pageok
pageok
pageok
[Cass Sunstein, guest-blogging, April 14, 2008 at 12:16pm] Trackbacks
Libertarian Paternalism and the "Yeah, Whatever" Heuristic

First of all, many thanks to Eugene for his kindness and generosity in supporting this discussion. While Richard Thaler and I have been working on the topic of libertarian paternalism for a long time, we continue to think of our project as a work-in-progress, and we are eager for suggestions and criticisms.

The basic idea is that private and public institutions might choose approaches that a) fully maintain freedom of choice (and are in that sense libertarian) but b) gently steer people's decisions in directions that will make their lives go better by their own lights (and are in that sense paternalistic).

In writing Nudge, Thaler and I have been much influenced by libertarian challenges to our earlier efforts. Many of our proposals operate as a kind "one-click paternalism," in the form of approaches that allow people to opt out at minimal burden or cost. Think, for example, of a default option for a cell phone setting or a rental car contract, which you can change with a click or a stroke of a pen.

We are hoping that in many domains, libertarian paternalism might be able to provide a real Third Way, one that goes beyond the mandates and bans often favored by the left and the simple enthusiasm for markets often favored by the right.

In the private sector, a poster child for libertarian paternalism is automatic enrollment in 401(k) plans. The idea is that when people start working, some percentage of their earnings automatically goes to savings -- but they can opt out if they want. Thaler is also a codesigner of the Save More Tomorrow plan, by which workers are asked if they want to enroll in a program that devotes some percentage of their future wage increases to savings. Both automatic enrollment and Save More Tomorrow have produced large changes in savings rates, and most workers are satisfied with the plans.

Of course a major impetus for libertarian paternalism is a series of empirical findings showing, not that people are "irrational," but that they are (in Thaler's term) quasi-rational, departing from rationality in predictable and specified ways. There's a ton of research on this question, but many of the proposals in Nudge rely on the immense power of default rules.

If a private company, or the government, offers something as the default, a lot of people will stick with it, even if a change is ridiculously simple to make. People often use what we call the "yeah, whatever" heuristic.

Why do defaults have such power? One reason involves inertia. People have a lot of things to think about, and especially when the choice is complicated, they're might well decide not to decide or to procrastinate. Another reason involves suggestion. When an option is described as a default, people often think, someone sensible thought that this is the way to go, and I might as well go along with what they thought.

Natural questions are: Who sets the default? Aren't they quasi-rational (or worse)? Why should we trust them? I'll turn to these questions soon. For now, the only point is that if we think creatively about the uses of default rules, we should be able to come up with many new possibilities for private and public institutions, which might be able to influence people's behavior without requiring anyone to do anything.

Terrivus:
Welcome Prof. Sunstein! I've read both your and Prof. Thaler's independent works, so I'm intrigued to see what you've come up with together.

And let me also welcome you to Harvard Law. As a conservative-moderate, I may not always agree with your conclusions, but I'm excited to have you at my alma mater. I think you'll find a wide spectrum of viewpoints there that will enhance your scholarship as much as your teaching will enhance your students' experiences.
4.14.2008 12:36pm
Cato (mail):
Default rules are interesting because they have both personal and societal preferences built in to the assumptions underlying them. When creating a default strategy, which should be chosen if there is a divergence?

Suppose, for example, that default rules for 401k plans set the default vehicle for investment as entities holding US government obligations. The inflow of money to the government would be broadly beneficial because the extra funds would drive down government borrowing costs, yet over long periods of time, equities would benefit investors more. What should the government do? Is that a default that should be set?
4.14.2008 12:42pm
Brian Mac:

"a series of empirical findings showing, not that people are "irrational," but that they are (in Thaler's term) quasi-rational, departing from [narrow norms of] rationality in...[vaguely specified] ways [that can be explained post hoc]"

Fixed that for you ;-)
4.14.2008 12:43pm
ithaqua (mail):
The concept of 'libertarian paternalism' strikes me, offhand, as (1) oxymoronic and (2) creepy. Easy opt-outs are better than the current arrangements, true; but the only genuinely libertarian 'default position', it seems to me, is to do nothing and not intervene at all.

Libertarian philosophy, to put it another way, is not merely the idea that people should have the right to decide their own actions (though that is an important part). Equally essential to libertarianism is the idea that any sort of top-down command economy or command society will make worse decisions, by whatever metric you measure by, than decisions made by the individual people involved. The very 'yeah, whatever' satisficing tendency that you seem to rely on for 'libertarian paternalism' to work means that people will often hand off their own responsibilities to these government 'suggestions', whereas if they were responsible for their own savings, for example, they'd pay closer attention to relevant matters and so make better decisions than some federal bureaucrat in a Washington office.

To put it another way: 'opt-out' government Social Security, for example, even if not imposed by force (as with current Social Security) has all the other problems associated with the government taking responsibility for individual retirement, and so is still a bad idea.
4.14.2008 1:10pm
Justin (mail):
Is this basically the same argument introduced by Samuel Issacharoff, et. al, in their Penn law Review article on Asymmetric Paternalism? If so, is the book supposed to popularize this argument, broaden it, make it more nuanced, etc?

More info here
4.14.2008 1:14pm
GWLaw08 (mail):
I always thought that the default for organ donation should be automatic enrollment--that citizens should have to opt-out if they did not want to participate. I think that's the way most of Europe works and they have an organ dontation rate > 70%. Is this is the sort of tinkering with default rules that Prof. Sunstein is referring to?
4.14.2008 1:15pm
A.C.:
The question for me is how the defaults get set -- how transparent is the process, and how susceptible is it to democratic control? I have my doubts about anything left entirely to technocrats, and about selecting public policy based on research into human behavior.

