George Will's recent article regarding libertarian paternalism has attracted some attention in the "libertarian blogosphere," for example, Jim Henley's post over at The Art of the Possible.

A few months ago, I wrote a brief summary of Libertarian Paternalism, identifying Thaler and Sunstein's "Choice Architecture" conceptual framework as 4th Generation Chicago School.The original motivation for my post, of course, was to dispel the myth that Obama was a welfare redistributionist. In fact, theoretically, "Libertarian Paternalism" is a departure from Paretoian Welfare Economics, as Sustain and Thaler argued in this paper. If you don't want to read the paper, the synopsis argument is that economic agents are only "boundedly rational," and that a menu-biased social choice theory (architected to "nudge" you to make the most rational choice) is more "libertarian" than a system that uses a litany of pigovian taxes or outright banning to discourage the "bad choices." Sustain and Thaler justify the "soft paternalism" of the "biased-menu" by arguing that menu-neutrality is an abstract myth, often resorting to using the "Cafeteria" as an analogy. Actually, I think the modern grocery food chain is a much better example, in that it literally spends millions collectively in studying how to "choice architect" the layout of a typical store to maximize sales.

Theoretically, "Libertarian Paternalism" derives in part from the implications of Sen's Paradox, which implies that even minimal libertarianism is generally incompatible with a weak form of Pareto Optimality.

Let's illustrate Sen's Paradox with an example. Consider a population of 2, P(1), P(2), constrained to a 4-tuple of choices {w,x,y,z}.
P(1)'s preferential ordering of the 4-tuple is: w,x,y,z.
P(2)'s preferential ordering of the 4-tuple is: y,z,w,x.

For a minimal condition of libertarianism to hold, both P(1) and P(2) must claim self-ownership or dictatorial powers over at least one pair of alternatives.

Therefore, we impose the following additional constraints:

P(1) has dictatorial powers of x over y.
P(2) has dictatorial powers of z over w.

The "libertarian constraints," however, only leaves us with x and z as the two remaining alternatives, but both P(1) and P(2) would prefer w over x and, likewise, y over z...in other words, neither x or z are Pareto Optimal.

Let's make this example a bit more concrete. Let P(1) = "a somewhat opinionated conservative boyfriend" and P(2)="his somewhat opinionated progressive girlfriend."

They have resolved to spend a Sunday afternoon together watching TV. Our 4-tuple of choices are:

w=watch conservative political show
x=watch football
y=watch liberal political show
z =watch romantic movie

The couple, however, are only now left with a choice between a football game and a romantic movie. It's a week before the greatest ideological election in the history of mankind between Brack Obama and John McCain, but the football game is between the the Jets and Falcons and the romantic movie is the 100th rerun of "You've Got Mail." How does our intrepid couple resolve this dilemma?

Of course, this is a overtly simplified model, but the general principle more or less remains valid. Let's see how this would be resolved from both an Anarcho-Capitalist standpoint and a Libertarian Paternalistic Standpoint.

First from the Anarcho-Capitalist standpoint, we have to recognize there are 2 strands of thought here. There is David Friedman's strand geared toward a pure economic analytical resolution and the Austrian strand, which is geared toward uncompromising natural rights.

In the case of Friedman, he would likely invoke Coase's theorem, and with alienable rights transfer and frictionless transaction costs, this could be efficiently settled by private bargaining by, say, the girlfriend promising sex that night as well as cooking dinner in exchange for being able to watch her preferred TV show(any feminists reading this blog, this is, in fact, a very real bargaining reality that I have traded in...sorry).

In the case of the Austrians, they don't much believe in "Utility Functions" to begin with and would likely prefer to engage in a diversionary argument of cardinal utility vs ordinal utility, but in the end, would default to the boyfriend and girlfriend calling up their respective DRO's to arbitrate the dispute.

The Libertarian Paternalists, who believe in the State and a menu-biased architecture, would have to take sides in favoring either a feminist or male bias in terms of remote control operation.

You can spend the rest of your life going down the rabbit hole of "libertarian consistency" and it's applicability to the real world, and in many respects, it reminds me of the same rabbit hole you can choose to descend into into terms of physics and the question of quantum de-coherence vs. Copenhagen wave-collapse.