The impossibility of a behavioural welfare economics

Anyone who has taken Econ 1000 learns that restricting soda consumption creates a deadweight loss:

At the price shown, consumers would like to buy Q₂ units of soda. A ban on extra-large sodas restricts their consumption to Q₁ units. At point Q₁, consumers would like to buy larger sodas, and firms would like to supply larger sodas, but they are not allowed to. The resulting loss in consumer surplus - the benefits consumers enjoy from consumption - is called a "deadweight loss", and represented by the red triangle above.

Analysis of deadweight loss is a very simple type of "welfare economics" - the study of how projects and policies affect people's well-being, or welfare. Implicit in standard welfare economics is the assumption that people know what is in their own self-interest, and act upon it. If people drink extra-large sodas, it must be because the enjoyment they get from drinking sodas outweighs the costs, for example, weight gain and increased risk of diabetes. Unless soda drinking harms other people, any analysis that begins with the premise that people's actions are based on rational self-interest must inevitably reach the conclusion that soda bans reduce welfare.

Behavioural economics, however, calls the basic premise of rational, self-interested choice into question. As Richard Thaler and Cass Sunstein argue in their book Nudge, people often don't make the effort of thinking through their decisions. Instead, they resort to "mindless choosing." For example, one experiment gave participants bags of five-day old stale popcorn, and then showed them a movie. Even though though the popcorn was "like eating Styrofoam packing peanuts", people ate it anyways - and the people with large bags ate more, even though they didn't particularly like the popcorn in the first place.

If behavioural economists are right, perhaps extra-large sodas don't really give people more happiness than slightly smaller ones. Consumer demand, seen through the lens of behavioural economics, looks more like this:

People are not totally devoid of sense; they cannot be persuaded to buy just anything. But, at any given price, a consumer might be persuaded by buy any quantity between Q₁ and Q₂, depending upon how the consumption choice is framed, or marketed. Any quantity within this range could be the observed outcome of a "rational choice" at the price shown. In this world, a soda ban simply nudges the consumer from point Q₂ - grabbing an extra-large soda - to Q₁ - grabbing a large soda. 

Yet this begs the question: how does the behavioural economist know which is better, Q₁ or Q₂? Sometimes, as in the case of sodas, the behavioural economist urges consumers to consume less. But when it comes to, say, savings, behavioural economists work in the other direction, creating nudges that get people to save more. But how does a behavioural economist know which direction the consumer should be nudged in?

One possibility is to replace the old tools of welfare economics - consumer and producer surplus - with health and happiness. Does drinking less soda increase health? Are people who drink smaller sodas happier, all else being equal, than people who drink extra-large ones? If so, then the ban may be good policy.

Health and happiness are not the only metrics by which behaviouralists judge choices. Thaler and Sunstein, for example, argue that each of us has an internal planner, a sensible person who would like to make rational choices, but who is defeated by our internal Homer Simpson. In this view, the behaviouralist is just helping us do what our inner planner would like. Yet how do we know what people's inner planners really want?

While I am sympathetic to behavioural economics, I worry about the consequences of viewing all policies through the lens of health. Some of the best things in life are also somewhat dangerous - like swimming across a lake with only the moon and stars for company. At the same time, I want to live in a world where I'm nudged towards healthy choices - where it's easy to get around by bicycle, for example, and the university cafeteria serves reasonably nutritious meals.

Welfare economics teaches us to respect people's choices. Behavioural economics urges us to protect people from making dumb choices that they'll regret later. I don't know of any way of resolving these two views.

Comments

[L_perrault]

Marc Fleurbaey has at least two papers that I know off on this issue. The first one that comes to mind is the one where people make bad choices and regret those choices afterwards. The paper is called "Freedom with forgiveness" (http://ppe.sagepub.com/content/4/1/29.abstract).

I saw a new one presented in Marseille that is more in line with social choice and standard welfare economics called "Behavioral Fair Social Choice" (http://ideas.repec.org/p/hka/wpaper/2012-012.html). It applies the Fair social choice developped by Fleurbaey and Maniquet to the issue of behavioral economics. It is pretty interesting. In this case what the planner cares about his fairness and resource equality and not so much on happiness and health per se.

[ianlee]

"But how does a behavioural economist know which direction the consumer should be nudged in?"

Frances - you have asked what I think is the central and most deadly question that must be asked of behav econ.

Are they not sneaking "god" in the back door, a deus ex machina - an unmoved mover - that presupposes what consumers, economic actors should do.

Sunstein's "inner planner" was first introduced in Plato's Republic as the "inner voice" - Socrates called it his daimon. Christians later called it the "guardian angel".

Plato at least came out of the closet when Plato prescribed guardians of gold, silver and bronze in support of the philosopher king. Ditto Marx who determined the grand decisions to guide and constrain the little decisions would be made by a vanguard of the proletariat.

When I read pieces on behavioral econ, it reminds me of my days in Sunday school far away and long ago with a very clear and distinct morality tale of good and bad.

Perhaps behavioral econ is a new version of metaphysics or possibly theology?

Calling Philosopher Rowe?

