Thursday, June 30, 2016

Some stuff economists tend to leave out

Chris Arnade got a PhD in theoretical physics, became a bond trader, and then became a photographer who documents drug addiction. He frequently writes on Twitter and elsewhere about social problems. Recently, he wrote a tweetstorm (series of threaded tweets) about his interpretation of Trumpism. I think Chris leaves out a lot of the reasons for Trump support - xenophobia, fear of racial conflict, misplaced nostalgia. I also doubt that the drug addicts that Chris spends his life documenting tend to be Trump supporters. But I think Chris makes some really good points when he writes:
Many people have no chance & they know it & reminded of it daily. On TV, on the web. They see what they are missing. And who is winning...The response to this is, we all have Iphones now. Everyone has all this stuff they never had before...Response 1: Maybe when we measure someone by how much stuff they have, those at bottom will never be happy. Because others have far more...Response 2: Maybe trying to get all that stuff is painful. Maybe the monthly payments needed to keep up isn't fun...Response 3: Maybe, crazy as it might sound, having an Iphone, and other things, just isn’t the key to happiness after all...Response 4: Maybe people need meaning beyond economic. Maybe they need to feel included, to have strong bonds to institutions & groups...Maybe the predatory, hyper rational society we have built has stripped institutions of their meaning...Maybe it has all left people, who used to get meaning from them, with very little, other than anger.
I think there's a lot to this. Putting it in econ language: 1) Social preferences are probably very important in reality. 2) Utility functions often contain inputs like inclusion, human connection, and a sense of meaning, for which no markets exist.

Econ rarely deals with these things. When is the last time you heard an economist talk about how minimum wage policy affects people's desire to feel like society cares about them? When is the last time you heard an economist talk about the need to include strong interpersonal relationships in GDP? When is the last time you saw a welfare analysis of tax policy that took inequality preferences into account?

The formal machinery of economics is totally equipped to deal with these things. You can put love in a utility function right alongside consumption, just call utility u(c,L). You can incorporate community stability into an urban model or a model of migration. You can put social trust as an input into a production function. You can give people a utility boost from having a job, and call it the "dignity of work". You can model the alienation of arm's-length market transactions as a transaction cost, and remove this cost for long-term informal contracts. Almost anything, except for some fairly exotic behavioral biases, can be dealt with using the standard tools of utility maximization and cost minimization that every econ student learns in school.

So why do economics classes, papers, and policy recommendations so rarely include any of these things?

Part of it might just be an accident - or maybe two accidents. Non-market goods like love, trust, connectedness, patriotism, social obligation, fairness, dignity, etc. is usually really easy to put into models - too easy, in fact. Assuming the existence of something really important but really simple is not intellectually or mathematically impressive. It might be the most important thing in the world, but it won't help you look smart for a hiring or tenure committee.

And these non-market goods are also very hard to measure empirically. How do you measure them? What economic data is a proxy for social connectedness, inclusion, or dignity? They'll usually wind up as unobservable parameters, meaning that the data used to evaluate the models will not be super-informative. (Sociologists probably know more about how to do this, but they don't even post their working papers or blog very much, so who knows what they're up to?)

Social preferences, on the other hand, can be devilishly hard to model. Some people do do it, in highly specific, stylized settings (see here and here), and some people do study social preferences empirically (see here and here; thanks to the excellent Ivan Werning for these examples). But they're pretty tough to put into policy evaluation models, and probably too tough for most undergrad classes.

So there are some technical and institutional barriers to putting these things in economic models and in econ classes and textbooks. But I suspect there might be cultural barriers as well. Econ began in an age when severe material deprivation was the main problem afflicting humankind (as it still is in many countries). Economists of the 19th and early 20th centuries correctly grasped that rising material prosperity would override almost any other concern for the destitute masses of the world, if that prosperity became available.

But as soon as countries get past the stage of abject deprivation, non-market goods and social preferences probably become a lot more important. By sweeping these things under the rug, economists are probably A) nudging policymakers to ignore crucially important stuff, B) developing a reputation for being out of touch, and C) opening the discipline to allegations of irrelevance from people who want to see econ replaced with other ways of analyzing the world.

Interestingly, I think a lot of the progress in the public discussion of these things has come from conservative-leaning economists, whose free-market bona fides probably give them the political cover to openly discuss things non-market goods like social trust. That's a positive development, but economists on the other side of the ideological spectrum should probably get in on that game. To their credit, left-leaning economists often do discuss inequality aversion, but this is rarely translated into formal models or official policy advice.