If a given choice is a good idea, it should be put out there as a topic for the democratic system to work with. If the vote favors a given default/opt-out provision, well and good. People can vote to impose things on themselves. But doing things centrally and relying on people to just go along is not the sort of thing I would expect a libertarian to favor.
4.14.2008 1:24pm
Thoughtful (mail):
I'm unclear as to the level at which the "paternalism" is installed. Take 401K enrollment, for example:

Is the idea that a coporation's financial committee decides it is better for their employees to be enrolled, and therefore makes the default option for their company that payments are deducted unless the employee chooses otherwise. In which case there will be marketplace competition for employees that will be based, in part, on how they favor or disfavor this option.

Or is the idea that the government passes a law forcing all companies with 401K plans to choose the "pay-in" default? In which case the system is "libertarian" in name only...
4.14.2008 1:27pm
PatHMV (mail) (www):
Ease of opt-out is crucial factor. May one do so anonymously, via the internet? Can one simply check a box on the employment form? Must one call a civil servant and hope the poor soul enters it properly into the computer? Will that civil servant be instructed to simply do what's asked, or will they act like an AOL call center employee talking to somebody who just wants to quit their account, without spending 2 hours being lectured to about the many wonderful benefits of the service?

How public will the decision to opt-out of a default rule be? Will it be public record? Will your employer know, or your spouse? Societal pressure to conform is better than government force, but not necessarily by much.

How revocable is the decision to opt-in or opt-out? Can you opt-out of the default 401(k) choice a month later? A year later? I've always found that the first day of employment requires dozens of decisions, the consequences of which are scarcely knowable at the time. If one fails to opt-out when the decision is first made, how difficult is it to change your mind later?

IF the opt-out were a private decision, easy to select, with few time limits on doing so, then I might support such a proposal in appropriate policy areas.
4.14.2008 1:35pm
Sasha Volokh (mail) (www):
I'd be interested in seeing how this plays out in terms of particular examples. So far, though, I think you're using "libertarian" in a different way than libertarians use the term.

For example, what should the default be for enrollment in 401(k) plans? I can imagine various defaults, but as long as the rule is announced to the employee ahead of time, I don't think that default could ever be un-libertarian. Even if 100% of the money is put into the 401(k). Even if opting out is in fact very difficult. From a libertarian point of view, the employee agreed to the contract, so provided no one was forcing him to sign the contract (this is a hard standard to meet according to standard libertarian views), anything in that contract is consistent with the contractual freedom that is a hallmark of libertarianism.

This is probably true even if it's a government employment contract, because from the employee's point of view, the government is just like a private employer. (I think there's room in libertarianism for an "unconstitutional conditions" doctrine, so the government shouldn't be as free to make any sort of contract as a private person is; but whatever the limits of that doctrine, I think it's arguably not implicated in the 401(k) question.)

So to the extent it's employers deciding for themselves what sort of policies to offer their employees (and thus what defaults to have), I think it would be fairly hard to fail the test of "libertarianism." The most naive-faith-in-markets libertarian would endorse all these things as just one example of the many things private employers should be allowed to offer.

On the other hand, once we're talking about the government imposing some condition -- for instance, requiring employers to offer this easy-opt-out 401(k) enrollment? -- then we get into an area of dubious libertarianism, regardless how easy the opt-out is. If the opt-out is easy, we might call it "paternalism that's less oppressive than it might otherwise be," or perhaps "paternalism that's more libertarian than the less-libertarian type of paternalism," but just modifying it with the unqualified adjective "libertarian" seems wrong.
4.14.2008 1:36pm
byomtov (mail):
Is the idea that a coporation's financial committee decides it is better for their employees to be enrolled, and therefore makes the default option for their company that payments are deducted unless the employee chooses otherwise. In which case there will be marketplace competition for employees that will be based, in part, on how they favor or disfavor this option.

That doesn't seem plausible. A prospective employee who is interested in the details of the 401(k) will normally understand that he can opt out if he so chooses. He is unlikely to turn down a job offer just because he has to fill out a simple form to do that.

The fact is that there is usually a default - non-participation - today. You have to fill out some forms and make some decisions to enroll in the 401(k). It's hard for me to understand why setting the default one way or the other is an infringement on liberty.
4.14.2008 1:36pm
Adam J:
ithaqua- neither oxymoronic nor creepy is really substantive criticism. The other problem is that with your critism is this statement "top-down command economy or command society will make worse decisions, by whatever metric you measure by, than decisions made by the individual people involved" can be wrong, because individuals frequently don't have the knowledge necessary to make the correct decision- individuals frequently have more important issues on their mind, like getting food on the table. Those that stay rationally ignorant can rely on the "experts" &stay with the default, or they can opt out. I can't possibly see how this presents a problem to libertarians, since it gives free choice to anyone with the initiative to grab it. Since libertarianism seems to based on the assumption that all people will have this initiative to make the best choices for themselves, I can't possibly see what the problem is.

A.C.- yeah, who would want public policy to reflect human behavior... that would just be nuts.
4.14.2008 1:42pm
Adam J:
critism = criticism
4.14.2008 1:42pm
A.C.:
Public policy designed by a small group to manipulate some facet of human behavior -- that's the problem. It's the lack of transparency that bugs me, not that it involves human behavior in some way. If someone wants to manipulate my incentive structure, I want them to say so up front and give me a chance to weigh in on the scheme.
4.14.2008 1:48pm
ithaqua (mail):
Adam J.,

"individuals frequently don't have the knowledge necessary to make the correct decision- individuals frequently have more important issues on their mind, like getting food on the table."