[Wonks Anonymous]

Economist Robin Hanson advocates seeking peace between your inner planner & Homer Simpson, rather than seeking to help one side win the internal war:
http://www.overcomingbias.com/2009/05/prefer-peace.html

[Dan F]

This analysis (and most economic theory, now that I think about it) ignores the very powerful impact on demand of Marketing and Advertising. Demand is not something that occurs suddenly in the mind of a consumer. Products are not normally created to fill demand, as much as we would like to believe that this is the case. More often, demand is created by marketing/advertising to consumers. Of course, as any individual and he will insist that he is immune to the effects of marketing, but any marketer knows that on the whole a campaign can have a significant impact on sales.
If the extra large soda had never existed, how many consumers would get to the bottom of their large soda and wish that there was a larger size? Decisions by consumers are usually framed by the available choices, although of course they become very upset when you take away a choice that they previously had.
Economic theory depends on perfect information, and marketing is opposite of information, since in most cases it is designed to create demand for a specific product and push the consumer towards that choice. Sometimes it does fall upon government to work on behalf of consumers to balance the effect of marketing on consumer demand.

[Squarely Rooted]

Here's an idea for some synthesis - people's choices are rational given the existing set of incentives (which includes social pressures and subconcious "behavioral" incentives) and their discount rate. But there is no "neutral" "pure" "natural" or "free" incentive set, so if you see the existing incentive set creating bad outcomes (morbid obesity, diabetes) then just change the incentive set.

The problem in practice is that in lots of cases this is probably really, really hard. Though I think we oft neglect the full implications of the theory of bounded rationality - that individual "computation power" and time is limited, so even if people are rational they only have so much bandwith to decide everything rationally. Might just be that, given popcorn and soda in the movies is such a "hardwired" choice, that the Pigovian solution of taxing the beverage may not work well and simply reducing choice is a better option. But who knows?

Also, I really like the "fuzzy" demand curve.

[Frances Woolley]

Dan F: "This analysis (and most economic theory, now that I think about it) ignores the very powerful impact on demand of Marketing and Advertising"

Darn. I thought what I was doing with that red demand curve is showing how demand is not fixed, as intermediate micro would tell us, but rather can be moved around within a range by framing - and what is marketing and advertising other than a way of framing choices?

Squarely rooted - thanks for the like. I'm trying to explain behavioural econ to my micro-for-non-majors class today, and this seemed to me to be the best way to do it.

L perrault, Wonks - thanks for the links.

Ian - On the idea that behavioural econ is a new theology - IIRC, people who believe in God are, on average, happier than non-believers (as Republicans are, on average, happier than Democrats). Yes, perhaps it is a little bit preachy - but people seem to benefit from a bit of preaching every now and then.

[jt]

I agree we should in general stay out of the refrigerators of the nation, but a subtle aspect, like second-hand smoking, is to what extent should society pay for the sins of others. I don't mind you skinny dipping in your favorite lake at midnight (mmmm, I hope your students aren't imagining it during your next class!) because most likely if something goes wrong it won't cost me very much.

[L_perrault]

Thinking about it, the paper more related to your interrogation is probably this one

Bernheim, D., and A. Rangel. 2009. "Beyond revealed preference: choice theoretic
foundations for behavioral welfare economics". Quarterly Journal of Economics 124, no.
1: 51-104.

If I remember correctly, the idea is to use the part of preferences where there are no ambiguities. From this they get back versions of compensating variation (equivalent variation), consumer surplus etc.. It is pretty clever.

[Sandwichman]

What is a "unit" of soda? Consumers can't buy whatever quantity of soda they want. They only get to choose between (e.g.) "Regular", "Large" and "Humungous." You are framing a false dichotomy between a hypothetically "free" market of exchange and state regulation. The market is already always and everywhere "regulated". Banning extra large sodas simply shifts the locus -- and presumably the objective -- of that regulation. Why do economists insist on perpetually inventing false problems at the same time as they studiously ignore real ones? You should be ashamed.

[Ryna]

There is only one possible form of reconciliation between respect for the voluntary choices of others and an urge to protect people from making choices we perceive to be wrong: persuasion. You know, sit down and talk to people and win them over to your point of view.

[ianlee]

Sandwichman - I do not get your point at all. I can buy just about any consumer product in bulk at e.g. Costco or any big box. Or I can buy a juice or a water or a soda or a milkshake or a cappucino and on and on anywhere at a anytime. We face a staggering array of product choices and substitutes that is almost bewildering at an astonishing number of stores.

Perhaps you were not around in the 1960s when we had 2 Cdn TV and 3 US TV channels, NO big box stores, no Macs Milk, no alcohol sold anywhere on a Sunday, banks open from 10 am to 3 pm M - F, and we often faced two choices of everything in at most 2 or 3 colours. We were only a few steps from Henry Ford, "you can have it in any colour you want so long as it is black".

I thought Frances' post was superb in capturing the incommensurates between the two approaches. I also thought Frances was brutally honest in noting how behav econ seems to slip in a predetermined set of choices or assumptions - like a magician with a card trick - that I characterized as "god" or metaphysics through the back door.

Given the intellectual popularity of behav econ today,and given that no one seems to want to challenge its assumptions or findings, I thought Frances provided an important contribution by initiating this debate.

