Here is yours truly again, from his latest book. I tried to condense the limits of welfare economics into a few simple sentences; here is what I came up with:
On the negative side, the economic approach considers only a limited range of values, namely those embodied in individual preferences and expressed in terms of willingness to pay. This postulate is self-evident to many economists, but it fails to command wider assent. It wishes to erect “satisfying a preference” as an independent ethical value, but is unwilling to consider any possible competing values, apart from preferences. It is hard to see why non-preference values should not be admitted to a broader decision calculus.
Typically economists retreat to their intuition that satisfying preferences is somehow "real," and that pursuing non-preference values is religious, mystical, or paternalistic. The rest of the world, however, has not found this distinction persuasive. They do not see why satisfying preferences should be a value of special and sole importance, especially when those same preferences may be ill-informed, inconsistent, malicious, or spiteful. The decisions to count all preferences, to use money as the measuring rod, and to weight all market demands equally must themselves rely on external ethical judgments. For that reason, the economist has no a priori means of dismissing non-preference values from the overall policy evaluation.