Saturday, December 10, 2011

The World According to Frank, Part II

This is a followup to my previous post. Robert Frank likes to talk about "positional goods." These are goods I like, not because they confer any direct benefit on me, but because I feel happy when I have more of this type of good than what others have. Similarly, I feel worse when others have more of this good than I do. Examples are large houses that are mostly empty or Beamers driven by people who couldn't tell the difference between riding in a Beamer and riding in a Hyundai if they closed their eyes. An extreme example is this house in India. Frank thinks that positional goods are ubiquitous and lead to social waste - basically, a lot of time and resources wasted over nothing. From his point of view the social welfare loss is large, and can be corrected with a progressive consumption tax.

But hold on. Aldo Rustichini likes to study envy and regret. He is an economic theorist who also has in interest in economic experiments and neuroeconomics. Rustichini likes to view envy and regret as elements of psychology that propel learning and economic development. Envy is a force that makes me want to replicate the successful behavior of others, while regret is a force that helps me to correct my own errors so that I can become successful.

To motivate these ideas, Rustichini likes to talk about how performance is evaluated in elementary schools in the United States. I went to elementary school in the 1960s in Canada, where we were graded in each subject, and were given a report card to take home with little else other than the odd short comment next to the grade - "shows promise," "improvement needed," etc. Grades were of course supposed to be private, but we knew what was going on. I felt good. Some others felt bad. Concern that some children in elementary school might feel bad led people in Schools of Education to push for changes in performance measurement in elementary schools. By the time my own children went to elementary school in the United States, grades had been ditched in favor of a lot of words that the teachers had to put together to describe performance on a report card. The parents get roughly the same message, but now the kids have a harder time figuring out who the top performers among their peers are.

Rustichini seems to think that grades in elementary school are a good thing, and that the modern wishy-washy approach to student evaluation is thwarting envy, which will actually reduce the amount of learning that takes place. If you know who the top performers are in your class, you can watch them, see what they do, and try to replicate it. Maybe you can get them to show you a few things.

What about "positional" goods? These can actually be performing a useful economic role, even if they are pure "waste" in the conventional sense. This is what signaling is all about. In a signaling model, acquiring a signal may be costly and a pure waste, but it reveals information because the signal is less costly to acquire for "good types." Sometimes people argue that there is a signaling component to higher education. It doesn't matter that you didn't learn anything in college, because employers know that only the smart and diligent ones actually finish.

To get envy working as a force for social good, we need to be able to correctly identify economic successes. How are we going to do that if the successful people hide their wealth? Someone who pushes her Beamer off a cliff in public is demonstrating to us that she is successful - so successful that she can waste resources. I see this and it makes me curious. How does a person get so rich that they would be willing to destroy their Beamer? Maybe I can replicate her behavior and get rich too, not because I want to destroy Beamers, but because I want to relax my budget constraint in a big way.

You don't want to tax the positional goods or pass laws against having them, for the same reason that you don't want to take grades away from the elementary school kids. In doing so, you are thwarting envy. When you do that, you will get less learning and less innovation, and on average we will all be worse off.


  1. stephen,
    david friedman debated Robert Frank, all posted on friedman's Ideas, after a Frank op-ed in the new york times in June 2010.

    the blog links are collated at

  2. Stephen,

    I think you're running together two different concepts. A positional good is one that, by its very nature, can't be provided to everyone, e.g., waterfront property, education at an elite college, one of Newton's original manuscripts, etc. Although access to positional goods depends on one's income relative to others, the motive for acquiring these goods need not be the quest for superior social status (though, of course, it could be).

    "Envy is a force that makes me want to replicate the successful behavior of others." This is just a stipulated definition that bears little relation to what we mean by envy. If someone asked, "why do you say he's envious of the the wealthy," which answer would you find more convincing: 1) "because he's working so hard to become rich himself" or 2) "because he's always putting rich people down"?

    (A lot of neuro-economics suffers from this kind of conceptual confusion. If you're interested, you can find a philosophical critique of neuroeconomics here:

  3. Steve
    even if we accept your arguments, the best it does is a theoretical argument that poses again at best a trade-off between two forces. To argue that we should not tax it you must argue that the positive role of envy or signaling outweights its cost, without a quantitative model that balance the two forces your arguments are not very convincing. While I dont have a quantitative model either, it seems clear that an MBA as a signaling device for example should be taxed. Taken as granted that the education value is zero and everything is signaling, using 50000 dollar as a mechanism to signal the smart and diligent seems excessive. Why shouldnt we use a standardized intelligence test or a test that require two weeks of hard studying to get the same signalling value? As a mechanism to signal it might have been an evolutionary strategy, for people that can actually invent a smarter mechanism your argument seems quantitativey, well, invalid. Have you ever considered to make your arguments quantitatively robust?