So your default assumption is that people are stupid and require guidance from liberal 'experts' to make the 'right' decisions. Well, you've got the paternalism part down pat, anyway; forgive me, though, if I have a higher opinion of individual capabilities than you do.

Creepy, I admit, is not particularly substantive. I stand by 'oxymoronic', though. Sasha Volokh put it gently: "I think you're using "libertarian" in a different way than libertarians use the term". I'd be more blunt: if a philosophy is paternalistic - that is, if it presumes to make the 'right' decisions for other people, if its default mode is to do something with someone else's money or property or what have you - then it isn't libertarian, and vice versa.
4.14.2008 1:56pm
Crust (mail):
Makes sense substantively to me. The one thing I don't get though is what is supposed to be paternalistic about "libertarian paternalism".

There is always going to be a default (at least in many of these situations, e.g. 401K selection). The proposal seems to be to give some thought to what the default should be rather than setting it more or less at random. It has been empirically observed that for whatever reasons people have a strong bias to chose the default (makes sense to me). My sense is there is a tendency to think that observation itself is paternalistic or objectionable. But (assuming the empirical foundations are accurate) that's just the way the world/human nature is.
4.14.2008 1:58pm
HipposGoBerserk (mail):
I think the problem people are having with this is that the lable doesn't fit the program being promoted. The proposal is that we should make a concerted effort to set rational defaults in a system so complex that it requires a lot of default rules. Given the (rational) ignorance and (rational) laziness of lots of individuals, it shouldn't surprise us that default rules have significant consequences and ripple through the economy in important ways. The connection to libertarianism is marketing - an attempt to make the paternalism of the proposal less threatening.

That said, I think there shouldn't be a lot of concern about this concept - so long as the opt-outs are simple check the box or state your preference requirements, there is no loss of liberty (no change in the process for making your own choices) and a gain in efficiency from moving to better defaults. The real issue is - can we identify the proper paternalistic default? With 401(k)s, from what's been reported, we have one. But can we generalize?

HGB
4.14.2008 2:00pm
Zacharias (mail):
Opting out is a big problem, in my opinion. If I were a European required to opt out of organ donation, I might have to deface a driver's license to eradicate the entire organ solicitation in order to secure my rights. As it is, I have to punch a 3/16" hole in all Amerikan bills in order to opt out of trusting in GOD.

It would be so fine to have a "libertarian switch" applicable to all queries soliciting an opt-out or an opt-in that would get me out of social security, public education, moments of silence, public prayers, parades, socialized medicine, marriage, breeding, Sundays and holidays and so on. Just taking a different day off than Sunday would expose me to more atheists and libertarians and protect me from having to deal with Texas Baptists and Methodists, for example.
4.14.2008 2:05pm
Scott Wood (mail):
I'd be interested in reading your thoughts about what happens when the new default doesn't change behavior as much as expected.

Take the ice cream vs fruit placement in the company cafeteria example. The libertarian paternalism approach is to put the fruit in the easier to get at spot. What happens if that policy is adopted with the expectation that fruit consumption will double, and it turns out that fruit consumption only increases 25%. Is the attitude going to be "oh well, I guess we learned something about the seriousness of the taste for ice cream" or is it going to be "well, maybe we should try hiding the ice cream behind the refrigerator case"?

My suspicion is that, in practice, most of the libertarian paternalistic policies aren't going to appear nearly as cut and dry as the opt into the 401k plan default, and will lead to all sorts of anti-libertarian actions.
4.14.2008 2:11pm
AK (mail):
I'm not a (big-L) Libertarian, so I'm not that worried about monkeying around with default choices. I would, however, like a promise that opting out of any new default will be available at any time and will be as easy as checking one box. (I've trying to change my direct deposit bank and even something that simple has involved so much work as to make it almost not worthwhile). Can you make me that promise? Any answer other than "yes" is "no."
4.14.2008 2:12pm
Thales (mail) (www):
Cass,

I'm interested in whether you have a public sector "poster child" or two from your book to illustrate how the Nudge thesis might work vis a vis government's relations with individuals, which is what tends to get hard core libertarians testy and defensive. Your opt-out of 401(k) enrollment example makes great intuitive sense. Private law is of course replete with examples of default rules--the Uniform Commercial Code is largely one giant set of default rules. What about public law? Judge Posner has blogged about municipal bans on trans fats, indicating that such fats actually, by any consumer metric (taste, health, consumer price) add no value (they subtract it), and when people are informed about how toxic the fats are, they voluntarily avoid them in favor of other fat substitutes. If these are true claims, such a ban can be described as paternalistic, but in a pretty harmless and choice-enhancing way. It's hard to say your freedom has been materially infringed if you are prevented from ingesting poison that doesn't have any pleasurable side-effects (so this is different from alcohol and tobacco prohibition). I think libertarianish people get more nervous in situations where there is some strong merit for each alternative in a decision situation, and the government nudges them in one direction based on motivations that are presumed to be in the interests of the individual.
4.14.2008 2:14pm
Adam J:
ithaqua - you equate ignorance with stupidity, which is obviously wildly inaccurate. We are ignorant of a great many things in life, because we only have so much brainpower to go around- and we have to decide how to spend it best. Some decisions, particularly ones that impact us financially, are enormously complicated and involve calculations we are incapable of performing, not because we are stupid, but because we're busy working &what free-time we have (if any) is too valuable to spend figuring it all out. Maybe you think everyone should get a finance &a law degree to ensure we can make these decisions, but that really isn't practical.

