« Preparing to Teach Open Source Civ Pro Part III: Compiling the Materials (or, Why I Have the Best RAs in the World) | Main | Tragic cases and Fed Courts »

Friday, August 30, 2013

Judging Nudging

Imagine that, in order to discourage obesity, New York had hired a band of professional wrestlers to throw on their masks and rove the City body-slamming anyone caught drinking a large sugary beverage.  What are the main economic problems with that approach?  Obviously the administrative costs would be pretty high, though chiropractors would be happier. 

Most commentators would identify two other big problems.  One, the body slam imposes a pure welfare loss on slammees; in comparison, if we had used a soda tax or a fine, we would putatively have had resources we could have transferred for the betterment of society.  And two, we don’t have a good sense of whether the slams are the right level of deterrence.  Since we can’t easily observe the subjective cost of being slammed or the costs of cutting back on cola, we don’t whether we are over-deterring some soda sippers (or under-deterring Coke addicts).  In contrast, a deterrent that was denominated in dollars would give a clearer measure of subjective valuation.

These two features are the core of environmental economists’ case against command-and-control environmental regulation, of Bentham’s case against prison, and of Ed Glaeser’s case against nudges, but in a new working paper I suggest that there is more to the argument. 

On the informational point, I argue that we can use small-scale government experiments to match cashless incentives with their price equivalent.  A group of Harvard and MIT economists just did that recently, measuring the dollar-equivalent effect of a series of “sticky default” contract terms.  Once we know that, the informational advantages of cash incentives is greatly lessened.

 The first, transfer, point is more complicated.  One key aspect I’ll highlight here is that transfers are not free; when you impose a soda tax, you potentially distort people’s behavior in inefficient ways.  For example, if I’m motivated to go to work in order to be able to buy the things I like, and the City’s tax leaves me with fewer of those things, maybe I work less.  Though we don’t have any good evidence yet, it’s possible nudges don’t carry similar costs.   Maybe I’m only drinking all that soda because it’s fizzing so temptingly in its 32-oz cup.  Give me a 16-oz. cup, and I’m just as happy.  So while the nudge may bring in fewer revenues, it also may have lower social cost.  It’s theoretically ambiguous which effect is more important.    

Another aspect of transfers it that by their nature they move wealth from one group to another.  There is a complex literature on how to structure transfers---for example, should you use rewards (carrots) or penalties (sticks)?  Most choices come with some difficult tradeoffs.  In many settings, a policy with no transfers splits the difference, giving a third set of tradeoffs that might be preferable in some cases. 

So, long story short, we currently have no great reason to think that 16 ounces is the right default size for a cup of soda.  But the general idea of using nudges in place of other traditional policy instruments at least isn’t crazy.

Posted by BDG on August 30, 2013 at 12:53 PM | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Judging Nudging:


I'll be the lone contrarian here in the comments by saying that no government has any business (or legal right) to "nudge" me, nor take any interest in my actions unless or until I harm someone else.


Posted by: BackwardsBoy | Sep 5, 2013 9:29:02 AM

Actually, all it does in my case is reduce the amount of ice I get.

I used to have a 54(!) ounce "Super Big Gulp Plus" mug from 7-11 that I would fill up. My procedure was to fill it to the top with ice (and the 7-11 ice stacks fairly tightly).

I like cold beverages.

Once I filled the mug up at home with ice and poured in a 20 oz bottle, and it went all the way to the top.

In a 16 oz container, I can't pour an entire 12 ounce can of soda in all at once.

So all large drink size bans to for me is piss me off.

Posted by: blake | Sep 4, 2013 2:01:48 PM

One last comment, with the caveat that I (obviously) still haven't read your article yet: I agree that the intent of the Big Gulp ban is to reduce the amount of soda (and hence calories) consumed, not to ban 32+ oz cups of soda for their own sake. But inasmuch as nudges' normative justification is libertarian paternalism, I'm not sure we should be determining whether something counts as a nudge from the perspective of the regulator; rather, it seems to me we should be focusing on interference with liberty from the perspective of the regulated.

From that perspective, I agree that the ultimate goal of someone who wants to purchase a 32+ oz soda is to consume a large quantity of soda. But, as I said before, depending on the circumstances, it might, in fact, matter to a consumer whether they are free to consume that soda in the form of one large cup or two smaller cups purchased either simulataneously or sequentially. So I think it comes down to the nuisance costs, and I'm not sure if these will necessarily be sufficiently small to fall within the nudge space, as T&S have defined it. If you're on a road trip trying to stay awake with lots of caffeine and have to pull over a second time for your second 16 oz. cup, is that as easy and costless as having to check a box to opt out of the IRA? What about having to miss some portion of every $18 movie you paid to see to get a refill on the 16 oz? And it may be more difficult (in some cases, prohibitivley difficult) to carry and store multiple 16 oz cups at once.

