comparisons of utility could not be used to justify either redistribution of income or even free trade—both policies that he favored, very strongly in the last case. (See my anniversary review in the Winter issue of Regulation.)
Modern economic theory does not even need to assume diminishing marginal utility of income or consumption, only a diminishing rate of substitution between goods. Constitutional political economy à la Buchanan cleverly avoids the impossible mission of comparing individual utilities, as I point out in my review of James Buchanan and Gordon Tullock’s classic book The Calculus of Consent:
By focusing on each individual’s rational decision to adhere to a social contract that serves his interests, Buchanan and Tullock avoid the problem known as the interpersonal comparisons of utility. Such comparisons are scientifically impossible because each individual’s utility is subjective, in his own head. But, of course, an individual can decide whether he prefers one situation to another for himself.
There is no way to demonstrate that the poor get more utility from a given redistribution than the rich lose in financing the transfer. Does a rich woman from whom $50 is taken and transferred to a poor man lose less utility than the latter gains? Suppose the rich woman would have bought one more bottle of Bourgogne and the poor man buys three cases of beer with the transfer. Does your (or another external observer’s) opinion change if we suppose that the poor man buys three jugs of milk instead?
An external observer must evaluate what is better, according to his own preferences or perhaps values (“in an ideal society, the poor should drink more milk”). Interpersonal comparisons of utility are moral opinions (what society should be) or, if it is one of the parties involved who makes the comparison, self-interested claims (“I want more beer”). Hence the impossibility of scientific interpersonal comparisons of utility. There is no way to tell whether the transfer has increased or decreased “aggregate utility,” an expression that has no ascertainable meaning in economics.
Add to this that if money is redistributed from a few rich to many poor, the gain by any of the latter must be so much smaller, and the argument for redistribution is even more debatable.
There are extreme cases where nearly everybody would agree that little utility is lost on one side and much gained on the other: consider the transfer of $10 from Bill Gates to a homeless man in a tent near a Macdonald’s. But this still remains merely an intuition because the ordinal ranking of some good (more or less preferred to other goods) is not comparable across the preferences of two separate individuals. Anyway, moral intuitions about such extreme cases don’t seem to provide any general justification for coercion.
Some argue that parents implicitly make utility comparisons among their children. But is this really what parents do or should do? Don’t they instead compare different future opportunities for each child while assuming the equality of their children? Children, by definition, cannot make such evaluations, and their parents are the best placed to do it. Even then, some external observers may disagree with some parents’ decisions. The two parents may disagree among themselves. At any rate, the government is not our parent.
One might think he can read in a friend’s mind, but somebody else, including the friend himself or herself, may think differently.
When they are not constrained, politicians and other rulers may, implicitly or explicitly, make interpersonal utility comparisons among their subjects. This does not prove that their comparisons are not arbitrary or self-interested. As Anthony de Jasay wrote, “when the state cannot please everybody, it will choose whom it had better please.” Indeed, the redistribution often goes from the poor to the rich. Do interpersonal utility comparisons only break down when the results go the “wrong” way?
Some economists, including Nobel prizewinner Amartya Sen, have proposed counter-arguments against the rejection of interpersonal utility comparisons. One is that external observers or public discussion can form objective opinions on interpersonal utility. Another one is that utility as judged by the individual himself—that is, what he prefers—is not really what he prefers. Still another counter-argument is that there is more in life than what an individual likes in his own life. All these counter-arguments seem to crucially depend on assuming that external observers know, in a paternalist or elitist way, what other individuals prefer or should prefer. (For a summary of the counter-arguments, see Amartya Sen, Collective Choice and Social Welfare, Expanded edition [Harvard University Press: 2007], pp. 1-41.)
Influenced by Buchanan, I would argue that, if we hold the equality of all individuals to be a crucial normative value, public policy and the very existence of government cannot be justified by impossible interpersonal comparisons of utility. The justification must come from the enforcement by a limited government of the general rules to which individuals can be presumed to unanimously consent at some abstract social-contract or constitutional stage (see James Buchanan, Why I, Too, Am Not a Conservative [Edward Elgar, 2005]; and his seminal The Limits of Liberty  [Liberty Fund, 2000]). Perhaps we can fit into this approach the economic justification of government as a producer of strict “public goods” or the Hayekian theory of the rule of law as coordinating conventions. Otherwise, only anarchy seems to be morally justifiable.