Kudos to Frances' honesty and courage - for "running against the wind".

[Shangwen]

The puzzling thing about some of the proposed BE interventions, like banning soda and demonizing salt, is that they often seem to confuse the general causes of a problem with specific behaviors that drive the marginal cost. The growing economic burden of chronic disease is largely driven not by overall prevalence (the N of afflicted) but by low treatment compliance rates. For some chronic illnesses, treatment adherence can be as low as 50%. We are not specifically worse by gaining 100 new people with Type II Diabetes, we are made worse off by 50 people who won't manage their symptoms well, take their meds, eat better, etc. Same for many other conditions. So the problem is not so much sugar, as it is not dealing with your diabetes. If you manage your disease well, you are taking on a relatively small number of private costs for a huge gain.

The further problem is that we just assume the misbehaving 50% are in error. But maybe they aren't all. If you are 60 and already have foreshortened life expectancy, one more visit to the buffet that increases your chance of amputation by 3% 10 years out is not so costly. Why should you care? The fact that it may impose an extra burden on the public health care system should not guide your private choices.

My point is not to discuss diabetes, but to point out that the policy targets of BE seem so off-the-mark. A ban on sugary drinks is a very different policy from creating incentives for people to manage their illness better. But the excitement is all about banning stuff.

[whitfit]

There are a couple of points I want to address:

First, I was under the impression that behavioral economics was about understanding how people make decisions, and contrasting the ways in which people don't behave the way that homo economicus behaves in neoclassical economics, but instead have persistent biases, misapprehensions of facts etc... I guess in constructing "nudges", and more specifically bans, it seems to me that a behavioral approach would have to demonstrate that there is a persistent set of misapprehensions or biases that creates problems in decision making, which I suppose still requires that we have a metric about what a better choice would be.

Second, a ban isn't a nudge. A nudge sets out the default state, but allows people to switch to some other choice. Frances' examples of making bicycling easier might be a better example of a nudge. I think this is an important feature of some BE interventions, because it allows the individual to make a different choice they think is better, but forcing them to think through the issues in order to move from a default position to one they might prefer.

Thirdly, Ryan (I think it isn't Ryna, given the libertarian content!) says that we should sit people down and convince them, not constrain their choices. Well, from the perspective of choices and the law (which is Sunstein's background), studying the law you realize that many things, from property rights to contract law is the result of arbitrary choices that set out a default structure of who owns what and how transactions occur. To pretend that introducing new defaults is some kind of new constraint that is the beginning of the constraint on liberty denies the existence of the history of law and institutions.

Finally, I am not sure that marketing only works by moving people across the fat part of the blurry curve. I like the concept of a blurry curve, but it seems to me that if the curve describes preferences, only marketing that makes a product more salient is likely to be that kind of move, whereas marketing that creates desire would move the curve in the classic way.

[Ryan]

whitfit - Good catch. It is indeed I, not my evil doppelganger Ryna! As to your point, I am not sure what you're getting at, or how on Earth my comment could in any way be conceived as a denial of the existence of the history of law and institutions. Bit of a stretch, no?

[Sandwichman]

"I was under the impression that behavioral economics was about understanding how people make decisions, and contrasting the ways in which people don't behave the way that homo economicus behaves in neoclassical economics..."

Funny you should mention M. Homo, whitfit! The pseudo-Latin is obviously intended to signify a pseudo-anthropology. Unfortunately for economics, there is also a real anthropology against which economists' pretensions of anthropological (and economic) observation might be tested. It so happens there is a case study in the works that demonstrates the monumental, colossal lack of intellectual curiosity that prevails in economics. The second most cited article in anthropology between 1966 and 1982 -- "frequently cited as an example of how not to go about understanding peasant societies." -- just happened to examine "one of the best-known fallacies in economics" that economists accept unquestioningly as a truism. See "How not to go about understanding peasant societies..." (or any society for that matter): http://ecologicalheadstand.blogspot.ca/2012/09/how-not-to-go-about-understanding.html

[Sandwichman]

Damn that missing close italics mark up!

[Fixed - SG]

[Sandwichman]

I'll try again:

"I was under the impression that behavioral economics was about understanding how people make decisions, and contrasting the ways in which people don't behave the way that homo economicus behaves in neoclassical economics..."

Funny you should mention M. Homo, whitfit! The pseudo-Latin is obviously intended to signify a pseudo-anthropology. Unfortunately for economics, there is also a real anthropology against which economists' pretensions of anthropological (and economic) observation might be tested. It so happens there is a case study in the works that demonstrates the monumental, colossal lack of intellectual curiosity that prevails in economics. The second most cited article in anthropology between 1966 and 1982 -- "frequently cited as an example of how not to go about understanding peasant societies." -- just happened to examine "one of the best-known fallacies in economics" that economists accept unquestioningly as a truism. See "How not to go about understanding peasant societies..." (or any society for that matter): http://ecologicalheadstand.blogspot.ca/2012/09/how-not-to-go-about-understanding.html

[whitfit]

Ryan - good point - I think I read more into it than was there.

You did say: "There is only one possible form of reconciliation between respect for the voluntary choices of others and an urge to protect people from making choices we perceive to be wrong: persuasion. You know, sit down and talk to people and win them over to your point of view."

