Counter Intuitive Nudges

Incentives don’t always work in the way we expect and neither do nudges. The British found that different slogans to encourage organ donation had markedly different effects.

 …the least successful message was: “Every day thousands of people who see this page decide to register [as an organ donor],” which ran alongside a picture of a group of smiling people.

….The most successful slogan was one which read: “If you needed an organ transplant, would you have one? If so, please help others.”

Ex-post it’s easy to come up with explanations for these differences but ex-ante these are difficult to predict. The unsuccessful slogan, for example, lets people know about a social norm; an approach that has been said to be very successful at reducing binge drinking and electricity consumption, so why didn’t it work for organ donation?

Here’s another peculiar nudge in NYC:

A counterintuitive “pilot program” aimed at reducing garbage in subway stations by removing trash cans appears, against all logic, to be working.

The idea of removing the trash cans came to the MTA in 2011 as a possible method of combating rats…But when the cans were removed from 10 stations, the agency found that not only did rat populations decrease, the amount of litter decreased, too.

My suspicion is that if this is true (and not random noise) then it’s a non-generalizable partial-equilibrium effect. If the trash cans have been removed only in some stations then people may be holding on to their trash knowing that they can dump it at the next station or in a trash can on the street. If you were to remove all or most of the cans this won’t happen.

Do you have other examples of counter-intuitive nudges? And what other explanations can you offer for the trash can effect?

Hat tip: Holman Jenkins.


As a former long-term resident of NYC, I can tell you that the explanation regarding the subway trash cans is simple: they are often not emptied frequently enough. There is a behavioral tipping point whereby most people will not simply litter by throwing trash on the ground or tracks. However, most people will dispose of trash in an already overfull trash can. Many New Yorkers can describe the act of balancing their trash on top of the pile, seeing that the can already has a halo around it, and often watching their trash eventually tumble down the side. However, by that point personal responsibility has given over the systemic responsibility: "it's not my fault the city failed to empty its trash can." It's the same effect as littering but by far has a different ethical connotation, even when alone.

The moral is, do not introduce a systemic solution if it cannot be supported. In lieu of one, people will problem-solve. Offer a spotty solution, and people will simply blame you for both your own and their own faults.

This is my initial reaction, this and AndrewL's scavengers.

Proper emptying may not be feasible, but I'd like to see it experimented with before saying there's a mystery.

Yep, main cause of litter is overflowing garbage cans...what everybody does, top them off, I do it too.

I wonder, do they make compacting garbage cans? How expensive would they be. Fundamentally, the overfilling seems a low density of typical trash issue.

There are some compacting garbage cans around NYC, but the existing ones are solar-powered for maximum Green.

Yes, I believe you are correct that overfilling is the problem. Is the "trash" they are counting all of the trash or just the trash that must be picked off the ground?

However, to rescue the hypothesis, what if people bring trash to the station knowing there is a can but not knowing how full it is. Does this invite trash? Do I eat the candy bar because I know there will be a can to eject the wrapper and refrain otherwise, ie trash generator as opposed to trash attractor.

If the can is full, are people more likely to throw trash on the ground or likely to carry it to the next can?

I observed different cultural norms in my life, and all I have to say is that places full of trashy people have lots of trash. It isn't about the can.

I went to a Halloween party where there were no trash cans but they were serving food and drink. Most people piled dirty plates and empty cups on tables in the corner. Were all the people trashy? No. I just think that there was no expectation of any other place to put it. Circumstances matter.

Expensive compacting cans that need power? Wouldn't increasing the number and size of cans be a better solution?

Unless you put some sort of video game on the compacting can that rewards trash feeding. Like the urinal with the painted fly.

So more cans would reduce trash on the floor too.

I still think the decision to banish cans is just kicking the can down the road.

Ted Mack, then Mayor of North Sydney, banished trash receptacles there in the late '80's. It worked, and they're still not there.

That the volume of trash in the venue decreases with the number of recepticles is not particularly controversial or insightful. The interesting question is where that trash goes.

If it is merely displaced, then there is a spillover/free rider problem. Maybe voters would prefer that trash be centrally collected at the downtown transit center than having it blow from their street corners onto their lawns. No one prefers a dirty transit center to a clean one, but what happens elsewhere is important to the decision.

Alternatively, the existence of the cans might actually generate more rubbish, and removing them eliminates some of the global volume of trash.

Admittedly this is a more intuitive nudge but McDonalds trash bins in Japan (where land is a very limited resource) are rather ingenious. They are divided between liquids, organic, plastic and paper, removing the normal option to simply tossing all your waste in one heap. To see this in practice visit:

....the amount of litter decreased, too.

