Via Volokh Conspiracy, I see that Richard Posner and Gary Becker have a pair of posts on the idea of “libertarian paternalism” (a notion introduced a couple years back in a paper by Cass Sunstein and Richard Thaler). The idea behind this “libertarian” or “soft paternalism” is that government rules seek to shape or influence people’s choices for their own benefit without actually prohibiting disfavored choices. Examples might include creating a voluntary, temporary “casino blacklist” for people with gambling problems who wish to prevent themselves from succumbing to temptation, or maybe requiring candy and junk food to be placed away from supermarket checkout lines to discourage impulse buys. I always found the concept intriguing, in part because I’ve no real quarrel with helping people make better decisions (by their own standards) without actually restricting their options, but also in part because I’d much rather see whatever paternalistic impulses inevitably exist channeled into relatively innocuous measures of this sort than into bans and mandates.
Posner’s discussion, like a number of previous ones, seems to get unhelpfully hung up on the semantic question of whether this is properly called “paternalism” at all. Posner only wants to use the term to describe cases where “government wishes override the informed preferences of competent adults,” which seems a little odd insofar as that would leave it inapplicable to actual paters. I’d actually take issue with that definition on two points: both the “override” and the “informed preferences” components.
As to the first, there are all sorts of cases in which governments, groups, or individuals attempt to constrain another person’s choices on the premise that, left to their own devices, that person will choose poorly. For lack of another term, “paternalism” seems like a good way to describe that impulse, regardless of whether the means by which it’s exercised actually amount to coercive prohibition. Sure, we most often think of “paternalism” as the coercive sort, which is presumably why Posner thinks “libertarian paternalism” is an oxymoron. But then, if it weren’t distinct in some way from the more familiar sort, we wouldn’t need a modifier. So hey, whatever, call this “soft” version schmaternalism if you prefer.
The problem with the “considered preference” component is brought out well enough by Posner himself, since he rather counterintuitively avers that his defense of New York’s transfat ban was not, after all, a defense of paternalism. You see, paternalism requires overriding “considered preferences,” and transfats are so harmful that (almost) nobody with sufficient information would choose to consume them over very close substitutes. Perhaps that’s right. If so, then as I suggested before, its seems there’s a relatively easy solution to the putatively insuperable transaction costs: Require restaurants serving food with transfats to print those items in red (or whatever) with a big honking warning at the bottom of the menu describing all their dire health effects. If Posner is right about what people would choose, then it’s not clear why patrons would have to go through all the convoluted calculations he describes involving one’s total annual transfat consumption or other risk factors for disease, because there is just nothing to be said in favor of consuming any amount. If no reasonable person would want to eat a transfat meal, then even someone who consumed few other transfats and had few other risk factors for disease would presumably eschew it too. Restaurants would then simply prefer to ditch the transfats and the effect will be equivalent to a ban. But I hope it’s not churlish to point out that if there is some benighted person who might still want to consume a transfat, then a full ban still overrides this mythical beast’s considered preference, and is to that extent paternalistic on Posner’s own account.
Becker’s points are well taken on their own, but he seems to be taking “libertarian paternalism” to delimit a specific set of cases—cases involving problems of bounded rationality—where paternalism may be employed by any means. I had taken it, at least in its defensible form, to specify a set of means, to wit, measures that seek to guide or shape people’s choices (through warning labels or “default options” or strictly voluntary blacklists) without actually limiting them. Maybe I’m misremembering the gist of the original Sunstein/Thaler paper, but that seems like a very distinct notion.