I read that as nudges are not good - that one should only use persuasion to convince people to change their point of view. But, I think you were saying less than that - that there should be coercion vs. freedom to choose. I was saying that nudges can be a way of consciously designing institutions and laws to make a default choice that recognizes the planning limits of individuals, which is admittedly not a response to your point.

[W. Peden]

Dan F,

"Sometimes it does fall upon government to work on behalf of consumers to balance the effect of marketing on consumer demand."

(1) Does the same principle apply to political decisions in a democracy? For example, extremist fascist/communist/racist political parties exist, can conduct marketing, and persuade people to vote for them. Should government work on behalf of voters to protect them against these kind of parties? If so, one starts down a very slippery slope. (The same slope obviously exists for moral choices as well e.g. as regards things like homosexuality or religion.)

(2) Doctors work on our behalf. Cleaners work on our behalf. Policemen work on our behalf. Describing the constraint of choice by supposedly enlightened elites as "working on our behalf" seems rather strange: in this role, government is neither mandated by choice, nor providing the conditions in which free choice can be exercised (as police do).

(3) No-one denies that adverts get things done, otherwise businesses, charities* etc. wouldn't spend money on them. Advertising is powerful. However, there are different kinds of power. In particular, in a liberal democracy, we close off the slopes described in (1) by allowing [i]persuasive power[/i], provided no deception is involved. So if I want you to do X and I persuade you to do X, then I have exercised an acceptable kind of power. On the other hand, [i]coercive power[/i] is when I get you to do X by removing some of your existing choices.

An example of persuasive power is if I try to convince you to join the NDP. An example of coercive power is if I threaten to shoot you if you don't join the NDP, removing your existing option to not join the NDP & not get shot.

Advertising, provided it is not fradulent, seems to me to be the right kind of power. Now, one might argue that advertising is the wrong kind of persuasive power, because much of advertising seems irrational. To take a very simple example, why should a motel have a flashing sign? Why shouldn't it just say "Motel"? Israel Kirzner, I think, refutes this argument: the non-informative part of advertising draws our attention to products; it doesn't make us want them (I have seen many a distracting ad for iPhones on the internet and I still don't want one) but it does draw our attention. In a dark stormy night on the road, that flashing light is very useful.

(4) Re: creating desires, I think one has to be careful here. Adverts may make us discover desires. However, is that the same as creating them? I don't think so: people desired not to suffering from infections before pencillin, but adverts for penicillin made them discover an instrumental desire for penicillin because it was a means of avoiding some of this suffering.

* Would government act on our behalf to discourage us from giving to charity, confused as we presumably are by their advertising? If not, then it seems to me that what is being proposed is not "balancing" of the effects of advertising, but rather simple old fashioned state coercion to get us to make the "right" choices, like burning heretical books in the Middle Ages or jamming the BBC in the USSR.

[Bob Smith]

"More often, demand is created by marketing/advertising to consumers."

Sure, but does that matter? I don't recall my economics courses having a great deal to say about the source of people's preferences. I think conventional economics just assumes they exist and have certain (perhaps debatable) properties. Who cases whether they arise from genetic hardwiring, parental influence, education, culture neighbours or advertising? All that matters in the conventional framework is that they exist and have the required properties.

[Victor]

Richard Thaler and Cass Sunstein in their book nudge never said anything about banning peoples choice, what they endorses is "libertarian paternalism", the basic idea of "libertarian paternalism" is that private and public institutions might nudge people in directions that will make their lives go better, without eliminating freedom of choice. Quoting Sunstein "A core example of libertarian paternalism is Thaler's Save More Tomorrow plan, by which workers can sign up to devote some of their future wage increases to savings. Another example is the automatic enrollment plan, by which workers are automatically enrolled in a savings plan, but can opt out with no trouble and at no expense if they choose to do so. "

[Bob Smith]

"People are not totally devoid of sense; they cannot be persuaded to buy just anything. But, at any given price, a consumer might be persuaded by buy any quantity between Q1 and Q2, depending upon how the consumption choice is framed, or marketed. Any quantity within this range could be the observed outcome of a "rational choice" at the price shown. In this world, a soda ban simply nudges the consumer from point Q2 - grabbing an extra-large soda - to Q1 - grabbing a large soda"

I like that analysis, and I think it makes sense. Mind you, I'm not sure soda is the great example, because I suspect the big success of soda producers is selling much larger sizes at slightly higher prices, which suggests that they believe that consumers are sensitive to price and are looking for "bargains".

But let me ask you this, if consumers can be pursuaded to buy any quantity between Q1 and Q2, they can also be pursuaded to buy any Q between Q1 and Q2 at a range of prices higher than P (I think, but I might be wrong). Retailers might even prefer to end up selling at Q1/P2 - that's the logic behind "premium" branding, sell the same thing at a higher price. The conventional story is that competition drives retailers/producers to bring prices back in line with marginal cost. But how does competition work if consumers are indifferent (at least within a range) to price? What's the competitive process that gets soda retailers/producers to end up selling at Q2?

