David R. Henderson  

Thomas C. Leonard on Nudge

PRINT
Dr. Henderson in Africa... Overturning the Soda Ban: One...

See Update Below.

While web surfing the other day, I came across work by Princeton economist Thomas C. Leonard. There's a lot of good work there. In this post, I want to highlight his review of Nudge by Richard Thaler and Cass Sunstein. Up until now, my favorite review of their book has been my own [Scroll down]. I do what only a few reviewers do: get the exact argument made by the author completely accurate, never exaggerating, never playing cheap rhetorical tricks. Then, when I criticize, the author can't claim that I misstated him.

Thomas Leonard, based on this one review, at least, seems to follow the same rule. And, in doing so, he comes up with a more fundamental criticism of Nudge than my own. As a result, my review of Nudge is now my second favorite.

Like Thaler and Sunstein, I had taken it as given that a libertarian paternalist--I explain in my review why this isn't a contradiction in terms--should be trying to design the "choice architecture" so that people make the choices that their "rational" side--the "Planner"--would make. But Leonard points out the problem with this. He writes:

Set aside the problems of bad nudges and evil nudgers. Assume effective instruments and benign intent. What does mean to make a time-inconsistent person better off when her Doer and Planner incarnations have opposed views of what is better? Tom Schelling posed the dilemma this way: ''how do we decide which side we are on?'' (1984, p. 87). On what grounds should a paternalist privilege the Planner?
Say I prefer $100 today to $110 next week (Doer), and I also prefer $110 in 53 weeks to $100 in 52 weeks (Planner). My preferences are inconsistent: I am more impatient in the near future than in the far future. But there are two ways to nudge them toward consistency: promote impatience (Doer) or restrain impatience (Planner). Thaler and Sunstein acknowledge the dilemma for the paternalist facing a divided self: ''If the arrangement of [cake and fruit] alternatives has a significant effect on the selections the customers make, then their true 'preferences' do not formally exist'' (2003, p. 1164).
We certainly don't want Homer Simpson at the helm of Ulysses' vessel when the cost of succumbing to the Sirens' song is mass death. But the choice between Doer and Planner is rarely so stark. How much should a person save? What is the optimal tradeoff between cost of weight gain and the pleasure of tasty calories? Are workaholics, anorexics and misers--what happens with a dictatorship of the Planner--any less immoderate than slackers, the overweight, and the profligate?
Sure, more people make New Year's resolutions to lose rather than to gain weight, but that's just Planner talk. Many others resolve to spend less time at the office, and more time enjoying life with family and friends, Doer talk.

And, immediately following, the crescendo:
The irony is that behavioral economics, having attacked Homo Economicus as an empirically false description of human choice, now proposes, in the name of paternalism, to enshrine the very same fellow as the image of what people should want to be. Or, more precisely, what paternalists want people to be. For the consequence of dividing the self has been to undermine the very idea of true preferences. If true preferences don't exist, the libertarian paternalist cannot help people get what they truly want. He can only make like an old fashioned paternalist, and give people what they should want.

UPDATE: In response to commenter Ted Levy, here is a link to much of Thomas Leonard's other work. There's a treasure trove there.



COMMENTS (23 to date)
Ted Levy writes:

Based only on your selections, that is a massively powerful critique, which deserves more play. Thanks for providing it.

David R. Henderson writes:

@Ted Levy,
Thanks, and you're welcome. Your comment motivated me to post an update with a link to some of his other work.

Ken B writes:

A powerful objection indeed, as Ted Levy says. I always worry when I detect an engineer of human souls at work, to borrow a phrase.

Glen writes:

This is a criticism of libertarian paternalism that Mario Rizzo and I have been making for some time. Here are a couple of blog posts where I talk about the illegitimate jump from "inconsistency of preferences" to "choosing among those preferences":

http://agoraphilia.blogspot.com/2009/11/new-paternalism-on-slippery-slopes-part_09.html

http://agoraphilia.blogspot.com/2009/11/new-paternalism-on-slippery-slopes-part_14.html

And in my Cato Unbound debate with Thaler, I made the same objection; scroll down the section on "Choosing Among Preferences":

http://www.cato-unbound.org/2010/04/05/glen-whitman/the-rise-of-the-new-paternalism/

Jonathan Bydlak writes:

Glad to see my senior thesis advisor - and a great teacher - getting the credit he deserves. FYI, he goes by Tim.

David R. Henderson writes:

@Jonathan Bydlak,
Thanks. I just checked your web site. Pretty impressive. What was the title of your senior thesis?