Anecdotally, I recently returned from Tokyo and there I noticed very few public trash cans (subways or street level) and it is an extraordinarily clean city. With no trash on the ground there is the normative push to not be the first to do so; and with no trash cans I stored the litter rather than just holding on to it. Perhaps having accessible cans increases the odds trash would fall out or overflow to the ground which would impact the normative effect?

In Tokyo there are a lot of people employed to keep the streets clean, so that's going to help. And bins are quite rare. The vast majority of people don't litter, but that's true in many cities in the world and I couldn't say whether or not Japanese are exceptional. I do know a lot of cigarette butts end up on the ground, but they are soon swept up. Interestingly, I've been told the laws against littering are very weak (or perhaps just poorly enforced).

Rubbish bins are very hard to find in Tokyo, I think for two reasons: (i) It is very rare for people to eat/drink while walking along the street (you should sit down and concentrate/appreciate what you are consuming), so very little rubbish is generated on the streets. When people buy from a vending machines they tend to stand next to the machine while they drink it, and there are always small bins for the cans/bottles there. (ii) Home recycling, at least in Tokyo, is very strict - plastic PET bottles one day, other (clean) plastic another day, cardboard separate, glass and cans separate, general rubbish perhaps twice a week, larger items 1 day a month etc. Its simply impossible to have rubbish bins on the street that can deal with this complexity, so you have to take it home and sort it out yourself in the proper way.

I was born and raised in Japan. Trash cans used to be around till the Aumu Shinrikyo's sarin gas attacks in 1995. Authorities virtually took away all the trash cans in Tokyo Metropolitan area and we have stayed that way since then. Not sure if this has positive effects on our behavior.

This is the kind of ground truth that grinds speculative theory to dust.

Why are there almost no rubbish bins outside of Tokyo as well?

On the trash can effect, is it possible that by reducing to zero the visible trash cans, the economics of the free-riders problem changes? Suppose you are someone inclined to disregard the norm of "do not litter" and part of your justification is that "someone else will pick it up." When there is a trash can 50 feet away, you might assume that some person who obeys the norms of "do not litter" will pick up your garbage, and at nominal expense of their time, mitigate the cost of your action. Littering in a "no-trash-can" setting raises the implied cost of action by presumed good samaritans. Another potential explanation is that the lack of trash cans lessens the incentive for rule-breakers to litter because there is no visible example of the encouraged behavior to defy. Someone with a "break-the-rules" mentality might not be reminded to do so when an example of the rule (trash going into a bin) is absent.

Trash cans are not only a food source for rats, but also a draw for the city's homeless who scavange the cans for bottle/can deposits and discarded food. No trash cans means nothing for the homeless to overturn.

People will throw trash out onto the tracks, but rats on the tracks are less alarming than rats on the platform. Homeless people will also go down onto the tracks to collect bottles, and the track drainage system generally pulls all the trash to the drains where vaccuum trains can collect it all at one point (out of sight).


Some trash gets knocked out of the cans by homeless scavenging for bottles, causing normative effects etc.

too much trash in the US, period.

Having trash cans makes you take more paper napkins, less concerned about where you will dispose all the stuff you're carrying. Without trash cans, typically most people will take their trash with them and throw them at home/work. Who's going to bother seeking out stations because it has a trash can?

The homeless will move to the landfills to scavenge.

I never took napkins *because* there were trash cans. I only took napkins *becuase* the pizza sauce would drip onto my hands get get on my face. I don't throw trash out onto the street either.

and no. The homeless will not move to the landfills to scavange. Bottles and cans generally don't make it to landfills and discarded food is inedible by the time it makes it there as well. Also landfills are very difficult to get to and have permiter fences.

Wasn't being serious.

He said he wasnt serious, but there is at least some potential for an as if proposition. People often respond to incentives they are only subconciously aware of. For example, I grab a coffee before I get on the train. When I get downtown, there is a trash recepticle awaiting my arrival. If the city removed those convenient cans, I might either carry the cup to my office, bring my coffee in a thermos, or buy coffee when I arrive downtown. If I prefer my hands free downtown, say to open an umbrella or buy a newspaper, my subconcious desire to rid myself of the cup might grow.

Same may be true with taking napkins. All else equal, people (as a group) might take more if they have a convenient place to dispose of them.

On the London Underground there are no rubbish bins (after use by terrorists for bombs years ago), there is also very little litter.