Also, if you are really so confident of people making the right decision, then you should have absolutely no problem with defaults, because they can opt out if the decision is wrong. Did you read the rest of Sasha's comment- he said Cass isn't using libertarian correctly because there's nothing unlibertarian about a default- which is certainly logical because there's still freedom of choice with a default.

A.C.- fair enough, but good luck with getting any government to do that, liberal or conservative. At least with a default you get the option to analyze it on your own &then make an informed decision if you so choose.
4.14.2008 2:14pm
Oren:
So your default assumption is that people are stupid and require guidance from liberal 'experts' to make the 'right' decisions.
It is quite rational for most people to be ignorant about the arcane details of matters not directly in their expertise.
4.14.2008 2:30pm
A.C.:
So here's the question -- should anyone be allowed to remain ignorant about the need to save for retirement, or should policymakers be forced to pound the drum on this issue until everyone takes in the message? The default here is presented as a way to get the desired behavior without doing the hard work of educating people into it.

This might have side effects. The desired behavior really isn't just that people put money into their 401(k) plans. They also have to leave it in rather than taking it out, they have to decide how to invest the money, and they have to know what to do with it when they retire. The thing you're shooting for may actually be financial literacy, not a one-time decision. And letting the one-time decision stand for the more complicated thing may not do much good over the long term.
4.14.2008 2:41pm
another anonVCfan:
I've ordered the book and I look forward to reading it, but unless you've got a good response to the public-choice critique, I think the idea that benign philospher kings will nudge us in the right direction is naive. Even if we accept that government can establish good default options for individuals--in the sense that the individual ends up healthier/wealthier/wiser than they would without the default--it seems an extraordinary leap of faith that any result of the legislative process will produce good default options.

In the 401(k) scenario, for example, it would probably benefit most employees to be invested in a plan that was widely diversified and held minimal (if any) stock in the corporation itself. But corporations that offer 401(k) plans are a more concentrated interest group than employees who partake in them. The way I see this playing out, corporations will influence the eventual legislation so that the default option simply reinvests the employees' money into the corporation's own stock. And when some of these companies go belly up and their stock plummets, these employees will lose their retirement savings and end up worse off than they would have been without the default.
4.14.2008 2:41pm
eeyn524:
Two different bases for libertarianism: opposition to coercion, and belief that free markets lead to greater efficiency. If you're in the first group, there isn't much problem with defaults; if you're in the second it still constitutes tampering.

The 401(k) example isn't very clean, because it mixes two issues- whether the employer is being libertarian with respect to employees, and whether the government is being libertarian with respect to the employer.

A related issue: should a libertarian government force private employers to behave in a libertarian way? For example, right to work laws?
4.14.2008 2:47pm
Aultimer:
I think we're going to need more examples or a better definition. The reason an employer chooses enrollment by default for employees is self-interest. Highly compensated employees (who make such choices like default enrollment) benefit from broader participation in 401(k)s. The plan sponsor also benefits from reduced employer-side payroll tax. If that's "paternalism" in your world, your Dad is real jerk.
4.14.2008 2:48pm
K Parker (mail):
Oren,
It is quite rational for most people to be ignorant about the arcane details of matters not directly in their expertise.

I hope you're not arguing that this means we're OK with handing over all authority in those areas of our life to some "expert".
4.14.2008 2:48pm
another anonVCfan:
"A related issue: should a libertarian government force private employers to behave in a libertarian way?"

Government has an interest in making sure employers use neither force nor fraud against their employees. Beyond either of those, how can a corporation act in an un-libertarian way? A major premise of libertarianism is that private parties have the right to alter their obligations to one another through contract, even if the government would be prohibited from coercing similar behavior.
4.14.2008 2:57pm
runape (mail):
"[T]he only genuinely libertarian 'default position', it seems to me, is to do nothing and not intervene at all."

This is a standard libertarian critique, but one that fails to recognize an essential building block of Sunstein's approach.

The idea that trips people up the most is that there is no way to "do nothing" and "not intervene at all." There is a default rule embedded in every agreement. To take the 401(k) example, the default can either be "give, unless employee opts out," or "don't give, unless employee opts in." There is no way to "do nothing"; the employer must choose one option or the other. And since the employer must choose, why should libertarians prefer the opt-in approach to the opt-out? After all, opt-in is more costly for those who prefer to participate, and is, therefore, equally an impediment to choice.
4.14.2008 3:01pm
SIG357:
"Libertarian paternalism" sounds like it might more aptly be named "somewhat less coercive liberalism", or even "big government libertarianism".

And I agree with the basic understanding of human nature underlying Sunsteins ideas. I just think that understanding can be applied in better ways, rather than as an altenate method for applying the liberal agenda.
4.14.2008 3:14pm
SIG357:
There is a default rule embedded in every agreement. To take the 401(k) example, the default can either be "give, unless employee opts out," or "don't give, unless employee opts in."



I think the libertarian position is "What the hell does the government have to do with peoples savings? Why is there even such a thing as a '401(k)' in the first place?"
4.14.2008 3:18pm
SIG357:
K Parker

I hope you're not arguing that this means we're OK with handing over all authority in those areas of our life to some "expert".