Jonathan Bydlak writes:

Thanks, Dr. Henderson. My thesis was titled "So You Want to be a Rock & Roll Star?", and looked at the popular music market through the lens of Sherwin Rosin's superstar hypothesis. Dr. Leonard also was my advisor for my junior year independent work, which looked at the relationship between economic conditions and presidential election outcomes.

David R. Henderson writes:

@Glen,
I just finished reading your Cato Unbound piece. It's excellent! Here are the two most relevant paragraphs:
There is some dispute as to whether all such behavioral inconsistencies reveal irrationality. But let’s say they do. Even so, that fact does not license a third party to choose among competing preferences. If a person is more patient when thinking about trade-offs in the distant future, but less patient when thinking about trade-offs near the present, which level of patience is “correct”? If you would sleep with a given person when you’re in a “hot” state but not in a “cool” state, which sexual preference is “correct”? Neither theory nor evidence provides a basis for answering these questions. As some new paternalists admit, behavioral inconsistencies may indicate that “true” preferences simply don’t exist.

Nevertheless, new paternalists have not hesitated to pick and choose the “right” preferences. O’Donoghue and Rabin, for instance, define “optimal sin taxes” in terms of a person’s most patient rate of time preference. Similarly, the new paternalists favor the preferences we display in a cool state (calm and sober reflection) over those we have in a hot state (fear, anxiety, arousal, etc.), even though arguably the “hot” preferences might do a better job of revealing our true desires.
I was very disappointed, although not totally surprised (Richard Thaler and I have been friends, or friendly acquaintances, for 37 years) by Richard's flippant response. He chose easy examples of where they're not trying to impose preferences (wine, Coke, T-shirts, etc.), thus avoiding the important issue that you and Leonard raised.

David R. Henderson writes:

@Jonathan Bydlak,
My thesis was titled "So You Want to be a Rock & Roll Star?", and looked at the popular music market through the lens of Sherwin Rosin's superstar hypothesis.
Cool! The bottom line being, I'm guessing, that it's like entering a lottery where the expected gross gain is reasonable but the ticket is very expensive? Did you get an article out of it?

perfectlyGoodInk writes:

Leonard:

So, Nudge defends three main claims: one, the architecture of choice greatly influences how people make choices; two, choice architecture is unavoidable (so why not design in ways that improve well being), and three, libertarian paternalism is not an oxymoron: paternalists can nudge while preserving freedom of choice.

The review doesn't seem to attack claims 2 or 3 at all, but mostly questions whether how to decide what's a good nudge or not.

However, on the question of retirement savings, if opt-out is manipulation, then the previous choice architecture of opt-in was equally manipulative of influencing people into saving nothing (regardless of whether that was the intent, per claim 3). Sunstein and Thaler contend that freedom is no more constrained under opt-out than opt-in (claim 2), and Leonard does not seem to argue this point.

However, Leonard is curiously silent on the question of whether he believes the nudge towards saving nothing does a better job enhancing welfare.

perfectlyGoodInk writes:

Sorry, I swapped claims 2 and 3.

Jonathan Bydlak writes:

Yeah, the point was that the unequal distribution of earnings in the marketplace exists because no one wants to listen to second-best artists. Unlike in many other markets, consumers place high premiums on perfection, so without an unequal income distribution, the goods consumers demand would not be produced.

From an empirical perspective, there's evidence that changing economic conditions alter people's preferences for music consumption, which ultimately gets reflected in the popular music charts (e.g. how long songs stay at the top, how many artists chart, etc.).

No article came out of it, but you can go down to the Mudd Library and check it out if you like!

Mike W writes:

From your review...even if old people did have a property right [in Social Security], why should the government single out young people to honor it? If the government is to honor it, should it not go after the people who made the commitment?

Yeah, Ok so far...but then,

One way to do this would be to attach all the property of all the congressmen and presidents who ever voted for or signed legislation to increase Social Security payments. This might sound extreme — no, this is extreme — ...[b]ut is it less just than having the government go after the future earnings of innocent two-year-olds?

Why are politicians to blame for responding to their constituents? Shouldn't the default be that we should not expect that much courage from them? Why wouldn't the better...and more realistic...solution be to reduce the unaffordable benefits or increase the taxes on the old people? These folks during their working lives made Social Security the "third-rail of American politics" and thereby avoided paying into the system an amount sufficient to fund their benefits without burdening future two-year olds. Shouldn't they pick up the tab?

I know, not on the behavioral econ point...

Tom West writes:

I think the differences here are close to axiomatic, and are best illustrated by Bryan's very illuminating line that he could not understand the idea of people denying themselves future choices (like locking the fridge and giving someone else the key).

If I understand correctly, to him (and perhaps most Libertarians), if you choose an action (i.e. it's unforced and you're properly informed), then by definition that choice was the best one for you. You weighed pros and cons along with your time discount and chose accordingly.