Anyway, I think Arnade has a point, and economists should listen - even though I'm not sure what exactly needs to be done.


  1. The economics of happiness deals with some of this, and has been extensively studied empirically now for several decades, although there are not much in the way of theoretical models about it. Of course, some would say that all that is being studied in this field is cardinal utility, not all these components such as social connectedness, and so forth, per se.

    Barkley Rosser

  2. A hidden dimension in prime-age, non-college educated male labor market drop-out -- usually left out?

    "The Crips and the Bloods" and my gang, American raised (former) taxi drivers, would enthusiastically work for $200 a week -- one hundred years ago when it would have been understood that that was the best a, then, much less productive economy could pay low skilled labor: the best in history, indoor plumbing, etc.

    They, we, won't work for $400 a week today. $800, yes.

    The money is there. The min wage was $440 in 1968 at half today's per capita. $600 a week would transfer 5% of income from the 55% who now take 90% to the 45% -- not counting push-ups. Push-ups: supermarkets pre-Walmart used to pay $800.

    Beautiful thing about collective bargaining is that (in the US make that "would be that") labor knows it has squeezed the most out of the consumer for the era it habitates -- and subjectively on top of the world.

    But as long as nobody else talks about re-unionization (as the beginning and the end of re-constituting the American dream) -- nobody thinks it is possible to talk about ...
    ... or something.

    Easy as pie to make union busting a felony in our most progressive states f(WA, OR, CA, NV, IL, NY, MD) -- and then get out of the way as the first 2000 people in the many telephone directories re-define our future.

  3. Good points but economists do not entirely push those things under the rug. Even though Dani Rodrik argues that economists have ignored ideas (social preferences), this is not entirely correct. They simply rely on their traditional axioms as sufficient proxies which they turn out not to be! Take the way Zingales takes to factor in culture (norms) in his work on capitalism and competition.

  4. In more simple terms, we have traded American manufacturing jobs to China in exchange for money and moved some of the suffering of the Chinese workers to the US. The money has gone to Wall Street, while the suffering has gone to the American workers.

    In the aggregate, there is more money, lower prices, more jobs and less suffering. People, however, do not suffer in the aggregate, they suffer as individuals and macro does not measure it.

    I think that it is exactly this suffering that is driving this election, so measuring it may quickly become relevant.

  5. When is the last time you heard an economist talk about the need to include strong interpersonal relationships in GDP?

    Stiglitz et al. "Mismeasuring Our Lives"

  6. Anonymous12:59 PM

    I wonder how much of these "non-market" goods could actually be modeled as market goods, or included as part of utility in a rigorous way if people simply backed away from the idea that utility is wealth primarily and other things only secondarily.

    For example, trust, respect, friendship, the pride of others, etc are all different forms of non-monetary good, but that doesn't mean there can't be a market for them. Being respectful towards others, being trustworthy, being loyal, these things have costs associated with them and the benefits can be modeled as a non-monetary good like other signalling mechanisms are. You can trade them too, it just isn't zero-sum (it isn't purely endogenous either - if I give you a public sign of my respect, the affect on the respect others have for me is based on their existing level of respect for you). Political science probably has more call to study this sort of thing, but perhaps it should really be folded into econ models.

    Gift-based economies seem like the place to start for this type of model. And in fact, I think there's always been a weakness in models of gift-giving that assume receiving a gift worth $X places a debt of ~$X on the receiver. The monetary value clearly has some effect, but there's a minimum price associated with the gesture in the first place, which could be modeled as respect. In fact, there are many occasions where the debt incurred by receiving a gift is discharged by an appropriately-formal thank you.

  7. Well, some of this is buried in the now much-studied economics of happiness. Many would argue that the latter is really a revival of cardinal utility, but without any specification of how these various elements such as social connectedness precisely enter in. Indeed, that has been much of what the research agenda of happiness economics has been about, just how strong is this or that non-marketed element in our happiness or utility functions.

    Barkley Rosser

  8. One of your best.

    Barely sub sub-text: Economists should start acting like the social scientists they are, gathering far more empirical data on these topics they ignore. Self-reported, social-science-style surveys -- the primary or at least obvious method for gathering -- are obviously problematic, but there are many methodologies for minimizing that curse that econs could learn and practice.