Yes, I'd like some clarification on that point myself. There is a reason why "classical liberalism" devolved into todays coercive liberalism, and this sort of thinking is it.
4.14.2008 3:21pm
Adam J:
A.C.- considering the amount of money our government puts into education, no... we apparently don't want to do the hard work of educating people. Regardless, there's still decisions where a greater degree of expertise is required than deciding you should save some money for retirement- a default (hopefully) offers the best solution to these decisions.

another anonVCfan- you're certainly right, I expect the "philosopher kings", as you call them, to get the defaults wrong with some degree of frequency. However, when the people who opt out have more success than the ones that blithly follow the kings, you can expect a revolution. To take your example, people will likely be quite pissed when screwed by following the "king" default, and as a result lose both savings and job, while those who avoided the default lose only job, but not savings. With a few sad stories picked up by the press the rules should change for the better. Also, you should recognize that since the private sector has little interest in ensuring you make the right financial choices (for you at least), and if you lack the resources to make the right choice, this seems to represent a market failure- and market failures are a time when even libertarians should recognize a need for government intervention.

K Parker- if you still have the ability to opt out of the expert's decision, you're hardly handing over all authority.
4.14.2008 3:25pm
frankcross (mail):
I think this proposed approach is potentially quite libertarian. If the default is zero government action, then it might appear contrary. But that is not the default in today's society. If you conceive the "nudge" as the alternative to coercive government action, then it is quite libertarian.

These proposals should not be evaluated in a fantasy world but instead in the real world, where coercive action is a reality.
4.14.2008 3:26pm
ithaqua (mail):

The idea that trips people up the most is that there is no way to "do nothing" and "not intervene at all." There is a default rule embedded in every agreement. To take the 401(k) example, the default can either be "give, unless employee opts out," or "don't give, unless employee opts in."


There is, indeed, a default position for everything. Where you make your mistake is in claiming that all default positions are equally libertarian if an individual is free to choose another position. In your examples above (for example ^_^ ) the former option could be phrased as "take money from my employee's paycheck without permission, unless he complains", and the latter "give my employee the money he earned and offer him this other option". They're both defaults, yes, but the one is less "libertarian" because it presumes to dictate the 'best' way for the employee to use his money.

That isn't to say that I would complain (too much) if individual corporations and do-gooder employers wanted to write this sort of paternalism into their employee contracts. Libertarianism, though, has nothing to say about relationships between employers and employees except that the government should stay out of them. Libertarianism, that is, is an exclusively political philosophy, and to call the idea 'libertarian paternalism' implies a political solution, ie, that governments should require these opt-out 401(k)s and so on. And that sort of top-down management has, again, all sorts of problems.

Two different bases for libertarianism: opposition to coercion, and belief that free markets lead to greater efficiency.

No, libertarianism is about freedom, which is more than an absence of coercion. The government should be the servant of society, not its master or parent or nanny or what have you; anti-drug public service announcements, though less offensive than armed SWAT raids on pot smokers, are still unlibertarian. It's not the government's business to keep people from making stupid decisions, because it's not the government's business to decide for its citizens what decisions are stupid in the first place.
4.14.2008 3:50pm
another anonVCfan:
Maybe it sounds libertarian when we're discussing the 401(k) option because securities are heavily regulated, but there are whole huge sections of our lives that are almost completely unregulated now that could become subject to regulation under this theory.

Let's look at something else, like combating obesity. I suppose it would fit in the libertarian paternalism mold to prohibit grocery stores from stocking candy at the checkout aisles, or to require them to put sugary cereals on the bottom shelf and bran flakes at eye level. A libertarian paternalist would say that consumers still have the freedom to buy those things, they're just being gently discouraged from doing so. But the other side of that coin is that we are now telling grocery stores where and how to stock their goods. Apart from the lost sales and additional compliance costs these businesses will face, we will presumably be sending inspectors into private businesses to see whether the Snickers bars are too close to the cash register. I don't know about everyone else here, but that sounds utterly Orwellian to me, and yet it seems completely compatible with libertarian paternalism.
4.14.2008 3:56pm
Deoxy (mail):

I think the libertarian position is "What the hell does the government have to do with peoples savings? Why is there even such a thing as a '401(k)' in the first place?"


This is true. But, back here in reality, the government IS involved in this kind of thing, and libertarians aren't sufficient in number to do anything about it... but we MIGHT be able to reduce the burden a bit in some practical ways (this being one of them).

In any area where the government demands to be involved, the "default" needs to be one of two things: the option that is, essentially, no government interference, or (if that's just not a choice) the option most people want (as that lowers the total amount of work people have to do because of this intervention).

Interestingly, people's penchant for taking the default makes that second choice above very difficult to really determine, which sucks.
4.14.2008 3:56pm
Dan Hamilton:
It is not stupid, or ignorant. People don't opt in to 401Ks and many other things

BECAUSE THEY ARE LAZY!!!


Everyone knows this to be true. This is not in anyway a putdown but it is why selecting what the default will be is so important.

It is ignorant to allow only the LEFT to take advantage of it. The RIGHT and LIBERTARIANS should wake up and make use of this well known fact. People are LAZY. Make use of it.
4.14.2008 4:03pm
eeyn524:
Ithaqua:

Agreed that freedom is more than absence of coercion, and that ideally the government wouldn't do anti-drug public service announcements. But it would be vastly better than what we have now, and if the idea of defaults can get some non-libertarians to lighten up it is well worth it.
4.14.2008 4:43pm
Pon Raul (mail):
another anonVCfan has the correct answer. Sunstein theory might be libertarian paternalism to the consumer, etc., but not to the company. To the company, it is just plain regulation. How about allowing each company to decide if each of their employees needs to opt in or opt out of automatic savings?

I am not sure about situations where a individual deals directly with the state. I guess that it would depend on the particular situation. I would want to be very careful that we don't impose substantial costs on people electing to opt out, and that might include the cost of having to learn about all of the programs that we need to opt out of.