In short, we can, as you say, describe an IRA default option as a ban on spending that money unless you opt out -- in which case, the reverse default option must be described as a ban on saving money unless you opt in. So perhaps the more important question is whether there's a meaningful difference between checking a box in order to reverse the effect of the regulation (whether we call it a nudge or a ban/shove), on one hand, and whatever effort is involved in reversing the effect of the Big Gulp ban/nudge, on the other. And I guess I'm not sure that there isn't a meaningful difference between the two -- which isn't to say that a soda ban/nudge couldn't be justified on other grounds, only that one might need to rely a bit more heavily on paternalism and/or prevention of negative externalities than on libertarian paternalism.

(There's also an interesting subsidiary question about how we should assess the effort needed to reverse a nudge when that effort will differ from person to person and context to context. The IRA default nudge asks more or less the same effort of everyone who wants to reverse it: check a box (although some might have to spend more cognitive effort understanding what and why they're reversing than others?). The soda ban/nudge, by contrast, will require different levels of effort to reverse and involve different degrees of nuisance without a reversal on different people, depending on whether they're the road tripper or movie goer I described above or someone for whom two cups or a refill is no big deal.)

Posted by: Michelle Meyer | Aug 31, 2013 12:51:44 PM

Thanks for the response, Brian. You're kind to say so, but my forthcoming piece (assuming it doesn't get killed in favor of more important breaking news) is just a WSJ piece keyed to the announcement of Obama's Nudge Unit; I'd be surprised if you learned anything from it (although obviously we hope a lay audience will). I will, however, look forward to reading your Texas L. Rev. article draft when I get a chance and thinking some more about these distinctions.

And yes, "Judging Nudging" would have been a great title for your law review article. It also sets you up nicely for the inevitable sequel, "Nudging Judging," proposing methods of debiasing judges.

Posted by: Michelle Meyer | Aug 31, 2013 12:11:44 PM

First, as always, kudos to Schleicher, il migglior fabbro, who supplies the blog-post title (and yeah, probably shoulda-been article title).

Michelle, I categorize the cup size as default rather than "shove" because I think of the behavior that 's being regulated---the thing people are trying to do that the law affects---is the volume of soda consumed. You can't buy more soda at once, but you can consume as much as you want (at some nuisance cost). Which is why Thaler's tweet ("a ban is not a nudge," he says) doesn't make any sense to me. Look, a sticky default that deposits money into your IRA is a ban on spending that money...unless you opt out. I get why politically Thaler wants to distance himself, but he's relying on semantics, not real analysis.

In any event, one goal of my paper is to make clearer that these really are just semantic distinctions. Analytically, the soda "ban" shares essentially every feature of Thaler's other nudges -- it's asymmetric, opt-out costs are low, it's transferless, and unpriced.

Of course, if you think that the behavior people really want to engage in is ownership of a giant cup filled with soda, rather than consumption of a particular volume, then we have a different story. That's a "ban" in the sense that the opt-out costs are high (leave NYC) and it isn't obviously asymmetric.

In any event, looking forward to reading your work.

Posted by: BDG | Aug 30, 2013 3:12:10 PM

I don't get a title-related shout-out? I'm glad to see you used it here!

Posted by: D.Schleicher | Aug 30, 2013 2:45:57 PM

Hi Brian,

I don't think the following takes away anything from your argument, but I'm curious about your assumption that the Big Gulp ban, as I like to call it, is a nudge, rather than a form of regulation that has been inspired by behavioral economics/cognitive psychology. It's clever to characterize a 16 oz cup as a default size, as switching the default is a classic nudge. But nudges by definition (at least if you go by Thaler & Sunstein's definition--they didn't invent the concept, but they did, so far as I know, invent the term, and hence its definition) involve (1) removing no choices from the option set and (2) allow those who want to reverse the nudge (i.e., make a different choice than where the nudger is nudging you) to do so easily and cheaply. The proposed soda ban flatly prohibits 32 oz and larger sodas (in certain venues), thus removing an option from the choice set. True, you can go back a fill up your 16 oz cup twice, or buy two 16 oz sodas at once and attempt to juggle them or squeeze them into those tiny car cup holders. But depending on the situation (do I really want to leave the movie in the middle of the chase scene to refill my 16 oz. cup?), these workarounds may not be virtually costless, as nudges, per Thaler & Sunstein, require (e.g., checking a box to opt OUT of a 401(k)). In a short forthcoming piece, in fact, my co-author and I actually use the Big Gulp ban as an example of a shove to make the not-novel point that those who object to menu counts and other nudges as nannying should in fact welcome them, compared to shoves (although we end up objecting to menu counts, as they've been prematurely introduced in NYC and in Obamacare, because that nudge is not (yet) evidence-based). Thaler himself has rejected the characterization of the soda ban as a nudge: https://twitter.com/R_Thaler/status/208273339507150849

Like I said, I don't think this takes anything away from your broader points. But as what counts as a nudge has generated some discussion and controversy (my co-author and I spent quite a bit of time arguing over whether displaying calories counts is a nudge), I thought I'd raise it.

Posted by: Michelle Meyer | Aug 30, 2013 2:25:14 PM

This hypo remarkably ignores the positive economic effects (including Keynesian and multiplier effects) of employing presumably-otherwise unemployed professoinal wrestlers.

Posted by: Joseph Slater | Aug 30, 2013 2:03:15 PM

The comments to this entry are closed.