The only theory that I wondered about is some sort of addiction/habit theory. That the P/Q2 bundle might not be optimal in a one-off game, but if people develop brand loyalties or preferences over time as their consumption increases, producers/retailers might be keen to encourage long-term consumption of their products (think of this as a rational addiction model, from the perspective of the drug dealer)

[Bob Smith]

"(think of this as a rational addiction model, from the perspective of the drug dealer)"

It's at least interesting that two core components of soda pop, sugar (or sugar-like substances of some sort, i.e., corn syrup) and caffeine are highly addictive.

[Frances Woolley]

Victor: "what they endorses is "libertarian paternalism""

Thaler and Sunstein seem to take the approach that if they are really up-front about the inherent contradiction of libertarian paternalism, no one will call them on it. But the question remains: who are they to nudge more towards more savings? Perhaps I enjoy living for today. Certainly the way that the tax/benefit system is set up in Canada, anyone who earns under $40K a year is probably making a rational choice but *not* saving - any investment or RRSP income will simply reduce the amount that they receive in Guaranteed Income Supplement. Save More Tomorrow might be great for governments, who can reduce future income support obligations - I'm not so sure of its benefits for the average consumer.

[Frances Woolley]

Bob: " I don't recall my economics courses having a great deal to say about the source of people's preferences."

Perhaps that's a limitation of economics - perhaps some preferences are in some sense "better" than others.

[Ryan]

Every preference is always in some sense better than any other preference.

[nemi]

"But how does a behavioral economist know which direction the consumer should be nudged in?"
How does welfare economist know that people do not want to be nudged (since they conclude it by locking on some lines on a blackboard)?

Most people think that e.g. forced retirement savings is a good idea - even though neoliberals insist that they would be better off (or at least not worse of) if they had full control over their resources. Is that “respecting people's choices” or an attempt to force their own preferred ideology on other people?

[Victor]

"But the question remains: who are they to nudge more towards more savings? Perhaps I enjoy living for today. "
There is no problem if you enjoy living for today, libertarian paternalism doesn´t ban your choice of living for today, it simply shows possible choices in a different way(framing effect) or chance the default choice (status quo bias if I remember correctly). If you want to live for today you could easily choose that.
Thanks for the answer.

[Frances Woolley]

Victor - people choose option A when framed one way, option B when framed another. So the behaviouralist says we can't use these revealed preferences to order options A and B. How then do we order them? The libertarian paternalist's answer (as I understand Thaler and Sunstein) is that your planner has preferences over A and B, and the aim of nudges is to align your actual choices more closely with what your inner planner would like.

But it still doesn't resolve the problem: how does anyone know what my inner planner wants? Without some kind of meta-preference ordering by which to evaluate choices, we're no further ahead.

[Brad]

One way to reveal meta-preferences might be to give people meta-choices. How would you like to be nudged?

[Peter T]

Impossible compared to what? The idea that my choices will faithfully reflect my preferences requires heroic assumptions: for a start 1. that I know all (or most) of the significant consequences of my choices; 2. that price plus the ordinary experiences of life will tell me all I need to know; 3. that I have the time and mental ability to compute and compare all the likely consequences of the choices I perceive. And there are other assumptions yet, equally heroic. In other words, a rational calculus economics is just as impossible as a behavioural welfare economics. Sorry, but constant uncertainty is a condition of life.

To illustrate: in buying a larger soda I am also buying (to the extent that other people do, and that this and related choices lead to greater obesity) more expensive airfares, less comfort on the subway, higher healthcare premiums and so on. Many of the consequences will be unknown to me or anyone, some will be unknowable, others will only be apparent after extended research. Price did not tell Ford Pinto drivers that they were also buying a chance at a fiery death. Neither your inner nor your outer planner can be sure that there are not Pintos somewhere in life - they just have to make the best guess they can.

When I had to make the final choices on buying expensive machines, I had a civil engineer, a physicist, an accountant and a lawyer to advise me, plus a set of regulations, guidelines and laws. Which of my planners was in operation? And how much was the final decision my preference?

[Phil H]

The theoretical discussion is all very well, but I still can't understand the example that you've tried to construct there.
"At the price shown, consumers would like to buy Q2 units of soda. A ban on extra-large sodas restricts their consumption to Q1 units. At point Q1, consumers would like to buy larger sodas, and firms would like to supply larger sodas, but they are not allowed to."

This really doesn't make any sense. Either (a) soda is fungible or (b) it's not, and we treat each cup size as a separate product. If soda is fungible, then the ban on sales of different cup sizes doesn't make any difference to how much people buy. If soda is not fungible, and the commodity in question is "a large cup of soda", then the consumption is not limited to Q1, but to zero.

If the cup sizes count as separate products, then you've already pretty much conceded the behavioural economics point: that consumers have no clue how much they want, they just buy the packages that big organisations tell them to.

If soda is fungible and cup size has no impact on consumer purchases, then your whole argument makes no sense, because there will be no drop in consumption.

Now, I don't know if this is just nitpicking. I'm aware that you were just using the example to lead into a bigger question. But it feels important to me, because I think you've stated the issue completely backwards. Behavioural economics doesn't depend on elevating some other set of welfare to godlike status; it is just about pulling down market fundamentalism from the godlike pedestal on which economists have put it. I understand (I think!) the value of elevating a simplified view of "consumer preference" to being the only good - it makes the maths of economics possible, and has produced a lot of great results.

