Friday, July 10, 2009

Libertarian Paternalism: Appearance vs. Reality

In the comments section of Russell Roberts’s link to my previous post, “Charlie” makes some thoughtful points. I was going to respond in the comments there, but then decided a new post would be worthwhile. Charlie begins:
Libertarian paternalism seems to aim to take paternalism and give it choice.
The key word here is “seems.” The advocates of libertarian paternalism have taken great pains to present their position as one that does not foreclose choice, and indeed even adds choice. But this is entirely a matter of presentation. They always begin with non-coercive and privately adopted measures, such as the ski-slope markings in Thaler’s NY Times article. And when challenged, they resolutely stick to these innocuous examples (see this debate between Thaler and Mario Rizzo, for example). But if you read Sunstein & Thaler’s actual publications carefully, you will find that they go far beyond non-coercive and private measures. They consciously construct a spectrum of “libertarian paternalist” policies, and at one end of this spectrum lies an absolutely ban on certain activities, such as motorcycling without a helmet. I’m not making this up! See these two previous posts for details, with direct quotes and page citations. (In a forthcoming article, Mario and I examine the “libertarian paternalist” spectrum in greater detail.)

Charlie continues:
For instance, the 70s Democrat solution to this problem would be to mandate that only one type of "plain vanilla" loan can be given out by all loan offerers. Now it's just that the plain vanilla loan must be offered as a choice.
This is actually an excellent example of Sunstein & Thaler’s rhetorical approach. Yes, it’s true, this policy offers the consumer greater choice than a more restrictive policy. BUT! Notice, first, that the more restrictive policy is no longer in place, which means the actual change advocated here would constitute a greater restriction than the status quo. And notice, second, that the policy in question most certainly does restrict the freedom of the other party – the bank – by requiring it to offer a certain kind of mortgage. Now, there might be great arguments for this approach; as I’ve said, I’m not a finance guy. But can we please not pretend this is a policy that fully respects freedom of choice?

Furthermore, as Sunstein & Thaler’s published work clearly indicates, this kind of policy is the thin end of the wedge. The next step, as outlined in their articles, is to raise the cost of choosing other options. In this case, the government could impose more and more onerous requirements for opting out of the “plain vanilla” mortgage: you must fill out extra paperwork, you must get an outside accountant, you must have a lawyer present, you must endure a waiting period, etc., etc. Again, this is not my paranoid imagination at work. S&T have said explicitly that restrictions like these would count as “libertarian paternalism” by their definition.

Charlie continues:
I find it hard to fault Thaler for not using more gov't anecdotes, because not that many exist. Taxes, tax credits, and subsidies have long been recognized as better ways to achieve public goals than top down regulation. Such as carbon tax > technology regulation, tax free IRA > mandating people save, school voucher > direct school spending. All these are libertarian paternalistic policies. They aim to achieve public goals without destroying choice.
Most libertarians, myself included, are quite willing to say that some interventions are worse than others – and when our favored policies are off the table, we will choose the second best. For instance, most libertarians I know prefer mandatory savings to government-run Social Security. Most libertarians I know prefer education vouchers or tax credits to the public school monopoly. Most libertarians I know prefer medical-marijuana-by-prescription to a total ban on marijuana.

The problem is that S&T’s “libertarian paternalism” is used almost exclusively to advocate greater intervention, not less. I have never, for instance, seen S&T push for privatization of Social Security or vouchers in education. I have never seen them advocate repealing a blanket smoking ban and replacing it with a special licensing system for restaurants that want to allow their customers to smoke. If they have, I would love to see it.

In their articles, S&T pay lip service to the idea that libertarian paternalism lies between hard paternalism and laissez faire, and thus that it could in principle be used to expand choice. But look at the actual list of policies they’ve advocated on libertarian paternalist grounds, and see where their real priorities lie.


Charlie said...

I appreceiate you considering my comments. I have seen the Thaler v. Rizzo debate before, but I reread it since you linked to it. I thought these two points summarized the debate.