  4. This is under the assumption that in fact these positional goods exist. How am I supposed to tell if something someone owns is a pure waste, if they really like it, or they are just holding it as an asset? I can't look inside their minds. Further, what am I to make of high heels and makeup? They certainly look positional, don't they? Do you want a tax on women?

  5. Neither Frank, nor I, want to eliminate position, rank, context, or prestige. We evolved for these things. People who cared about how well they did, or ranked, or consumed, relative to other members of the species, were motivated to do better than other members of the species, and thus usually did better than other members of the species at reproducing, and having children who survived. And therefore, their rank-consciousness genes became more and more relatively common in the species as time went by.

    We get valuable pleasure from relatively high rank, context, and position, and it motivates us. The goal is not to eliminate it, simply to get the same amount at a far lower price. It's about efficiency, not elimination. Do you see Frank calling for Maoism, where everyone is supposed to have the exact same status and consumption? He's calling for a 70% top marginal rate. This is the rate we had – or higher – during the post-war decades until the era of Republican domination began in 1980. And that post-war era had an explosion in innovation and discovery, with growth much greater than in the Republican era with top marginal tax rates half, or even a third as high.

    Frank and I simply realize that we can achieve approximately the same amount of status, rank, and position, all the benefits, all the pleasure, all the motivation, at fraction of the price through highly progressive taxation which does nothing to change the relative and obvious rankings.

    Suppose the top marginal rate is 70%, instead of 35%, or 25%, or zero like the Republicans would really like – the marginal rate drops as income goes up, maybe even going negative! Then, if someone innovates and works hard, they can look forward to 30% of $100 million, $30 million, instead of $65 million, or $100 million.

    Now, please really think about this. If $30 million was just as rare, just as prestigious, just as much of an achievement, same rank, as $100 million, would it really give much less motivation and pleasure? Your car is just as rare and prestigious; it seems just as gorgeous and amazing, same for your house (which is located on the same beach, because you have the same bidding power as before for that land. Ranking hasn't changed; everyone you bid against pays the same higher tax rate.) Are you going to really feel much less pleasure and motivation?

    Did the Magnum P.I. Ferrari of 1979 really give much less pleasure to its owners than a Ferrari that's equally exclusive, equally rare, today, but costs five times as much? Or is it the case that the Magnum P.I. Ferrari achieved very close to the same level of pleasure and motivation, but at a fifth the cost? Now that would be tremendously efficient. The new Cadillacs of the 70s with their plastic wood and pimpstripe velour seemed as beautiful and enjoyable and prestigious and motivating to me, and it seems everyone at the time, as the cars of equal exclusivity today costing many many times as much, with their British Walnut and leather, and four million horsepower that almost no one will ever use anything close to.

    And what does it suggest? That with highly progressive taxation, say at the 70% top level, like we had for many extremely high growth decades (higher than with lower rates), we can still have about the same level of motivation, and pleasure from the rank and position of achievement, same good stuff, but now we get these trillions per decade raised from the tax that can go into basic scientific and medical research.

    Which would result in more innovation?

  6. Now before you answer, let me let you know that in 2007 basic scientific and medical research from all sources, public and private, was $64 billion/year ( So we're talking about something like tripling basic scientific and medical research – and on top of that, this is government sourced, so it's not inefficiently hidden from everyone so no one can use it for free. It's posted on the net, for frictionless use 24/7, worldwide.

    What would increase innovation more? What would increase long-term growth more?

    And if you're still unsure, keep in mind the income and substitution effects, the backward bending labor supply curve, the historical relationship between wage/hour and number of hours worked (with wage/per hour skyrocketing from 1900 to 1970, but hours worked per week DECREASING), and this graph:

  7. Williamson: "Further, what am I to make of high heels and makeup? They certainly look positional, don't they? Do you want a tax on women?"

    Not entirely positional. Even if we assume that women are in some zero-sum competition with each other, it is not zero sum to men, I think.

    Would we (men) be happier if all women in society were 10% better looking? Yes.