For instance, libertarian paternalism is somewhat used with the tax forms and the $3 federal campaign financing thing, yet most people still say NO. Sure, the form still says Yes or No, but I doubt that the numbers would really change if you just has a NO box. Accordingly, even if there is automatic enrolement for organ donation when you get your driver's license, I am not sure that very many extra people would say YES if they are given a real opportunity to say NO, and know that they can say NO.
4.14.2008 4:51pm
Adam J:
another anonVCfan- obesity is something where education should be the solution, since people can fairly easily puzzle out the pros and cons of having another snicker bar if they take a little time to get educated. Enforcing defaults should probably only be a solution for complex problems where many rational people cannot afford to spend the time educating themselves to make the best decision.
4.14.2008 5:00pm
runape (mail):
"In your examples above (for example ^_^ ) the former option could be phrased as "take money from my employee's paycheck without permission, unless he complains", and the latter "give my employee the money he earned and offer him this other option". They're both defaults, yes, but the one is less "libertarian" because it presumes to dictate the 'best' way for the employee to use his money. "

I think I understand your position (although I may be about to misstate it). I think you are conflating (1) libertarianism as a philosophy with (2) the policy preferences that most libertarians hold.

We can agree that opting in or opting out has some small cost (call it x). The question is who should have to pay x. Take the 401(k) example again. There are 2 groups of people: one group that wants 401(k)s, and one group that does not want 401(k)s.

The question Sunstein and Thaler raise is: which group should have to pay x for their choice to be honored?

Here is where I think you err. Libertarianism as a philosophy seeks to maximize choice (or freedom). A philosophical libertarian has no problem letting people choose to participate in a 401(k).

That said, many people who hold libertarian viewpoints think participating in a 401(k) is stupid (or inefficient, or both).

The true "free market" approach to the 401(k) problem would be to allow both groups to follow their personal preferences. If participating in a 401(k) is inefficient, market forces will work to make participants reconsider their decision. If it is efficient, those who do not participate will choose to do so.

This leaves the question who should bear the opt in/opt out cost, x. You suggest that it should be 401(k) participants who bear the cost. But why should that group have to pay to have their choices honored? Why shouldn't those who don't want to participate have to pay instead?

What is missing from your account is a principled basis for placing the costs on one group versus another. What you implicitly suggested was that participating in a 401(k) is stupid, so we should force people who want to participate to pay x. But that is not a principled objection, because many people think that not participating in a 401(k) is stupid. Reasonable minds disagree.

One principle you might try is majoritarianism. This works well in extreme cases. (Consider: if one voter thought the South should secede, would it be reasonable for the South to secede unless everyone else voted otherwise?) But in close cases, majoritarianism does not work. Reasonable minds disagree on the merits of 401(k)s.

I'm interested to know whether there are other grounds for your objection to paying to have your choice honored. (Incidentally, I tend to agree with you that 401(k)s are often a bad idea.)
4.14.2008 5:47pm
another anonVCfan:
Interestingly enough, voting is a classic example of a high-cost activity where many people remain rationally ignorant. Since it's safe to assume everyone wants to be healthier, wealthier, and happier, maybe we just need to give all elligible voters default votes in favor of libertarian-paternalist candidates. If you prefer a different candidate, just send in an absentee ballot (we'll even give you the stamp!). That way we don't prevent anyone from voting for non-paternalists, we just give them a nudge in the right direction.

/snark
4.14.2008 5:49pm
another anonVCfan:
"What is missing from your account is a principled basis for placing the costs on one group versus another."

Libertarians don't (or shouldn't) care which group pays this cost. If an employer thinks that 401(k)s are great and wants to nudge their employees to participate, libertarians believe they have that right, so long as the employee consents. What libertarians care about is the coercion inherent in the government mandating that nudge.
4.14.2008 5:57pm
PatHMV (mail) (www):
Prof. Sunstein, I wonder if your argument might benefit from providing specific examples about the plethora of "default" laws we have today, and their very common application in day-to-day life by almost all of us. Perhaps you do in your book.

The most common example, I would think, is leases. Under the laws of most states, almost every statute provided for renting property is a "default" rule, which can be easily altered by the parties by contract. We have to have such default rules, because we know from experience that people very often do NOT make sure that all possible problems are covered by the contract. Disputes regularly arise over issues that were never contemplated by the parties, and thus were not provided for in the contract.

Many times, the parties to the transaction do not wish to even try to contemplate all eventualities. Some rely on handshake deals, contracts scribbled on the back napkins (literally), or simple time-honored customary rules. In such cases, a judge resolving the dispute cannot simply examine a contract to determine the parties' intent and expectations. Yet the court must STILL resolve the disagreements that arise. The judges can't simply say: "shame on you parties for not hiring a $1,000 an hour lawyer to create a perfect, all-contingencies-addressed lease agreement."

Yes, default rules have a long and vital role in our society.
4.14.2008 6:09pm
runape (mail):
"Libertarians don't (or shouldn't) care which group pays this cost. If an employer thinks that 401(k)s are great and wants to nudge their employees to participate, libertarians believe they have that right, so long as the employee consents. What libertarians care about is the coercion inherent in the government mandating that nudge."

I think this is right - the crucial idea is that someone is "mandating the nudge" (read: setting the default) regardless of which group pays the cost. But there is no philosophical libertarian objection to switching the default.