Thaler: "Instead we stress that when governments do write laws, especially those mandating -- rather than nudging -- some action, they should do so with an eye toward making people better off. We do admit to liking some mandated -- and thus non-libertarian -- cooling-off periods under certain circumstances, as when buyers are especially likely to have made decisions under undue selling pressure. Who amongst us has not bought something under pressure that he would like to undo the next morning?

Mario's main misconception is that government can avoid nudging. It can't. The rules of the common law are legal rules that governments write. Whether governments are more or less corrupt than the private sector is an empirical question, and there are surely many examples of dumb or unethical behavior in both sectors. But this is beside the point. We favor better government, not more government. We urge both sectors to adopt libertarian paternalistic policies."

Rizzo: "Richard wants to use the word "libertarian" to differentiate his paternalism from the traditional variants. Yet he uses the word in a fuzzy way. He wants to define libertarian along a continuous variable -- the cost of exercising the exit option. However, libertarianism, as every libertarian understands it, uses a bright-line test -- who imposes the cost?"

To most outside observers, Rizzo's idea that there is a bright line libertarian test seems very silly. Because most people have met libertarians have noticed they span the spectrum from anarcho-capitalists to fiscally minded republicans. Undoubtedly, Rizzo would pick some group and identify them as true libertarians and everyone else as something else, but then he is really upset that the whole world uses libertarian to mean something different than he does, not just Thaler.

In fact, some libertarians think collecting taxes to pay for a military is paternalistically restricting our choices. Others think collecting taxes for public schools or mandating people spend money on education is A-Ok. When the outside world hears "Libertarian Paternalism" they don't think like libertarians that "this is an assault on my culture, my worldview, my intellectual foundation." They think, "oh, you mean paternalistic policies that Milton Freidman might propose like school vouchers, the Earned Income Tax Credit, and pollution taxes."

Thaler argues that their is a continuum between libertarian and paternalism and somewhere in between there is a smaller continuum that is called libertarian paternalism. To prove him wrong, you'll need to come up with a definition for a policy to be called libertarian or paternalistic. To do so, you'll end up with statements like, "Mandating people raise money for courts and national defense is a paternalistic policy, but almost all libertarians agree that it should be done." This is where Rizzo tries to take us. Of course, it's silly cause now everyone is a libertarian paternalist wanting some libertarian polies and some paternalist policies, and we are denying that policies themselves are on a scale, so mandating health food labels is no more paternalistic than the military draft. Sunstein and Thaler say things like "more paternalistic" and "more libertarian," Rizzo wants a bright line, but can't draw one.

If Sunstein and Thaler had called it market-oriented paternalism or Chicago paternalism it wouldn't receive anything like the reaction it has, but libertarians see it as an assault on the word libertarian. I guess I can't blame them, they already lost the word liberal, what will they be called if they lose another word?


Glen Whitman said...


I appreciate your points, but I also disagree. I think you, too, may have been fooled by Thaler's disingenuous way of presenting his ideas. I encourage you to read his actual published articles in order to see what he really advocates, rather than reading only his op-eds, blogs, and debates.

Upon rereading the Thaler-Rizzo debate, what struck me most was Thaler's willingness to misrepresent his own position. When he says, "We have never suggested that any particular contractual form be imposed, including automatic enrollment," he's lying. He has undoubtedly suggested a whole range of contractual forms under the rubric of libertarian paternalism; again, read his articles. And when he says, "It is both wrong and misleading to characterize libertarian paternalism as primarily an activity of governments," he is being highly misleading at best. Look at his articles, and you'll see that he and Sunstein start with one or two examples of private paternalism, then rapidly transition to a lengthy list of paternalist government interventions. If libertarian paternalism were primarily about private action, why would they use a politically loaded term like "libertarian" to refer to it?

As for whether paternalism is inevitable, two things. First, even with default rules, paternalist motivation is not inevitable. Default rules in common law have typically been selected by "ratifying" what is conventional practice -- usually what is the most commonly chosen option. No paternalism is required for that approach, but paternalism is required to overturn it. The new default rules would put the onus of opting out on those who would deviate from the paternalist's favored option, rather than on those who deviate from the most typical behavior.

Second, even if you agree that some paternalism is inevitable, Sunstein & Thaler go very far beyond the inevitable. Again, read their actual papers.