Conscientious Londoners? Effective litter patrols, who knows?

DC's Metro tends to be fantastically clean - I think arresting people for eating in the system has an effect, a policy that has been in place since the first Metro stations opened.

From a ten year old article (I doubt the rules have changed that much - nor have the cameras become less present) -

'Stephanie Willett is a 45-year-old scientist for the U.S. Environmental Protection Agency from Bowie whose skirmishes with the law had largely been limited to a couple of speeding tickets.

Until she was caught chewing inside a Metro station.

About 6:30 p.m. July 16, Willett was eating a PayDay candy bar while riding the escalator from 11th Street NW into the Metro Center Station. Metro Transit Police Officer Cherrail Curry-Hagler was riding up.

The police officer warned Willett to finish the candy before entering the station because eating or drinking in the Metro system is illegal.

Willett nodded, kept chewing the peanut-and-caramel bar and stuffed the last bit into her mouth before throwing the wrapper into the trash can near the station manager's kiosk, according to both Willett and Curry-Hagler.

Curry-Hagler turned around and followed Willett into the station. Moments after making a remark to the officer, Willett said, she was searched, handcuffed and arrested for chewing the last bite of her candy bar after she passed through the fare gates. She was released several hours later after paying a $10 fine, pending a hearing.

"We've been doing our best to crack down on people who are consuming food and beverages in our stations because we get so many complaints about it," said Lisa Farbstein, a Metro spokeswoman. "In this instance, the woman was given a warning, which she ignored, and she jammed the rest of the candy bar into her mouth and continued to chew."'

That's what makes mass transit so great compared to cars. Car makers still think cup holders are a good idea!

And the trash can door is electrically operated so you can toss out your drink container easily into the trash gutter.....

No trash cans on the roads in NH do not reduce the littering.

But perhaps that's why NH has Litter For Free or Die or something like that on all its licence plates. Libertarians aren't going to be denied the liberty of dumping their trash on my yard and that of my neighbors.

Not surprised you celebrate the arrest of a person for disposing of food into their mouth. Lays bare your preference for oppressive totalitarianism.

Life in prison for 11 items in the express checkout lane?

I'm a nothern Virginia native whose very first experience on Metro was during the days it was free to ride on the first two open stations. To say I'm 'celebrating' Metro's extraordinarily strict approach seems strange - but that was just the first incident that came up. One that many people here could not easily dismiss, unlike some of these -

'A 12-year-old girl found out the hard way that there’s no snacking allowed in the Washington, D.C. subway.

Seventh-grader Ansche Hedgepeth was handcuffed, booked and fingerprinted for eating French fries in a northwest Washington subway station.

Ansche told police she knew she wasn’t supposed to eat in the station but didn’t think she would get arrested. Ansche’s mother Tracey Hedgepeth, who has written a complaint letter to the Metro Transit Police Department, said police went too far.

“I can’t believe there isn’t a better way to teach kids a lesson,” she said. “The police treated her like a criminal.”

But Metro Transit Police Chief Barry J. McDevitt is unapologetic about the girl’s arrest last month and others like it.

“We really do believe in zero tolerance,” he said.

Commuter complaints about unlawful eating on Metro cars and in stations led McDevitt to mount an undercover crackdown on violators. A dozen plainclothes officers cited or arrested 35 people, 13 of them juveniles. Only one adult was arrested.


In 1987, Iran-contra figure Fawn Hall was given a $10 fine for eating a banana in the Metro Center subway station. She was not arrested.'

Trying to explain DC Metro to outsiders is always interesting - it is like a taste of Singapore's freedom in the midst of the anarchy that rules in such places as Philly, NYC, or on the MTA. (And 'Singapore's freedom' is a sarcastic oxymoron based on the reader actually knowing the sorts of rules that are enforced in public settings in Singapore - such as the forbidding of gum chewing - It is not a celebration of 'Singapore's freedom.')

Especially trying to explain Metro to people who feel that pointing out out just how strict DC Metro is in terms of using police and cameras to keep people from eating a banana is 'celebrating' one of the more prominent examples in experiencing a police state right in our nation's capital. For four decades.

You dug yourself deeper, not out.

All of these are frivolous reasons for a man with a gun to impose a fine. There is a huge difference between eating a wrapped candy bar and eating a takeout container of curry chicken.

Zero tolerance policies are zero intelligence policies. Fining people for potential harm rather than harm is oppressive unless the potential harm is likely, intolerable, and reasonably forseeable. That said, there are limited circumstances where a bright line criteria is necessary to delineate an unambiguously bad state from a good one. This usually involves difficulty defining the proscribed behavior and ensures equal understanding of the law.