  8. What I tell my wife is that she looks fine in running shoes.

  9. Stephen,
    I second Richard's comment. If you want a good description of why taxing pure positional goods doesn't harm purchasers read Ng, Y K, Diamond Goods, AER (some time pre 1989 when I first read it in his class). If you believe positional goods exist you should be advocating taxing of them as the solution to US deficit. The tax makes the buyer no worse of and does not have the effect you describe on the envious. If the price of a diamond is doubled the envy does not disappear - it just kicks in at a size that is half as big. The creamon the optimal tax take is that we save half the resources currently wasted digging diamonds out of the ground.

  10. The key idea here is that you see someone consuming something that you view as socially wasteful. But who is to say when the thing I have bought is really doing me no good, and society should take it away? How does Robert Frank know that the dollar he takes away from the rich person and gives to the poor person is going to be spent on something that Robert Frank deems socially useful? The more I think about this, the less I want Robert Frank determining what I should and should not have.

  11. Another thought: This is the down side of envy. Robert Frank is so green with envy that he's willing to spend an enormous quantity of time (a social waste, in my view) concocting a contorted theory, and marketing that theory, in order to convince people to use the power of the state to whack rich people.

  12. Stephen: "Another thought: This is the down side of envy. Robert Frank is so green with envy that he's willing to spend an enormous quantity of time (a social waste, in my view) concocting a contorted theory, and marketing that theory, in order to convince people to use the power of the state to whack rich people."

    Are you kidding, Steve? The issue of how hard to whack the rich might be the most common subject in Western politics.

    Let's ask a politically connected, retired academic for his views on the subject...

    "Kill all the rich people. Break up their cars and apartments. Bring the revolution home, kill your parents, that's where it's really at." -- Bill Ayers

  13. Stephen,

    If my girlfriend has taught me anything, don't tell a woman she looks 'fine'.

  14. "The key idea here is that you see someone consuming something that you view as socially wasteful. But who is to say when the thing I have bought is really doing me no good, and society should take it away? How does Robert Frank know that the dollar he takes away from the rich person and gives to the poor person is going to be spent on something that Robert Frank deems socially useful?"

    Stephen, so you're saying that something that's hard to estimate precisely, like positional externalities, should be assumed to be zero. It's hard to estimate precisely, so we won't estimate it; we'll assume it away. We'll call it zero.

    Well, I have news for you, zero's an estimate. And if you look at the qualities we typically want in our estimators; statistical consistency, efficiency, unbiasedness, arbitrarily choosing the constant zero stinks. We can often do far better in estimating than throwing all of the information away that we do have, as imperfect and fuzzy as it may be, and saying we'll just choose the constant zero as our estimate.

    Hey, pollution is hard to estimate precisely, so let's assume it's zero. That way no people like Robert Frank can take away our polluting products and shift production to things they deem less socially harmful. That will be a lot more likely to maximize total social utils than intelligently trying to estimate what the harmful externalities of pollution are and taxing them. All hail the constant zero estimator! KISS!

    And let me add this, I feel little envy (in fact, we're fortunate to actually be pretty high up in income and wealth). It's a non-motivation for me (there are lots of things I admire, and would like more of; I wish I could dunk a basketball, but I don't feel anything negative for people who have more of them). I just want to use tax policy in response to positional externalities because it would create massively more utils for society, especially over the long run, if the money raised were invested well. From what I've read of Frank, he just wants to do good for society; that's his motivation, not to get back at rich people he envies.

    Also, he's not choosing specific goods to tax. He writes about the problems of this. He gives the example of historic regulations on jewelry. In response, rich people started making their buttons out of gold, and adding ever more buttons to their coats, which is why today we still have a number of buttons on the sleeves of our suit jackets. He instead advocates just a highly progressive tax, as, in general, the very high marginal cost goods are much more positional as a percentage of their utility.

  15. "If my girlfriend has taught me anything, don't tell a woman she looks 'fine'."

    No, it's OK. It has taken a while, but my wife now understands me, more or less.

  16. I like how Richard Serlin tries to make himself equivalent to Robert Frank. Not in the same ballpark, Dicky Boy, even if you're both batshit crazy. Just because you believe something exists doesn't make it so.

  17. Eric Falkenstein has an interesting article on envy and its role in econ & finance.

  18. I think someone who believes "whatever rich people spend their money on" is a usable definition of positional goods probably shouldn't be too scornful of people who believe that zero is a usable estimate of positional externalities.

  19. "He instead advocates just a highly progressive tax, as, in general, the very high marginal cost goods are much more positional as a percentage of their utility."

    Note the words, "in general", as in, they tend to be.