  9. Actually Noah, there has been rather a lot of work in this area over the past decade. Much of this is empirically driven, reflecting the fact that many of the elements of the utility function not captured by the market are becoming more empirically tractable as measures improve. I suspect that this work gets a higher profile among economists working in public policy than in academia. This is partly driven by the fact that most of the policy issues faced by the governments of developed countries involve the provision of goods and services not captured well by market prices.

    A good example is the OECD's work on well-being (by which they clearly mean something very close to utility). They explicitly mention most of the elements you cite above (social contact, quality of work, meaning and purpose in life) as part of their framework. If you dig round a little you'll find quite a lot of examples of attempts to both measure these outcomes, and apply them to policy issues:

    The chief economist of the New Zealand Treasury recently released a working paper putting an OECD style framework into a formal economic model. The working paper does more or less what you describe, and can be found here:

    For a more academic approach, French economist Mark Fleurbaey is a good starting point. See for example:

    Finally, its worth noting that interest in these issues certainly drives a lot of the recent work on subjective well-being (i.e. life satisfaction etc) and modern stated preference (e.g. Kimball, Heffetz, and Benjamin, 2014 Even with a very large caveat around some of the methodological issues associated with these sorts of measures, the value of being able to observe the utility function directly outweighs the challenges.

    Conal Smith

  10. Anonymous12:06 PM

    Have you tried looking at research from other social sciences?

  11. Awesome, insightful post.

  12. ken melvin5:06 PM

    Demand stems from want and need. Marketing is all about getting consumers to want. To have the wants and not the wherewithal leads to all kinds of problems.

  13. Back in the 1960s I remember Ashley Montagu, the anthropologist, criticizing American feminists. He argued that what women wanted was cultural respect, not better pay, better jobs, and so on. Ignoring the mansplaining, he might have had a point, but the glory of the US was that it wasn't about you. It was about your money. I think Ted Morgan, a French duke who became an American citizen, noted that the American dollar was one of our greatest inventions in that it ignored class and all those other distinctions that the old world was so fond of.

    I agree that economists should consider things besides money. It pays to factor in risk. An $8 an hour job with a guaranteed 40 hours a week might be worth more than a $20 an hour job where you often get 20 hours, but frequently get 10. Economists talk about risk in pricing investments, but not in employment or purchasing things like living quarters. A lot of what money can buy is a lower level of risk. A few million spent lobbying Congress can get you that $100B credit line when you really need it.

    All that community, meaning, purpose stuff is just a cheap way for the upper class to pretend to give those beneath them something without actually having to give them money.

    1. But businesses can't get Congress to fund investment because conservatives have gone extreme Milton Friedman.

      Friedman opposed the policies of the 20s-60s for investing too much in too reliable electric power, too much telephone service, too much enabling of home building with roads, water, sewer, electric, telephone, schools built to serve homes yet to be built. Those cost too much in his view, paid too many workers too much, driving demand for more costly investing.

      Friedman started winning with Reagan's election, so we have the triumph of Friedman over Galbraith, slow grow, high economic inefficiency, vs high growth and high economic efficiency.

      Efficient allocation of factors of production means zero profits. As Galbraith notes, in the 60s, managers and employees planned the corporate spending for themselves which mean high growth to provide opportunity, but high growth means zero returns to shareholders, or at least the barest minimum. And paying taxes was not a problem, or rather, high tax rates made growth planning easier in that paying workers to grow the economy cut tax bills more than profits, and when taxes were paid by businesses who refused to invest, that paid for government high risk R&D or teaching future employees to read, write, think, do science, etc.

      Galbraith understood what it took to create a good society and circa 1990 knew what was wrong and how it was going to end in weeping.

      Even Krugman dismissed Galbraith, siding more with Friedman in grabbing hold of "theory" without foundation and proof.

  14. I wouldn't blame the economists so much as the social psychologists and the other social sciences.

    As the recent replication problems indicate even the most basic measures one would need to try and model this stuff aren't particularly solid.

    However, I would count various work studying how the availability of alternative partners affects marital happiness and the like to be attempts to make headway here.

  15. If one looks at the other social sciences what one will get is the happiness research, also referred to as life satisfaction research as mentioned above. Veenhoven who runs the most in-depth and updated international data base ouf of Rotterdam is indeed a social psychologist. The sociologists are all over this, but so are the psychologists and political scientists. Among psychologists interested in economics, Daniel Kahneman has done a lot of work on this, some of the most rigorous and important.

