OK I am supposed to be balancing my checkbook tonight, so that’s why I continue to harp on this issue of Dan Klein and interpersonal utility comparisons. Gene Callahan and Daniel Kuehn have chimed in. I have just two more clarifications, and then it’s double entry bookkeeping time, baby. Goethe, eat your heart out.
(1) I have no problem if Dan Klein thinks “a dollar means more to a poor person than to a rich person.” I have no problem if he goes so far as to say, “Look, we’ve all heard the standard shtick from Rothbard to Varian about ordinal preference relations blah blah blah. But c’mon people, we all know perfectly well that there is a legitimate sense in which a dollar means more to a poor person than to a rich person.”
To repeat, if Klein says that, fair enough. I have little problem with it. What I do have a problem with, is Klein putting that on a test to see if people know basic economics. That would be like asking, “Can the government raise GDP by spending a dollar?” and then dinging people for saying yes. That would be a very unfair question to ask people, since they were probably taught that government spending can raise GDP. And by the same token, some of the people Klein surveyed may very well have been explicitly taught that it is a meaningless statement to claim that a dollar means more to a poor person than to a rich person.
So that was my objection, that this was an awful question for the purpose Klein intended. I don’t care how Klein himself would answer the question.
(2) Daniel thinks we can all get along if we assume “other things equal.” No, that won’t cut it, at least not in the way I am thinking about utility. It’s not that I think “a dollar means more to a poor person than a rich person” is sometimes true, sometimes false, and thus I don’t want to say it is true in general. For example, my objection isn’t that hey, maybe we’re comparing a miserly billionaire with an aesthetic monk who took a vow of poverty.
No, that’s not the problem. The problem is that utility has to do with ordinal preferences. To say “an apple gives Mary more utility than a banana, but vice versa for John,” simply means that Mary would choose the apple over the banana, while John would do the opposite if faced with the same choice. There is no way you can take such language and come up with a comparison of Mary’s utility from an apple versus John’s utility from a banana. Mary and John are separate, acting agents. It is simply nonsense to start comparing one person’s utility with another, if you are using “utility” the way mathematical economists have used it in price theory since (at the latest) the 1930s, and the way the Austrians were using it even earlier.