But your complaints here about theology seem utterly misplaced, and a bit of a symptom of the economists' arrogance that people have been writing about lately (mostly sociologists who can't sell popular sociology books any more because economists have nicked their niche). Sure, when you do behavioural economics, you do have to select a bunch of criteria against which to judge behaviour (and thus far, the methods used to pick those criteria have been ad hoc, because no-one knows what true human value is). But those criteria are not set in stone. I'm reminded of development economics, where various definitions of development can be applied and argued over, and rated as important alongside efficient market equilibria.

[Christiaan Hofman]

"Unless soda drinking harms other people, any analysis that begins with the premise that people's actions are based on rational self-interest must inevitably reach the conclusion that soda bans reduce welfare."

There's an important wrong assumption in this sentence, something that I think goes to the hart of the problem. The problem is that one important category of "other people" is *always* there, but seems to be always ignored. That "other people" is you in the future. What you do right now affects you in the future. But in your decision today you typically don't take into account these effects, or not at the appropriate cost. This is a market failure just as much as external costs to third parties. And it is precisely these externalities that soda bans try to correct for.

[Greg]

Awesome post, Frances!

I think we need a subdiscipline called axiological economics.

[Frances Woolley]

Phil H: " Sure, when you do behavioural economics, you do have to select a bunch of criteria against which to judge behaviour"

That's the #1 point I wanted to make here. I agree the example wasn't as beautifully set out as it could have been. An alternative lens through which to view the ban on extra-large sodas would as an example of the "deadweight loss of overconsumption" - see a few posts down. On theology - all I said on that subject was that people needed a little preaching now and then.

Christiaan - "That "other people" is you in the future." This argument is pretty familiar to parents, isn't it? "Put on sun-screen, you'll get cancer/wrinkles/end up looking like a worn leather shoe" "Don't have time, sorry." We care about our kids' future selves more than they do. At the same time, this doesn't answer the question *how much* the present self should be expected to sacrifice for the future self - see Shangwen's point above. There Smiths fan's excuse for smoking has some merit: "I smoke 'cause I'm hoping for an early death" - someone might rationally choose to place a low value on their future well-being.

Greg - thanks.

Peter T - this post was written for my intermediate micro for non-majors class. After going through choice theory, indifference curves, income and substitution effects, derivation of demand, I showed them Dan Ariely's TED talk. Huge mistake. They were totally convinced by Ariely, and left the class with the impression that all of this micro choice theory is useless. Now have to figure out how to salvage intermediate micro theory from the wreckage.

[Frances Woolley]

Shangwen - the behavioural economists do have some really good personal-life-management nudges, e.g. putting cookies in opaque containers. But it does seem that anything that is viewed as an intervention in the kitchens of the nation is a hard sell.

Brad: " How would you like to be nudged?" Ask a hypothetical question, get a hypothetical answer! One argument made by, say, Dan Ariely is that we really don't know what our preferences are, and it's hard to work it out.

[Ryan]

Isn't "nudge" just a new word for "increased transaction cost?" The thing I dislike most about the concept of nudging is that it is old theory dressed up in new terminology, and the new terminology is designed to paint the nudge as being something other than what it is.

Basically, the idea that if you increase transaction costs by units smaller than the smallest available unit of account, then people won't complain about it. But if you subject a large enough number of decisions to nudges, then you're looking at a significant cost increase after a while.

Sort of like how nitrites in small enough quantities are perfectly safe for human consumption, but if all you eat are processed foods, then you start consuming enough nitrites to cause real damage to your body.

But anyway...

[Bob Smith]

Christiaan"But in your decision today you typically don't take into account these effects"

But that's a hefty assumption, isn't it? Maybe soda drinkers do take into account the effects on their future health, but heavily discount them. Isn't that just another way of saying that their preferences are wrong? I agree with you that people don't take into account their health effects an the appropriate cost, but that's a separate story resulting from third-party financed health care (and, an irrelevant one if people don't take into account future health effects).

[Alice Woolley]

Frances I think the problem I have with some of your analysis is that, as I think Whitford pointed out earlier, we don't have a non-nudging society anyway. We nudge people's behaviour all the time, and we do it in ways that create choices that may or may not be the ones that a person would make. There is no neutral. Take for example Thaler and Sunstein's story about the order you put food in in a buffet. People take more of the food that is earlier, whether it's healthy or not. There is no way not to present the food in a buffet in a particular order, and so the only question is what order you are going to put it in. This is particularly true when you look at the intersection between law and people's economic choices; the law consistently creates incentives and structures for people to behave in particular ways, and the only question is what those incentives will be, not whether.

Also, while I share some of the dislike of paternalist decision-making expressed here, I also don't see all decisions that could be characterized as paternalist which, when made in a moderately well functioning democracy, as equivalent to quasi-fascism. Yes, Bloomberg banned big sodas, and did lots of other clearly paternalistic things. That's a pretty good campaign platform for the other candidate if s/he thinks it has any traction.