    Note the words, "are much more", which are different than, are 100%.

  20. "Note the words, "in general", as in, they tend to be."

    "Note the words, "are much more", which are different than, are 100%."

    Neither of these statements come with any evidence -- how exactly do you "know" these things exist? Serlin is the economic equivalent of Michelle Bachmann -- ill-informed and believes in fairy tales.

  21. All the qualifiers in the world won't change the fact that Serlin, and Frank, really are using "whatever rich people spend their money on" as an operational definition. I'd think that even people who share their premises might be taken aback by the contrast between the extensive and detailed knowledge of positional externalities they claim in the course of selling their idea, and the rough justice they're willing to settle for when it comes time to prescribe policy. If it's so bleedin' obvious that granite countertops are a wasteful status symbol, why "throw... all of the information away that we do have", and resort to added income taxation on people who wouldn't have any idea what their countertops are made of unless they took it into their heads to ask the help?

  22. Some questions for the critics of Robert Frank who are willing to set aside policy issues for a moment:

    1) Are there some goods that, by their very nature, can't be made available to everyone, e.g., having one's children attend a better-than-average school?

    2) Does access to these goods oftentimes depend on one's income relative to others, e.g., owning a home in a better-than-average school district?

    3) If all households increase their expenditure in hopes of acquiring positional goods, and all households remain in the same position relative to others, are these expenditures wasteful?

    4) Do concerns about social status (relative position) influence decisions regarding occupational choice, work effort, saving, expenditure, and investment?

    If you answer "yes" to these questions, then, leaving policy questions aside, I suspect that standard optimization models (if I may be permitted to generalize) are missing something of considerable importance.

  23. If I might come at this in a different way:

    It seems you don't like what you view as the lack of evidence that there are a lot of goods where a large percentage of the utility an individual gets from that good (or disutility he avoids) is rooted in the position, the context, and/or the prestige of that good.

    If such factors in the utility function of typical individuals really existed, then this would mean that highly progressive taxation (with the money raised well spent on little positional goods and investments) could greatly increase total societal utils.

    So, you (and many of your commenters) think how much pleasure someone gets from a McMansion, or mansion, would be about the same in a world where that home is in the 10th percentile of cost, as in one where it's in the 90th. And people would think of the house as just as high quality. Amazing to me you think that, but let's just move on.

    So we assume positional externalities are zero. We use the mighty constant zero estimator to estimate them.

    Now, the utility functions used in economics for centuries assume that the utility a typical individual gets from a dollar decreases as he gets more dollars, and not just a little, exponentially. The amount of utility from the millionth dollar is typically assumed to be a tiny fraction of the first. This also implies that highly progressive taxation can greatly increase total societal utils, and yet I don't hear you and your commenters saying, how do you know decreasing marginal utility of dollars (dramatically decreasing) exists? Why are you foisting your preferences on others? Why don't we just use the mighty constant zero estimator for change in marginal utility, and have all utility functions be straight lines – it's simpler too! (actually, I think a lot of your commenters would love this)

    And if you read Frank's books (and really just opened your eyes and thought about this and were honest), you'd see there's a lot of evidence.

  24. "...and really just opened your eyes and thought about this and were honest..."

    Richard, I resent that. Closed-minded and dishonest. Thanks very much.

    The idea here is that you see some element of human behavior - people buying what you view as useless, but visible, stuff - and jump to the conclusion that there is an externality that you want to correct, in Pigouvian fashion.

    Well, an open-minded and thoughtful social scientist sees some puzzling aspect of human behavior and starts thinking about what useful purpose that behavior could serve, before jumping to conclusions about how we should curb that behavior.

    After some thought, there are two far more plausible explanations for what you think you are seeing:

    1. You're not seeing what you think you are seeing. That is not "useless" stuff those people are buying. They actually like it, for reasons that have nothing to do with what other people think.

    2. It's signaling, as I outlined above. In that case it's not an externality, and you don't need to correct for it.


  25. Regardless of the other issues, the idea that there's something wicked about declining to act on inadequate information needs to be buried as deep as we can possibly bury it.

  26. I apologize if I've offended you, but given the resistance to seeing this, especially among your commenters, seeing something that seems very obvious to me, I have to wonder if they are really thinking about this and being honest with themselves.

    It's not that these products are necessarily useless, just that a high percentage of the utility they give people comes for positional reasons, and that's zero-sum-game utility for society as a whole, and the same positional amount can be provided anyway, by maintaining the rank, but lowering the cost of achieving that rank (i.e. with a higher tax on all wealthy people, the rank order stays the same). In other words, wealthy people would lose little utility if they went to much lower cost alternatives, but everyone else did also, so rank stayed the same.

