Continuing with my ongoing theme of puzzling over regret aversion by proxy, I'd like to try to compare counterfactual states of the world where:
a) regret aversion can either be generalized or not
b) paternalism acts as either a substitute or a complement for individual regret aversion
c) the world is either well-ordered or non-ergodic (by non-ergodic, I mean that surprising, unpredictable [as distinct from unpredicted] events happen)
What does it mean if a lesson can be generalized? This is one of the dilemmas of education: as educators, we want our students to be able to draw larger conclusions from our examples, to be able to apply the basic framework to a wide range of problems. This is the point of Sir Issac Newton getting bonked on the head by an apple while reading in the quad--he didn't develop a theory of the mutual attraction of apples and noggins, he described the Law of Universal Gravitation. Unfortunately, the literature suggests that most students aren't of Newton's caliber and they seem to have a hard time generalizing from word problems about trains leaving from St. Louis. Choo choo. That's for classroom instruction, however. Perhaps it's possible that by helping people make better decisions through force in the market for sex and drugs, they'll make the right decisions in the rock-and-roll market all on their own. I'm not sure what the literature says on this, so I'm obliged to consider that it's at least possible.
Does paternalism erode virtues like diligence, forbearance and prudence? Moral hazard arguments rest their analytical power on the answer being "yes". I can't think of any one of my friends, on the left or the right who would argue otherwise. The disagreement is rather over whether or not the marginal benefit of the particular intervention outweighs the marginal cost. For big stuff like giant bank bailouts, conservative economists will shout that the state has rewarded bad decisions made by politically well-connected elite entities. My libertarian economist friends will point out that these big bailouts will provide more implicit government guarantees for the next round of imprudent overexposure. My liberal pals snicker at this overwrought hand-wringing and point out that, hey, would you rather the entire banking system and global economy collapse? From their point of view, the conservative perspective is cruel and the libertarians (beyond being irrelevant) offer a red herring, since if what you want is to eliminate moral hazard, the system just needs more oversight and regulation. So for individuals, does paternalistic regulation help people who just need a leg up on making the right decisions, or does it act as a barrier for people who really do honestly and truly want to engage in transactions that offend the moral sensibilities of the paternalist? Can it be an admixture of both?
Consider alcohol prohibition. Without a doubt, there's an element of parentalism in there: some alcoholics have a hard time weaning themselves off the bottle without some external help. Other folks have no trouble going their whole lives drinking lightly in social settings only. Even ignoring the terrible social dysfunction associated with prohibition, the 18th Amendment cared for the first group at the expense of the second. Moreover, over a person's lifetime, they de facto abdicate the onus of having to carefully consider the costs and benefits of alcohol (or marijuana or prostitution or cigarettes or what-have-you) consumption. Instead of asking "do the discounted net lifetime benefits exceed my required threshold", they ask "is this illegal?"
Naturally, this isn't the entire story. There's this phenomenon sometimes called "time-inconsistent preference reversal" or something similar. Once you clear away the math, the idea is simple, clear and consistent with intuition. At one particular time in one particular context, a person may not be especially considerate of the future. Give that same person some time or change the context and hey-presto, the future state of affairs may all of a sudden be especially relevant. Think of having kids: one day you're the happy-go-lucky grasshopper, playing your fiddle, not giving much of a care for the coming winter, but when baby makes three, there's a Kafkaesque metamorphosis into a diligent ant. This effect is probably more common than we think. People's time preferences change with age, with lifecycle changes (distinct from strictly age, of course), with psychological health, with career health, with... well, with lots of stuff. People are dynamic, including their expectations for the future. This is hardly revelatory, but it does pose a bit of a problem for paternalism: how to distinguish between the grasshopper and the ant when they wear the same skin and hum the same tune.
To get the most bang for your buck (so to speak), you'd want people to be able to apply lessons learned from a minimally invasive paternalistic policy to other areas where they might otherwise make potentially regrettable decisions (this helps avoid the problem of throwing the baby out with the bath water). You'd also want either a well-ordered world ruled by Gaussian distributions and normally distributed error terms or people to retain (or develop) healthy, functional regret aversion mechanisms. The breadth, scope and growth of the regulatory state suggests that these conditions probably don't hold.
I think it's probably fair to say that we can safely discard the assumption about ergodicity. The systematic failure of soothsayers and forecasters is strong evidence in favor of non-ergodicity. If regret aversion could be generalized, then the only justifications for a large paternalism apparatus would be sinister: wealth appropriation or tyrannical social engineering: regulations are no longer in the interests of the citizen, but in the interests of the ruling elites. While I think this may happen in some dictatorships and totalitarian states, I am not prepared to claim it's a big part of the political motivation in social democracies. As far as complementarity is concerned, my prior belief is that moral hazard arguments are probably correct. Furthermore, if the world is non-ergodic and paternalistic regulation is a substitute for regret aversion and lessons are non-transferable, regulation is logically unable to accomplish its goal of protecting people from their own lousy decisions even within its own scope. Bans on drugs have not stopped people from getting high. Bans on toys sold with meals haven't stopped kids from eating hamburgers. There is relentless pressure to skirt the best laid plans of mice and men. Human ingenuity and non-ergodicity is why they oft go awry.
So where does this leave us in terms of moral obligation? If you'll indulge my unfortunate habit of beating the same drum yet again, I'd like folks to reconsider what the optimal regulatory zone is. I doubt it's the entire US for most stuff. I'd forward the proposition that the ideal regulatory zone for pharmaceutical products is probably no larger than the family and social circles. Even if consuming Schedule I drugs are non-euvoluntary for some folks, it strikes me as unethical to punish non-regretful users to prevent potential regret in others. Ditto for bacon-wrapped hot dogs and helmet-free bicycle riding and walking back and forth from the rec center to the athletic field or, heck, pick your poison from any of the nanny-state-of-the-month award winners. Moreover, to the extent that people aren't really all that indifferent between default options, the new paternalists will have to move from nudges to shoves to the inevitable truncheon of the state when voluntary choices fail to conform to the wishes of the nudger, which are, naturally, the simulated future agent of the patronized.
And perhaps more importantly, how willing are we to force others to conform to our own deontological preferences?