My strong suspicion, however, is that this is not a viewpoint shared by the majority of humanity. I think most human beings have felt that their rationality is limited, and while saddled with a high time discount, some part of them wants to maximize happiness over the long-term.

Such people (such as myself) will welcome externally imposed nudges (or in fact, actual limitations) in order to be guided/forced into a course of action that will result in maximizing happiness over one's lifetime.

Thus for such people, the idea of an authority nudging them in a direction which seems likely to increase their net happiness doesn't seem a bad thing. It's why most students will, given a choice, choose *not* to have all the assignments due at the end of term when given the option. They understand their irrationality, and choose to have their freedom curtailed by having assignments due earlier than strictly necessary (i.e. a nudge that as the assignment is due, they may dislike.)

As I said, I think disagreement here devolves down to base definitions about whether you *can* make a sub-optimal choice.

perfectlyGoodInk writes:

Tom West: If I understand correctly, to him (and perhaps most Libertarians), if you choose an action (i.e. it's unforced and you're properly informed), then by definition that choice was the best one for you. You weighed pros and cons along with your time discount and chose accordingly.

Tversky and Kahneman demonstrated empirically that framing identical choices differently caused people to choose differently. Presenting a scenario that, out of 600 people, Program A saves 200 people and Program B has 1/3 chance to save all and 2/3 chance no one will be saved, 72 percent of respondents chose A -- despite the expected value being equal. What was more striking was that when they restated the same choices as "400 people will die" and "there's a 1/3 chance nobody will die", 78 percent to choose B.

This finding holds up with the changes from opt-in to opt-out for 401(k) plans. If framing effects didn't matter, nobody here would be arguing about nudges because they would be completely irrelevant.

The left/right argument has instead shifted from "do people systematically make irrational decisions" -- which most economists concede does occur -- to "is the government any more likely to make rational decisions for people." Of course, my argument: most definitely not if you nudge people towards zero savings. Note, I think Thaler and Sunstein's most compelling argument is that nudges are unavoidable even if you do not intend to make them.

Ted Levy writes:

"Thaler and Sunstein's most compelling argument is that nudges are unavoidable even if you do not intend to make them."

I do not see why this is compelling, at least if that means "compelling as regards the need for government intervention". With "libertarian paternalism" you have the government making a unilateral choice because "nudges are unavoidable". With actual plain old libertarianism, you have a decentralized process where numerous market actors make multiple options available, creating multiple nudges for you to choose among. A market in nudges.

Jim Rose writes:

Vernon smith, when asked about behavioural economics, wondered how so cognitively flawed a creature made it out of the caves.

The answer had a lot to do with the institutions that emerged to overcome human limitations. The double auction has been clearing markets for 1000s of years.

The behavioural economics is an excellent example of how engaging in JS Mill’s truth that engaging with people who are partly or totally wrong sharpens your arguments and improves their presentation.

People have a better understanding of rationality such as through the work of Vernon smith on ecological and constructivist rationality and of how people deal with human frailties and correct error through specialisation, exchange and learning.

• George Stigler’s in his Existence of x-inefficiency paper opposed attributing behaviour to errors because error can explain everything so it explains nothing until we have a theory of error.

• Kirzner wrote a response saying that error is pervasive in economic processes. Rational Misesian human actors are human enough to err.

What is inefficient about the world, said Kirzner, is at each instant, enormous scope for improvements exist in one way or another and is yet simply not yet noticed.

The lure of entrepreneurial profits harnesses the systematic elimination of error and points the way to the institutions necessary for the steady social improvements.

Behavioural economics is a clumsy way of discussing the pervasiveness of errors because insufficient attention is paid to decentralised, emergent market processes that correct them, often long ago.


Thomas DeMeo writes:

"Set aside the problems of bad nudges and evil nudgers. Assume effective instruments and benign intent. What does mean to make a time-inconsistent person better off when her Doer and Planner incarnations have opposed views of what is better?"

Nudging is a tool. It doesn't need to address every dilemma. Why not just be selective?

perfectlyGoodInk writes:

perfectlyGoodInk: Thaler and Sunstein's most compelling argument is that nudges are unavoidable even if you do not intend to make them.

Ted Levy: I do not see why this is compelling, at least if that means "compelling as regards the need for government intervention".

Because Leonard's argument seems to be that nudges are a bad idea because it is difficult to determine the optimal level of savings. Not only is this making perfect is the enemy of the good, the unavoidable nature of nudges renders this argument irrelevant. There was already a nudge towards zero savings. This means that the argument isn't about whether or not there should be a nudge, because nudges are already everywhere.