I note that Sunstein in legal circles is known far more for his work on democracy and legitimate regulation than on behavioural economics- and that's in part where I think his view of the legitimacy of "paternalist" choices would come from.

Finally, if the state doesn't want to pay to look after you in your retirement, I don't see that as somehow malicious. Why shouldn't we as a democratic collectivity avoid paying for people who are looking to free ride. That may be your individual choice, but it seems to me like your choice is to be the grasshopper, and if I'm the ant I might be more inclined to nudge you to save than to give you my food.

[Frances Woolley]

Alice - interesting. I'd wondered if you would make a deliberation-type argument - that it is possible, through some kind of deliberative process, to work out what one's underlying preferences really are (I'm not convinced that it is, but I think there's an argument to be made there, but perhaps that's what you're hinting at when you talk about democracy).

My point is not that nudges are bad - as you say, we're nudged anyways. Sugar/salt/bulk flour is almost always on the bottom shelf at the grocery store, in plain boring packages, while cake mix is at eye level, with mouth-watering images on the front.

My point is that - once one accepts the basic premises of behavioural economics i.e. that it's difficult to know what our true preferences are - it's impossible to carry out a straightforward welfare economic analysis of the type taught in Econ 1000. As L Perrault pointed out, there may be ways of merging the two - I haven't read those papers yet - but it's not straightforward.

Do you think that democratic processes can provide an alternative (by which I mean, alternative to ECON 1000 type welfare analysis) way of evaluating policies?

[Alice Woolley]

In answer to your last question - no. If democracy reveals preferences, it is only in an incredibly imperfect way, and a way impossible to track back to any individual person. But why can't the answer simply be that it *is* impossible to carry out a straightforward welfare economic analysis. or, more accurately, that if you run that analysis you get an answer, but the answer may be truthy and not true?

I think that the claims of empirical psychology (which behavioural economics uses/rests on) are often themselves fairly soft (they show some of the truth but not all of it) and I think most of the time they more complicate other forms of analysis/decision-making than offer any real replacement for it.

[Frances Woolley]

Alice: "But why can't the answer simply be that it *is* impossible to carry out a straightforward welfare economic analysis.:

I guess the response "But welfare economic analysis is the only thing I know, and without it my life would lack purpose and meaning" is not acceptable?

More seriously, yup, I recognize the flaws in the current paradigm, but I'm uncomfortable with the alternatives. We've had this conversation before, haven't we - if I recall correctly, you tend to be more comfortable with rights-types paradigms and I tend to be more comfortable with utilitarian-type paradigms? And basically welfare economics is just one big souped up utilitarian calculus - with the large and very significant exception of the capabilities approach pioneered by Amartya Sen, Martha Nussbaum and others.

[Gepap]

I am sorry, but why should public policy care about an individual's "internal planner's" desires? We nudge people to save because we want there to be reserves of resources available in times of scarcity. We nudge people towards eating less non-nutritional things in order to minimize disease and illness and their related costs. Public policy should not care about whether an individual might want to eat or drink to excess and thus get ill - if that is what an individual wants, they will overcome the obstacles that have been erected in their idiotic path.

[l_perrault]

"Peter T - this post was written for my intermediate micro for non-majors class. After going through choice theory, indifference curves, income and substitution effects, derivation of demand, I showed them Dan Ariely's TED talk. Huge mistake. They were totally convinced by Ariely, and left the class with the impression that all of this micro choice theory is useless. Now have to figure out how to salvage intermediate micro theory from the wreckage."

Ariel Rubinstein has two good "essays" why economists should be careful of behavioral economics and experimental economics. Because he has been actively doing research with experiments he is an especially credible voice on the debate.

Easiest to read: Discussion of “BEHAVIORAL ECONOMICS” - http://arielrubinstein.tau.ac.il/papers/behavioral-economics.pdf

Another one in the European Economic Review: "A theorist's view of experiments"- http://arielrubinstein.tau.ac.il/papers/61.pdf

[Patrick]

Say we call the adverse health effects of drinking fizzy sugar water externalities, and we then correct the market failure (somehow). Price increases and quantity demanded decreases from Q2 to Q1. Any hand wringing about DWL?

Now how is this really any different from the paternalistic nanny state 'nudging' us to demand Q1? Presumably supply will equilibrat and the price will increase and all will be well.

And how is it any different from a brilliant ad campaign that convinces people to switch to e.g. that vile vitamin water stuff (presumably healthier). Quantity demanded decreases etc.

We don't fret about prices 'nudging' us, nor do we fret about marketers 'nudging' us for their own ends, but we need to worry about Gov't openly (doing it secretly would be different IMO) nudging us not to to do blatantly silly things?

[Frances Woolley]

Patrick - see post about Pigou, paternalism and potato chips, http://worthwhile.typepad.com/worthwhile_canadian_initi/2011/12/pigou-and-paternalism.html. Calling the harm to the future self an externality is just a way of ducking the question: why should the nudger know better than the nudgee how to trade-off present and future well-being of the same individual.

L Perrault - more things to add to my must read list! Thanks.

[Gepap]

What "well-being" is derived from a 32 oz. soda?

