Libertarian Paternalism???!! What the...

A very astute-seeming reader of this blog asked me, as a behavioral economist, about my take on libertarian paternalism. Now, I taught an economics tutorial entitled "Sex, Drugs and Rock & Roll", where we discussed risky behaviors and various takes on the libertarian viewpoint. Nonetheless, this concept is (was) foreign to me, and in literal terms doesn't even seem to make sense. The nice people who contribute to Wikipedia, however, inform me that "soft paternalism, also referred to as libertarian paternalism, is a political philosophy that believes the state can 'help you make the choices you would make for yourself—if only you had the strength of will and the sharpness of mind'". Okay, so far so good, though I'm not convinced that the state knows how a perfectly rational me would behave.

Now, for my take on the matter. Let's take a person that smokes crack as an example. This person may be
a. Behaving perfectly rationally, but his discounting of future happiness is so large that he is willing to sacrifice a lot of long-term happiness for the immediate high
b. Behaving in a time-inconsistent (or state-inconsistent) way such that after he smokes some crack he's going to look back and really wish that he had made a different choice
c. Be unaware of the true costs and benefits of smoking crack (note that this can apply to both rational and time-inconsistent people)
d. Not considering costs and benefits at all, since some hard drugs can actually suppress the areas of the brain that analyze those tradeoffs

Now, consider that the aim of policy in general is to maximize social welfare. I would then argue that situations b and d could warrant some paternalistic intervention, since those crack smokers will almost certainly be thankful later. Choice c seems to call for some sort of education initiative. And choice a...well, that guy is just somebody that I can't relate to, but he seems to be doing okay for himself. A policy gets very tricky here because an outside observer, and perhaps the people themselves, cannot distinguish between possibilities a-d, so there would have to be a blanket rule that raises some people's long-run utility (b,d) through limiting choices and lowers that of others (a and potentially c). Do the benefits to the winners outweigh the costs to the losers? I at least think that this is the relevant question to consider.

I found the paternalism argument curious in general, since it is often a hard sell, and much of the justification for intervention can be made through an externalities argument as well. For example, I would be much more behind the government outlawing crack because of what it can do to children, both before and after birth, since children cannot choose who their parents are and presumably are not in the market for crack themselves.

Addendum: I added a comment to the Marginal Revolution post that initially started this discussion. Basically, one of the arguments being made was regarding paternalism for poor people since they are more prone to "irrational" choices. A colleague of mine did experiments on this very topic (paper source is given in the comment), and he finds that it is cognitve ability that drives biases, not specifically level of wealth. So even if somebody was into peternalism for those that can't help themselves, targeting poor people isn't the way to go. Also, when behavioral economists hear about increased choice being utility-decreasing, they usually think of bounded rationality issues rather than self-control ones.

Addendum #2: Apparently the Australian government doesn't have qualms about being selectively paternalistic.