Leonard also opens a can of worms here:

Leonard: The subtitle of Nudge is ‘‘improving decisions about health, wealth and happiness,’’ but it would be more accurate if it read ‘‘manipulating decisions about health, wealth and happiness.’’ After all, the consequences of manipulation depend upon the nudger’s intent, which may well be to exploit rather than to ameliorate, and also upon the effectiveness of the nudge in question

The traditional goal of libertarians is to further freedom and liberty by maximizing freedom of choice, but acknowledging that government manipulation can influence choices means that firms can also manipulate people.

Market competition is supposed to maximize societal welfare because firms face incentives to maximize profits by lowering prices and increasing quality. Presumably, this is why you think a marketplace of nudges would be desirable. However, this assumption is called into question if firms have the ability to increase profits by manipulating people into buying things they didn't actually need or want, and which may have long-term harmful effects.

Market competition indeed becomes a marketplace of nudges, where firms compete on being better manipulators of taste. Leonard is very much correct that the nudger's intent may well be to exploit, because the ability to manipulate choices means that power exists without coercion.

Kevin Driscoll writes:

@perfectlyGoodInk I think Libertarians have long been aware of the difference between experienced utility and decision utility. They know that firms lie and marketing matters. The debate is not about government nudges vs. the perfection of the market.

Rather, the question is one of tradeoffs. Given that nudges always occur and it is difficult (I think impossible) to determine what the 'correct' actions are, whose nudges do we prefer? The libertarian critique is and has been (since Mises?) that economic actors possess local information that would be impossible for a central planner to have. The fact that those actors can have malicious intent and be manipulated is no real objection if the same can be said of the central planner (and I think it can).

perfectlyGoodInk writes:

Kevin Driscoll: The fact that those actors can have malicious intent and be manipulated is no real objection if the same can be said of the central planner (and I think it can).

Yes, I alluded to this when I said: The left/right argument has instead shifted from "do people systematically make irrational decisions" -- which most economists concede does occur -- to "is the government any more likely to make rational decisions for people."

I don't think anybody here is truly suggesting malicious intent. What we have are individual actors acting in their own self-interest, and those interests are often at odds. To me, the pervasive existence of nudges/manipulation indicates the importance of focusing on power differentials -- which actors have the greater ability to manipulate behavior of other actors?

Kevin Driscoll: The libertarian critique is and has been (since Mises?) that economic actors possess local information that would be impossible for a central planner to have.

Well, since Hayek, actually. I'm hardly suggesting scrapping the information aggregating feature of the price system in markets. However, Hayek believed that the profit motive in markets would cause these distributed actors to allocate resources efficiently towards scarcity.

While I agree, he does not seem to factor into account that the same actors face similar incentives to allocate resources inefficiently towards manipulation as well as rent-seeking. It also seems to overlook that different actors have large differences in resources and ability to aggregate information, which lends towards increasing information asymmetries (another source of power).

So it strikes me a case of Hayek seeing what he wants to see. The left sees market failure everywhere and the right sees government failure everywhere. Really, I feel that the left/right divide oversimplifies the issue by presenting a false dichotomy between central planning and markets. Our current reality is that of a mixed economy, and the important debate ought to be what kind of mix, and not which extreme is less worse.

Indeed, governments and markets aren't entirely separate and independent entities because they influence/manipulate each other. Note that one implication of the pervasive existence of manipulation is that firms can manipulate voters as well as consumers.

The power of nudges and manipulation indicates to me that it is narrow-minded for libertarians to focus exclusively on coercion. Coercion is often a poor tool, as it is very visible and can provoke reprisals. Real power is getting somebody else to do something they wouldn't have (preferably in such a way that they thought it was their own idea in the first place, ala Inception).

Jayson writes:

The best critique is by robert sugden in 2008 in the journal: constitutional political economy. He makes the same point as in the Leonard quote and many more

Thomas Sewell writes:
Say I prefer $100 today to $110 next week (Doer), and I also prefer $110 in 53 weeks to $100 in 52 weeks (Planner). My preferences are inconsistent

I'm a little late to the party... but while I agree with the overall argument, I think this is a weak example.

If you take into account the fact that you have more certainty about the immediate future and more uncertainty about the more distant future, than it can be non-contradictory to prefer a certainty of what you'll buy with $100 today vs. the uncertainty of what you'll be able to buy with $110 next week, while at the same time, the uncertainty levels of 52 weeks vs 53 weeks are about the same today (when you're making the choice), so your time discount level could have changed enough to change your choice.

In other words, "today", or even "near future" is very different than "distant future" in terms of certainty around what you can do with that money.

Minor defect, to be sure, but it jumps out at me when I see time-based arguments about discount values being inconsistent.

Comments for this entry have been closed
Return to top