An interesting (but separate) question is who should set the default - the government or someone else. I agree it should often not be the government; but surely there are instances where, because of collective action and related problems, the government is the best choice.
4.14.2008 6:24pm
another anonVCfan:
There's a vital difference between the default rules you've just described and, for example, the opt-out 401(k): the rules you described only come into play if there's a conflict between the parties that they haven't planned for. In the 401(k) example, there's no conflict; employer makes an option available, the employee takes it (or not), and neither feels they have any claim against the other. It makes perfectly good sense to have default rules in the former because there's a live controversy that has to be resolved, and the rule of law requires that we resolve these things similarly for similarly situated parties. Injecting default rules into situations where there is no conflict is a different undertaking. There you're saying, even though you're both satisfied with the status quo, we think one of you could be happier, and we're going to require the other one of you to pay for it.
4.14.2008 6:28pm
cjwynes (mail):
The difference between an opt-out and an opt-in for 401(k) withholdings is that one of them requires an affirmative act for me to keep control of my money and one of them does not.

But as other commenters observe, the example which has been put forward here is designed to seem relatively innocent and minor. I suspect the social engineering types in the government would be unlikely to stop at boring things like 401(k) accounts. They would much prefer to get involved in the sort of anti-obesity campaigns proposed in some of the commenters' hypotheticals. Whether Prof. Sunstein endorses that level of paternalistic meddling or not, we don't know yet, but his ideas would probably be carried in that direction sooner or later by the left-wing fascists looking for a theory to act as cover.
4.14.2008 6:31pm
PatHMV (mail) (www):
another anonVCfan: While there are some differences, I don't think that they are that great, depending on how Prof. Sunstein proposes to implement his ideas. Many parties to leases intentionally omit provisions precisely because they know that the default rule is a reasonable one. Other times, they choose to write in the current default rule, simply because they like it and fear the legislature might change it at some point in the future. The point is, parties rely on their knowledge of the default rules in constructing their contract to begin with, even before there is any conflict. The default rules shape their behavior.
4.14.2008 6:45pm
Carolina:

Let's look at something else, like combating obesity. I suppose it would fit in the libertarian paternalism mold to prohibit grocery stores from stocking candy at the checkout aisles, or to require them to put sugary cereals on the bottom shelf and bran flakes at eye level. A libertarian paternalist would say that consumers still have the freedom to buy those things, they're just being gently discouraged from doing so. But the other side of that coin is that we are now telling grocery stores where and how to stock their goods. Apart from the lost sales and additional compliance costs these businesses will face, we will presumably be sending inspectors into private businesses to see whether the Snickers bars are too close to the cash register. I don't know about everyone else here, but that sounds utterly Orwellian to me, and yet it seems completely compatible with libertarian paternalism.



Well-said, sir.
4.14.2008 6:56pm
runape (mail):
"It makes perfectly good sense to have default rules in the former because there's a live controversy that has to be resolved, and the rule of law requires that we resolve these things similarly for similarly situated parties. Injecting default rules into situations where there is no conflict is a different undertaking."

What is an example of a situation in which there "is no conflict"? In essentially every interaction (save the truly extreme cases), reasonable minds will disagree about the optimal policy approach.

"The difference between an opt-out and an opt-in for 401(k) withholdings is that one of them requires an affirmative act for me to keep control of my money and one of them does not."

I think this is another confusion of libertarian philosophy and individual preference. You want to keep control of your money - that is, you prefer to place the cost of switching the default on those who would opt in. But what is the principled basis for letting you enjoy your preferences cost-free, while forcing other people to enjoy their preferences only at cost?
4.14.2008 6:58pm
another anonVCfan:
"What is an example of a situation in which there "is no conflict"? In essentially every interaction (save the truly extreme cases), reasonable minds will disagree about the optimal policy approach."

The fact that outside observers disagree about the correct policy approach doesn't mean there's a conflict between the only people we should be concerned about, namely, the folks who are interacting. There's no conflict in the 401(k) example because both parties voluntarily reached an agreement; neither coerced the other and both are apparently satisfied. The only person who's bothered is the libertarian paternalist, who thinks that one party would be happier if only they cared enough to make a "better" choice. But merely thinking that you what's best--even if it's demonstrably the best choice--doesn't by itself give anyone the right to inject themselves into other adults' affairs.
4.14.2008 7:26pm
Adam J:
"I suspect the social engineering types in the government would be unlikely to stop at boring things like 401(k) accounts."

Ah, everyone's favorite crappy argument, the slippery slope, rears its ugly head.
4.14.2008 7:33pm
gallileo:
"But what is the principled basis for letting you enjoy your preferences cost-free, while forcing other people to enjoy their preferences only at cost?"

The libertarian-principled basis is not hard to find here, really.

"Reducing the possible uses of a piece of property should require the owner's active decision."

The present 401K proposal *automatically* places money beyond the owner's control unless she affirmatively chooses otherwise. Once the money is in the 401K, the freedom the owner has is greatly reduced. If she finds the trade-off worth it, then more power to her, but libertarian principles state that it should be an active decision to decrease her options, not a passive one.

Today, if I erroneously choose not to opt in to my 401K, the money is still available to invest for retirement, or consume, or invest however I prefer.

Under the present proposal, if I erroneously choose not to opt out of my 401K, then the money's use is greatly restricted, and I suffer a significant penalty.

This is not just a personal policy preference--it is the libertarian principle that people should be able to dispose of their property however they want, for as long as they own it.

If I make an error today, I still own the property and can still dispose of it however I want. If I make an error with your system, I no longer own the property in the same way, and my options are greatly reduced.
4.14.2008 7:40pm
cjwynes (mail):
runape,

It seems to me is that the primary purpose of libertarianism is to protect our natural rights to life, liberty and property. The government should be protecting our ability to control the exercise of those rights -- that's it's only purpose from a libertarian perspective.