In many movie theaters now the smallest serving of soda you can find will likely be 24 oz. and that is in a plastic bottle. At the fountain the smallests might be 32oz. The largest can reach a full half-gallon, with the price difference being $1, between say a $4.50 "small" and a $5.50 large. Of course, the actual soda costs the theater next to nothing, which is why they can double the amount sold for only $1 more. But given that most of the profits for the theater chain come from concessions, it is necessary to get as much as possible. The theater can rationalize the relative large costs of even the "small" by pointing out the fact it isn't small at all, and having already made the consumer chose that item, the theater knows it is easier to sell the consumer even more by pointing out the great "value" of perhaps even doubling the amount for "just" $1. There is no real well-being being derived, anymore than people derived from the stale popcorn in the experiment mentioned before. And this is NOT something that was consumer driven - the theater chains themselves set the sizes, something that has a long production and distribution chain attached to it (someone has to make those cups and stock them). if they have increased the sizes so much it has everything to do with the everyday costs and revenue realities of their industry. Theaters get little from the box office in the first few weeks (most of that money goes to the studio) and they depend on movies playing for a long time to make a profit off them, but that hardly happens anymore - most movies experience very large drop-offs in ticket sales from the first few weeks. So, again, concessions and other schemes (like installing arcade games or other related business activities) must make up for this gap as theaters try to make up for their loss in box-office revenue.

Of course, none of those business realities can be modelled, so academic economists continue to willfully ignore reality in their analysis because the truth seems to expose nasty moral realities they would rather see disappear.

[Patrick]

Is it ducking the question? Say I only care about myself in the present. Other people's current and past dumb choices affect me now.

[Gfris]

I can think of a purely neo-classical defense of the soda ban that has to do with externalities and not behavioral economics. People drinking large sodas are more likely to become diabetics (and have other later health problems) that will increase the costs of (my) health care in the future. I noticed a comment above implying an externality between the soda consumer's current and future self. Why should the government arbitrated between the two seems to be the question. My point is that if YOU drink a lot of Big Gulps and get fat, it is going to cost ME money in the future. So it is not at all a simple matter of individual choice and paternalism.

[l_perrault]

Gfris- This can't be considered an externality in the classical sense. His consumption does not affect your utility directly, but it will affect prices for health services but that isn't considered an externality. Although you could always argue that there is some sort of congestion in privately provided health services than maybe you have an externality story.

[Frances Woolley]

To be clear: when soda creates a negative externality, restrictions/taxes on soda are welfare improving. There is no contradiction between traditional welfare economics and behavioural economics.

The conflict between welfare economics and behavioural economics only gets interesting when we're talking about behaviours that cause self-harm - smoking in some isolated wilderness spot, reading fmylife.com even though it makes you feel sad and depressed, etc. (And please let's not get into the "I suffer when other people hurt themselves by smoking" debate - smokers contribute so much to the well-being of others by paying taxes and dying before they have a chance to collect much by way of pension plan payments - plus as long as they die fairly quickly they don't cost the health system much).

[Determinant]

That was an episode of "Yes, Prime Minister".

[Peter T]

Frances

A useful post. A further point might be that any extended division of labour progressively removes both the ability to rationally form preferences and the ability to express them. In other words, the division extends to the range of choice available, the ability to know all the factors relevant to choice, and the ability to actually choose. So the more you trade, the less you can have preferences. An example might be my preferences for how safe the buildings I use are: I cannot know how safe they are, I cannot know (in most cases) how safe they could be at what cost, and anyway the choice has been removed from me to building code regulators, builders and so on. Even with the soda - I can choose size, but I can't choose the ingredients, or know how they are made ( I can and do often read the fine print, but I can't be sure it's telling the truth, I can't understand what all the chemicals do and so on without spending time and money that I can better spend on other things). I realise this doesn't help you salvage micro choice theory.

[Neil]

The economics view that people know what's best for themselves is inherently a problem. Humans are still an animal, governed by our instincts ("gut feelings,") and a herd mentality. When we step back from the problem, we can make a rational analysis of our choices, but actually making rational choices every day is not just really hard, but probably impossible.

Economics tells us a lot about why people act the way they do, and how to influence people's daily choices in a direction that decision makers have determined is preferable, but I think the attitude that economics can tell us anything about just what the ideal choice is is hubris on the part of economists. People are known to, as a group, place disproportionally little value on the future, yet economists treat this as a rational decision instead of an evolutionary leftover of being able to reasonably plan on snuffing it by the time we reach middle age. So consuming more sugar today is similarly treated as a rational choice instead of behaviour hardwired by our ancestors environmental conditions that no longer apply.

Those who are willing to step back and look at the problem from the perspective that maybe we don't always know what's best for us can come up with policies that will make not only our overall society better off, but very possibly most of the individuals within that society better off as well.

[Determinant]

Or looked at in another facet, classic micro assumes that information is costless and the ability to process that information is also without cost or limit. Behavioural Economics, OTOH, posits that information-processing is costly and time-limited. Macro had a similar debate between Hayek and Keynes over whose system was more "general". Keynes assumed costly information, Hayek used cost-free information and processing.

Any engineer will tell you that placing limits on processing ability drastically affects the outcome and the strategy chosen to reach it.