Yeah, the PATH trains don't have any trash cans and they're usually way cleaner than the MTA ones. I usually make an extra effort to hold on to my trash and throw it away when I'm off the train.

The trash method doesn't work in Philadelphia. They have almost no trash cans on the street and there is more trash on the ground in that city than I ever ever seen anywhere else.I think the problem with overflowing trash cans is the likely explanation. On the DC mall whenever the trash cans fill up, people just keep piling it on or throwing it next to the can -- they probably sense neglect by the city and its workers ("if they cant bother to empty the trash cans often enough, then why should I bother looking for somewhere else to throw my trash away?")

I think it only works if you do it selectively at some locations.

Obvious question: Why those locations?

So Tokyo is clean but NYC and Philly are dirty. I wonder what Steve Sailer would say about that.

What does NY and Philly have that Japan does not?

Guns, obviously.

This sounds like it is related to the crime reduction philosophy not to tolerate graffiti and such. People are reluctant to be first to trash an area that appears looked after. Overflowing trash cans send the opposite message.

Perhaps not exactly counterintuitive nudging, but this post reminded me of the unofficial "alcohol" policy at my college campus. It was extremely lax, and the college even subsidized kegs (for people of legal drinking age of course) for dorm parties and other events. Thinking back, it was probably harder to find water than alcohol if you were out and about on campus on a Friday or Saturday night. We always speculated that this was a better alternative to having young college kids going off campus, likely resulting in drinking and driving.

Wisconsin-Madison by any chance?

I won't identify the specific college, but will say its a top 10 liberal arts college (under U.S. News & World rankings).

National Guard armories used to have open bars as a way to keep soldiers working late and staying overnight as opposed to going out to a bar, driving drunk, and getting into bar fights.

This was thrown into the rubbish can when the army cracked down on alcoholism and drunk driving. Effective or not, the bars detracted from the overall message.

For years I thought "fantasy football" was, due to its name, not the sort of thing people who actually liked real-world football would be attracted to. Then I figured out what it actually is, basically a fairly normal variety of betting schemes.

A curious coincidence..

Here in Australia, the national parks service reduced littering any to a very low level in many picnic grounds by removing all rubbish bins.

The PATH train, a quasi-subway between Manhattan and the nearer parts of New Jersey, has never had trash cans, and it is far cleaner than the NYC Metro. There is also no eating allowed (though this is enforced about as tenuously as speed limits). People simply don't bring their trash into the stations.

A confounding factor is that the NYC Metro travels into all of the less savory neighborhoods of New York. The PATH mostly serves yuppie areas.

It may be that a lack of garbage cans discourages people from bringing/producing refuse on the trains. It may be that the PATH's clientele is less likely to litter. Speaking as someone who used the PATH for 3 years, I learned not to carry garbage with me that I didn't want to carry until the ride was over.

I can believe the trash can one. In a lot of places, including the Toronto subway system, garbage bins were removed for a while out of fear of terrorism. I think trash went down, though nobody was measuring.

Whether litter increases or decreases upon removing the bins might depend on the social norms (and laws) of the place in question.

Due to my fellow apartment dwellers leaving unbelievable amounts of festering food waste in the garbage cans of our garage, the management removed them. As a strong believer in drive-through food (although I eat all of it, thus my waste is mainly paper and plastic), I now must put my trash in cans at work or at a gas station. The problem has been moved, but not quite "solved".

Unexpected results in Drachten, a Dutch town:

They removed most of the traffic lights and signs, and traffic improved and there were far fewer accidents.

A frequent libertarian argument. Having lived in Eastern Europe where there were no signs or lights, I observed high accident rates.

Clearly there is a threshhold for the benefit of public goods exceeding their costs, and below that threshhold the project is inadmissible. Remove the public good and things improve.

In densely populated areas, no signs or signals is catasteophic. There is usually a good reason for putting them in in the first instance. But even without signs or lights, there can still be governing laws on who must yield, such as yielding to the car on your right, and left turning traffic must yield. Here, the public goods (laws and liability) are invisible.

The trash argument is sophistry. Take away trash cans here, and the trash goes elsewhere. From the point of view of the metro station, this is good, but from the point of view of the new recipients, it is bad. Clearly there is a fallacy of composition.

On the other hand, trash cans - like new bridges - might be congestion attracting. Is it possible that the existence of trash cans CAUSES more trash to be generated? I doubt it.