    The same signaling can be done at a far lower societal cost. As Tom Hird said in these comments, a tax can make it so you send the same signal with half the carots – and the paper that's from was from a winner of Australia's top economics honor.

  27. "Regardless of the other issues, the idea that there's something wicked about declining to act on inadequate information needs to be buried as deep as we can possibly bury it."

    I intend to use this phrase as often as I can.

  28. "It's not that these products are necessarily useless, just that a high percentage of the utility they give people comes for positional reasons"

    You're just restating the same thing, without evidence, that you've stated many many many times before. You believe it to be obvious, we say it requires evidence. You want to choose for other people, we don't without some reason to think they're creating an actual externality. How can you possibly know why someone buys something? Arrogant doesn't even begin to describe you, Serlin. You border on sociopathic.

  29. "How can you possibly know why someone buys something?"

    How can we possibly know how much utility people get from something, even in a relative way among bundles of goods? How can we possibly know that peoples' utility per dollar plunges exponentially with more and more dollars, that the utility from the millionth dollar is a tiny fraction of the utility from the ten thousandth dollar?

    Yet this is used in ALL economic models. We don't use straight line utility functions. And people who use, and qualitatively believe, utility functions with exponentially decreasing utility aren't all sociopaths (although maybe to an extreme enough libertarian they are).
    The answer for why economists believe marginal utils per dollar plunges exponentially is that there's tons of evidence, some of it is the formal kind (regressions, social laboratory experiments, surveys) but most of it is informal. And you know what, the informal can have just as flawless logic chains anchored to even milder, or far milder, assumptions than the formal. As the great growth economist, Paul Romer of Stanford, said:

    In evaluating different models of growth, I have found Lucas's (1988) observation, that people with human capital migrate from places where it is scarce to places where it is abundant, is as powerful a piece of evidence as all the cross-country growth regressions combined. But this kind of fact, like the fact about intra-industry trade or the fact that people make discoveries, does not come with an attached t-statistic.

    – Journal of Economic Perspectives, Volume 8, Number 1, Winter 1994, Page 20.

    The evidence for positional externalities being substantial is very similar to the evidence for exponentially decreasing marginal utility of dollars. It's mostly informal, without t-statistics, but often the "informal" depends on assumptions far milder than those depended on by t-statistics. You want to see the evidence formal and informal? You've heard some here, but it's a lot, and I can't recite Frank's books. I suggest you buy one.

  30. Is Richard Serlin a troll, or being trolled?

  31. "Is Richard Serlin a troll, or being trolled?"

    Both. He entertains me, so I pester him about evidence, which he cannot provide. But he is definitely trollish, or perhaps he really is an idiot.

  32. Richard has actually been kind enough to send me a holiday gift:

    "Falling Behind," by Robert H. Frank

    Thank you Richard. I will display this book prominently on my desk in order to position myself correctly. In return, I was going to send you the two-storey inflatable reindeer (see, but then envy got the better of me, and I have bought that for myself.

  33. I really wish I had more time for this.

    Sometime in the not too far future I'm going to put together a post, or long comment, summarizing the evidence that position, context, rank, prestige, etc. are a substantial part of the average person's utility function.

    In the meantime, I respectfully ask that you just sometimes really think about whether you would consider a product as high quality and enjoy it as much if it were in the 5th percentile of cost and exclusivity as if it were in the 95th (or if you wouldn't consider the quality a lot higher and enjoy a lot more a product that's in the 5th percentile, if it were instead in the 95th). If your home was considered ghetto, and the people in your neighborhood were in the 5th percentile of income, would you enjoy it just as much, even though it was exactly the same in absolute material terms.

    And then ask yourself about the people you've known in your life, or read about, and whether science says our DNA varies that much that they could be so different from the mean, or whether you've moved among and observed people that segmented off greatly from the mean, how greatly? how rarely? what's plausible given mild assumptions like what we see is actually there. Try to estimate these things and their strength unbiasedly, without a strong bias towards zero, as much as the constant zero estimator is so attractive to many.

    I'll leave you with two interesting papers:

    From Nobel Prize winning economist Gary Becker, "Evolutionary Efficiency and Happiness", mimeo 2004.

    From the August 2005 QJE, "Neighbors as Negatives: Relative Earnings and Well Being", by Erzo Luttmer.