With regard to the definition of libertarianism and non-coercion: It's true that libertarians differ amongst themselves on a variety of matters. But on this matter, there is remarkable agreement. Libertarians don't define coercion in terms of cost alone. No libertarian I've met would say you're "coerced" by the costliness of traveling 5 miles to the nearest bakery, unless perhaps that 5 miles resulted from (say) a zoning restriction. Nor are you coerced by the baker's choice to charge high prices. What matters is not the presence of a cost, but how the cost is imposed and by whom. Again, this is not a serious point of debate among libertarians. Sunstein & Thaler are just confusing matters by trying to subtly shift the definition.

As for bright lines: Every line involves some degree of fuzziness. But some lines are clearly fuzzier than others. Libertarianism, with its emphasis on rights, generally leans toward brighter lines. Sunstein & Thaler, on the other hand, studiously avoid drawing any lines at all. They deliberately ignore the line between public and private, and between coercive and non-coercive. Indeed, they do their best to obsure those lines.

Finally, on the issue of libertarian hard-liners versus moderates. As libertarians go, I'm personally fairly moderate, and I think Mario is similar. There are a number of policies that I support, despite their lack of 100% libertarian pedigree, because on net I think they're desirable -- but I recognize when I'm doing so. This debate isn't primarily about how hard-core we want to be. It's about being honest about the lines being crossed by the policies we adopt, and it's about being careful in our arguments. I want S&T to be honest that many of their so-called "libertarian paternalist" proposals are not very libertarian at all, but in fact limit choice in various ways. And I want them to recognize the fuzzy and slippery nature of their argumentation. Then we can have an actual debate about these policies on the merits.

Anonymous said...

I'm not a 'libertarian' of any sort, but expecting modern states to be able to forsee and economically implement any of these measures, whether or not we think they are theoretically good ideas, is absolutely crazy. The nation state has to be the most ass-backwards political organization in history. It's also the most dangerous and hardest to control. Sovereignty is bad, mmkay.

See van Crevald, 'The Rise and Decline of the State'; Jeffrey Friedman in Critical Review 'Bureaucracy'; Carl Schmitt, 'The Concept of the Political' and 'Political Theology'; Anthony de Jasay 'The State'.

One of the great failures of modern politics (and human beings in general) is taking too much for granted, and assuming what 'everybody knows' has any relationship to what is actually the case. As much as libertarians and market-statists like to criticize the 'mainstream' for this failing, they demonstrate it in ample measure by pretending that there is something either 'natural' or exclusive about the alternatives of anarchy and state. Most political organizations in human history, and most social organization at present, can not be reasonably described as either sovereign civil states or anarchistic. The social structures that define most human beings past and present are a heterogeneous network of contractual, semi-contractual and customary norms and institutions.

Clint said...


I'm largely unaware of the internal debates about libertarianism, though I'd consider its objectives to be quite in line with a lot of my political thinking. That said:

"I have never, for instance, seen S&T push for privatization of Social Security or vouchers in education. I have never seen them advocate repealing a blanket smoking ban and replacing it with a special licensing system for restaurants that want to allow their customers to smoke."

I know these are tangential to your main point here, but it seems to me that these cases (SS, vouchers, smoking) could all be justified from a libertarian perspective. While all three deprive people of freedom, I could argue quite well that each enables far greater freedom.

A smoking ban strengthens the freedom to live. A public school system strengthens the intellectual freedom that comes from education. Social Security strengthens retirement freedom.

That said, I don't agree with paternalism in any political context. The decision to institute a smoking ban must be arrived at democratically.

Ran said...

You might enjoy what I take to be a video-essay on the perverse effects of paternalism:

Jon said...

A smoking ban strengthens the freedom to live. A public school system strengthens the intellectual freedom that comes from education. Social Security strengthens retirement freedom.

No. A smoking ban decreases freedom by taking away choice. That one might conceivably live longer -- if the ban works to actually prevent a person from smoking, thus increasing her life -- under such a ban, thus increasing the sheer number of things that one could potentially get to over her life, doesn't alter the fact that she's less free because of it. One of those things she'd like to do has been prohibited to her.

Put differently, what we do with our freedom isn't the same thing as freedom itself.