Rizzo and Whitman distinguish between what they call “old” and “new” variants of paternalism. Old paternalists justified interference with people’s choices on the grounds that doing so would bring those choices into line with what the paternalist thinks is best. New paternalists, such as Thaler and Sunstein, justify interfering with people’s choices on the grounds that doing so will bring those choices into line with what their own beliefs about what is best.
Rizzo and Whitman argue that the new paternalist program requires vastly more information than its advocates realize. They think that this information cannot be accumulated and that as a consequence the new paternalism will be indistinguishable from the old paternalism in practice. That is, those engaging in paternalist interference will do so based on their own preferences rather than the preferences of the people who are being interfered with.
Their article is daunting in length, but they gave us a helpful guide at the beginning that identifies six basic areas (Rizzo and Whitman 2009, 910).
Prof. Brown started us off with the theory of revealed preferences. This involves inferring utility functions from observed choices. Research in behavioral economics, alas, suggests that the assumptions needed to make this work are false. Bummer.
So where do we go from there if we want to know what people really want? That is information that we need to have in making all sorts of decisions that affect others, after all. Robert suggested that the critical question concerns regret, whether people wish in retrospect that they had made a different choice than the one that they, in fact, made.
Rizzo, Mario J., and Douglas Glen Whitman. 2009. “The Knowledge Problem of New Paternalism.” BYU Law Review 2009 (4): 103–61. http://digitalcommons.law.byu.edu/lawreview/vol2009/iss4/4.