The law mostly presumes that people will prefer to maintain control over their own property. There are exceptions like adverse possession and trademark dilution (which may be opposed by some libertarians, for that very reason), but in general we are not required by the law to take any affirmative steps to protect our natural rights. Indeed, surrendering our rights generally requires some affirmative act on our part. This is basically an extension of the distinction between acts and omissions, a distinction which is central to libertarian philosophy.

If it's all just a matter of whose preference should bear the costs, then how about this: instead of letting people opt out of their right to remain silent, let's have them choose whether to opt in. From now on, you have NO right to remain silent while in police custody unless you fill out a form saying that you want to preserve that right. After all, in my experience as a publice defender and now a prosecutor, a majority of suspects who are questioned choose to give statements. Why should the MAJORITY who chooses to blab to the cops about their misdeeds have to fill out that Miranda waiver? Why shouldn't the minority who actually exercise that right have to bear the cost?

Yes, I suppose you could reverse the presumptions our law typically makes and instead create a world where people who pay less attention to their affairs are routinely having their natural rights taken away from them at every turn. You could require that people repeatedly have to take specific actions to say "yes, I really do want to preserve my rights." That world would certainly be paternalistic, but it wouldn't be libertarian.
4.14.2008 7:49pm
Adam J:
another anonVCfan- "The only person who's bothered is the libertarian paternalist, who thinks that one party would be happier if only they cared enough to make a "better" choice." I'd love to see your empirical evidence for that one. Personally, I know alot of people who are worried they've made the right decisions regarding taxes, mortgages, saving, etc., because they know that they lack the expertise to make the best decision.
4.14.2008 7:51pm
Lily (mail):
Dear Busybody Do-Gooders:

Leave me alone to make my own decisions.

Thank you,
Lily
4.14.2008 11:29pm
CDU (mail) (www):
1. People are lazy.

2. Thinking is work.

It's really astonishing how much these two statements explain about the world.
4.15.2008 12:56am
wuzzagrunt (mail):
Adam J:
"I suspect the social engineering types in the government would be unlikely to stop at boring things like 401(k) accounts."

Ah, everyone's favorite crappy argument, the slippery slope, rears its ugly head.

Let's set up a demonstration project concerning union membership. What is the default position? What is the opt in/opt out procedure?

In the real world, that would be very entertaining.
4.15.2008 12:58am
David Schwartz (mail):
I don't think it's useful or sensible to try to consider public actors and private actors at the same time. And, in the case of private actors, I don't think it's useful to consider what they may do and what they should do at the same time.

So long as we are talking only about what private actors may do, they are free to nudge as they please and they cannot ever be unLibertarian. When we talk about what private actors should do, perhaps sensible defaults can decrease the costs of getting what you really want in some cases. And people who get it wrong cannot impose costs on people who are unwilling to bear them.

However, when you talk about public actors imposing costs on choices, then this "libertarian paternalism" is actually totally unlibertarian. For one thing, very little within the scope of government that libertarians accept is subject to default rules.

And will the government just set the default default? Or will individuals be free to choose their own set of defaults?
4.15.2008 1:32am
e:
sig357 raised a point I was hoping would be addressed by others: that the 401(k) construct is not libertarian in the first place.

Rather than recognizing a more complete personal liberty, the gov't grants a limited choice of how you can characterize your earnings for tax purposes. The gov't allows a category of limited individual control for a marginally better tax treatment. Perhaps this is an illusory freedom against more socialist alternatives, but libertarians would likely favor alternatives which reduce the overwhelming regulation within our tax code.

401(k) opt in/out defaults seem more like "moderate paternalism." As a moderate I'm okay with either of those binary (and socially coercive) policy choices, but I sure don't claim that either is anywhere near some ideal of either individualism or collectivism.
4.15.2008 2:30am
Dan Hamilton:

Or will individuals be free to choose their own set of defaults?


Did you really want to say that? Or did you just have a brain short?

The defaults are set either by the government (almost always) or by the Corp. Never by the individual.
4.15.2008 11:13am
LarryA (mail) (www):
Ah, everyone's favorite crappy argument, the slippery slope, rears its ugly head.
Please name one federal government program started more than twenty years ago that has not metastasized beyond what they swore would never happen. Show your work.
The other problem is that with your criticism is this statement "top-down command economy or command society will make worse decisions, by whatever metric you measure by, than decisions made by the individual people involved" can be wrong, because individuals frequently don't have the knowledge necessary to make the correct decision- individuals frequently have more important issues on their mind, like getting food on the table.
There is no "correct" decision. As an individual, I have far more knowledge about my personal situation than even my Federal Congressman does, and we know each other. In this case, for instance, blithely deciding that "everyone needs an employer retirement plan" isn't nearly true. The government simply cannot make individual decisions for 200 some odd million individuals.

Opt out/opt in? Who cares?

Wouldn't it be nice if you could go to a bank or other financial institution, set up a retirement/health insurance account with the same tax breaks as the government-favored employer-related plans, so you and your employer wouldn't have to worry about all that stuff every time you change jobs?
4.15.2008 12:02pm
Toby:
Let's posit that the opt-in, opt-out has been made almost completely cost free, and eliminate the costs part of the discussion.

1/3 always opt in, fine
1/3 always opt out, fine
1/3 to lazy/confused/whatever to decide.

If that 1/3 shifts toward the personally controlled savings side of the scale, they will have habits of thought and expectations that are less alligned with the Leviathan state as supplier of all benefits. Politically, this is a good thing for liberty.
4.15.2008 6:02pm