OT: Did you see the article about the Paradox of Choice? I can't find the link right now.

In my view, the Paradox of Choice is hogwash. Choice is unambiguously welfare improving, * ceteris paribus. *

But ceteris usually isn't paribus. More choices create more information to collect and evaluate. The increase in information increases analysis costs. The information is also collected in a serial, non-random fashion. So if there is a paradox, it is a Paradox of Information where less information MIGHT result in you being unambiguously worse off. The first question is to determine whether there is a contrived example to prove existence of the paradox. Then to determine the conditions and likelihood of those conditions. For example, it might take a rapidly rising information cost relative to the marginal benefit of the information.

But you may just end up at a constrained maximum which is not really paradoxical.

I think a lot of confusion around the paradox of choice arises because it a) assumes we want to maximise, not satisfice; b) because it assumes that our decision-making is entirely individualistic and c) it assumes complete trust.

We like owning things which lots of other people own - and for perfectly sensible reasons.

If there were only three kinds of pension you could buy, we would know that if one of thethe three turned out to be a bit of a rip-off then someone else - someone who checks these things more assiduously than I do - would probably notice. They would make a fuss about it. We could petition the government for compensation.

As things stand, my pension is effectively unique. I have no confidence in it, or in the people managing it, because it is not subject to collective scrutiny. I cannot discuss it with friends who know more about pensions than I do without going into an insane amount of tedious detail.

In markets where there are three or four dominant players, people who wish to satisfice (ie most sensible people, most of the time) buy one of those three or four leading brands. If you have a market like the pensions market which is just a hideous mess, this simple default option does not exist.

In Japan, the lack of trash cans is annoying but most Japanese just assume they have to carry their trash with them for a while.

The situation isn't hopeless though since there are usually recycling canisters in vending machines or next to them. And convenience stores often have trash cans, recycling either inside or next to the store.

Once you know an area it isn't too bad, but I can see how tourists would be annoyed..

Interesting insights into trash behaviour around the world. Here in South Africa, the amounts of litter on the roads and public streets is substantial, regardless of whether there is a trashcan nearby or not. The seemingly accepted way to get rid of your trash is to simply throw it aside.

I've driven behind taxis (here in SA, a Taxi is generally a 10-12 seater Toyota Quantum) where KFC boxes with their used contents have been casually discarded out the window of the moving, or not moving at a traffic light, vehicle. Tins, bottles, basically anything which has finished being used or consumed is discarded without a thought.

Perhaps the "social norm nudge" didn't work exactly because it didn't correctly quantify the number of people signing up. Huge numbers get thrown around in policy discussions all the time -- "millions, billions, trillions." If I hear "thousands," maybe I actually think that it means a very small portion of the relevant population is signing up. Lets say "thousands" means 2,000, and the population of Birtain is ~ 63 million. Then 2k/63million = 0.003% of the population is signing up. Not enough to convince me a social norm is happening. Even if that number is in fact daily (instead of the highest daily number so far, which I often suspect of such statements), then we still get roughly less than 1 percent: 2k/63million*300 = 0.9%.

I say this not because I think everyone does careful calculations, but because when I read the first slogan, my first gut reaction was "thousands a day really doesn't sound like a lot of people are doing this." I.e. not really a social norm. A small bit of calculations shows that to be true.

If they want to communicate the existence of a social norm, it seems that they need to think about a different way to phrase it.

*Britain, whoops. No edit button, huh?

Also, the 'people who see this page' bit seems to confuse the message surely it would be better to focus on the total number of people donating than just those who see the advert. Or it doesn't get over that there is a problem, because loads of people donate already, where the second slogan neatly highlights the free rider problem.

Some literature (Goldstein et al. 2008) could suggest that saying those who see the page sign-up could be successful. However, what prompts an individual to adhere to a norm is multifaceted and is not as straight forward as X works every time.

The article is woefully lacking in particular what "successful" means. How can you begin to gauge the messages without knowing more details of the actual results?

There are only three things you need to know to understand to practise nudging.

1) Understand that small interventions can have surprisingly large effects. Therefore you should probably test small things more than you do at present.

2) Understand that many of our decisions are influenced by unconscious mental processes which are opaque to introspection. Therefore the way "we think we decide" is not the way we really decide - but may be a post-rationalisation. Hence you should also test interventions which may seem "counterintuitive".

3) In many cases the unconscious instincts driving our decisions have a hidden, second-order intelligence to them. They are only "irrational" when rationality is defined in very narrow terms.

Comments for this post are closed