    The founder of this in economics was Richard Easterlin, dating back to a famous paper that was published in 1974 as a book chapter because no journal would publish it. It was another 20 years before the subject got going again, and is now a major industry in several disciplines, well past the cottage stage. Easterlin's original finding that people care much more about their relative income standing than their absolute one continues to be controversial and debated and studied.


  16. BTW, it does matter whether you ask people if they are "happy" or "satisfied," with the former appearing to reflect more moment to moment moods, whereas the latter involves longer term views. There is roughly an 85% correlation between the two, which is high, but not perfect. One variable that differs is indeed relative socioeconomic position, which is much more important for life satisfaction than it is for moment to moment happiness.

  17. The reason is, conservatives, with the help of progressives have erased them from economics.

    "When is the last time you heard an economist talk about the need to include strong interpersonal relationships in GDP?"

    Well, FDR dwelled in the need for work to provide respect and dignity in his 1935 State of the Union:

    "But the stark fact before us is that great numbers still remain unemployed.

    "A large proportion of these unemployed and their dependents have been forced on the relief rolls. The burden on the Federal Government has grown with great rapidity. We have here a human as well as an economic problem. When humane considerations are concerned, Americans give them precedence. The lessons of history, confirmed by the evidence immediately before me, show conclusively that continued dependence upon relief induces a spiritual disintegration fundamentally destructive to the national fiber. To dole our relief in this way is to administer a narcotic, a subtle destroyer of the human spirit. It is inimical to the dictates of a sound policy. It is in violation of the traditions of America. Work must be found for able-bodied but destitute workers.

    "The Federal Government must and shall quit this business of relief.

    "I am not willing that the vitality of our people be further sapped by the giving of cash, of market baskets, of a few hours of weekly work cutting grass, raking leaves, or picking up papers in the public parks. We must preserve not only the bodies of the unemployed from destitution but also their self-respect, their self-reliance, and courage and determination. This decision brings me to the problem of what the Government should do with approximately 5,000,000 unemployed now on the relief rolls."

    I grew up when Galbraith and Friedman were arguing, Galbraith for the corporate planned economy driven be government policy, and Friedman arguing for a totally soulless profit driven economy where everyone sought to maximize profit without spending a dime more than needed.

    Galbraith argued part of his case 50 years ago, back when there was so much demand for workers that wages rose driving up prices, and basically the fear was of way too much growth. Friedman argued for policies to reduce growth.

    30 minutes of Galbraith:

  18. While Noah is being slow about approving stuff here, I shall add a bit more to what I have said earlier.

    So, just to be clear, in the happiness/life satisfaction literature, now enormous across several social science disciplines, including economics where it arguably started with Easterlin's 1974 paper, pretty much every conceivable non-market variable one can measure has been thrown into various empirical pots to test how it affects this "happiness" or "life satisfaction," with a short list of suspects including health, marriage, religious belief and practice (and type of religion), unemployment status (arguably market related), age, gender, numbers of organizations one belongs to, one's level of "generalized trust" (often identified with "social capital" as is the number of organizations one belongs to), nationality, socioeconomic status, and a lot more.

    So, this is hardly a case of economists having "left out a lot of stuff," although this stuff does not show up in standard textbooks for sure.

    Barkley Rosser

    1. Chris Purnell12:54 PM

      The idea that this policy will make 'me' happy appears fanciful. We can be pretty sure that something will have a negative impact- catastrophic illness, divorce, persecution, extended unemployment, feebleness from age based disability, unmitigated loneliness and so on. But what makes 'me' happy? absence of these hindrances? It all feels trite.

  19. Sure you can find a utility function that takes any variables as inputs and can still be plugged into any model. But can you plug any utility function into these models?

    What if my utility function does not map to Hausdorff space? Are there any models that can actually handle this (an honest question)? What if you can't even impose a Topology?

    1. If you cannot impose a topology, Gilberto, then you are not a real man and should go to the Econ Job Market Rumors site to express your anger, :-).

    2. On a more serious note, we know that there are all sorts of preference structures that do not even allow for utility functions at all, with lexicographic ones being the most famous. The problem is not mapping into Hausdorff or other spaces, it is measuring in any reliable way these non-marketed variables.

  20. AnalyticalYogi4:42 PM

    First of all, I am so stoked that there is a Can't Buy Me Love thumbnail at the beginning of this post.

    Second of all, I am just a mere lower division undergraduate econ student and I have already been reminded to consider these non-monetary factors in my economics classes. I think that it is very important definitely, and hope that econ students and rising economists continue to remain aware.