Skip to content
 

“One of the easiest ways to differentiate an economist from almost anyone else in society”

I think I’m starting to resolve a puzzle that’s been bugging me for awhile.

Pop economists (or, at least, pop micro-economists) are often making one of two arguments:

1. People are rational and respond to incentives. Behavior that looks irrational is actually completely rational once you think like an economist.

2. People are irrational and they need economists, with their open minds, to show them how to be rational and efficient.

Argument 1 is associated with “why do they do that?” sorts of puzzles. Why do they charge so much for candy at the movie theater, why are airline ticket prices such a mess, why are people drug addicts, etc. The usual answer is that there’s some rational reason for what seems like silly or self-destructive behavior.

Argument 2 is associated with “we can do better” claims such as why we should fire 80% of public-schools teachers or Moneyball-style stories about how some clever entrepreneur has made a zillion dollars by exploiting some inefficiency in the market.

The trick is knowing whether you’re gonna get 1 or 2 above. They’re complete opposites!

Our story begins . . .

Here’s a quote from Steven Levitt:

One of the easiest ways to differentiate an economist from almost anyone else in society is to test them with repugnant ideas. Because economists, either by birth or by training, have their mind open, or skewed in just such a way that instead of thinking about whether something is right or wrong, they think about it in terms of whether it’s efficient, whether it makes sense. And many of the things that are most repugnant are the things which are indeed quite efficient, but for other reasons — subtle reasons, sometimes, reasons that are hard for people to understand — are completely and utterly unacceptable.

As statistician Mark Palko points out, Levitt is making an all-too-convenient assumption that people who disagree with him are disagreeing because of closed-mindedness. Here’s Palko:

There are few thoughts more comforting than the idea that the people who disagree with you are overly emotional and are not thinking things through. We’ve all told ourselves something along these lines from time to time.

I could add a few more irrational reasons to disagree with Levitt: political disagreement (on issues ranging from abortion to pollution) and simple envy at Levitt’s success. (It must make the haters even more irritated that Levitt is, by all accounts, amiable, humble, and a genuinely nice guy.) In any case, I’m a big fan of Freakonomics.

But my reaction to reading the above Levitt quote was to think of the puzzle described at the top of this entry. Isn’t it interesting, I thought, that Levitt is identifying economists as rational and ordinary people as irrational. That’s argument 2 above. In other settings, I think we’d hear him saying how everyone responds to incentives and that what seems like “efficiency” to do-gooding outsiders is actually not efficient at all. The two different arguments get pulled out as necessary.

The set of all sets that don’t contain themselves

Which in turn reminds me of this self-negating quote from Levitt protoge Emily Oster:

anthropologists, sociologists, and public-health officials . . . believe that cultural differences–differences in how entire groups of people think and act–account for broader social and regional trends. AIDS became a disaster in Africa, the thinking goes, because Africans didn’t know how to deal with it.

Economists like me [Oster] don’t trust that argument. We assume everyone is fundamentally alike; we believe circumstances, not culture, drive people’s decisions, including decisions about sex and disease.

I love this quote for its twisted logic. It’s Russell’s paradox all over again. Economists are different from everybody else, because . . . economists “assume everyone is fundamentally alike”! But if everyone is fundamentally alike, how is it that economists are different “from almost anyone else in society”? All we can say for sure is that it’s “circumstances, not culture.” It’s certainly not “differences in how entire groups of people think and act”–er, unless these groups are economists, anthropologists, etc.

OK, fine. I wouldn’t take these quotations too seriously; they’re just based on interviews, not careful reflection. My impression is that these quotes come from a simple division of the world into good and bad things:

– Good: economists, rationality, efficiency, thinking the unthinkable, believing in “circumstances”

– Bad: anthropologists, sociologists, public-health officials, irrationality, being deterred by repugnant ideas, believing in “culture”

Good is entrepreneurs, bad is bureaucrats. At some point this breaks down. For example, if Levitt is hired by a city government to help reform its school system, is he a rational, taboo-busting entrepreneur (a good thing) or a culture-loving bureaucrat who thinks he knows better than everybody else (a bad thing)? As a logical structure, the division into Good and Bad has holes. But as emotionally-laden categories (“fuzzy sets,” if you will), I think it works pretty well.

The solution to the puzzle

OK, now to return to the puzzle that got us started. How is it that economics-writers such as Levitt are so comfortable flipping back and forth between argument 1 (people are rational) and argument 2 (economists are rational, most people are not)?

The key, I believe, is that “rationality” is a good thing. We all like to associate with good things, right? Argument 1 has a populist feel (people are rational!) and argument 2 has an elitist feel (economists are special!). But both are ways of associating oneself with rationality. It’s almost like the important thing is to be in the same room with rationality; it hardly matters whether you yourself are the exemplar of rationality, or whether you’re celebrating the rationality of others.

Conclusion

I’m not saying that arguments based on rationality are necessarily wrong in particular cases. (I can’t very well say that, given that I wrote an article on why it can be rational to vote.) I’m just trying to understand how pop-economics can so rapidly swing back and forth between opposing positions. And I think it’s coming from the comforting presence of rationality and efficiency in both formulations. It’s ok to distinguish economists from ordinary people (economists are rational and think the unthinkable, ordinary people don’t) and it’s also ok to distinguish economists from other social scientists (economists think ordinary people are rational, other social scientists believe in “culture”). You just have to be careful not to make both arguments in the same paragraph.

P.S. Statisticians are special because, deep in our bones, we know about uncertainty. Economists know about incentives, physicists know about reality, movers can fit big things in the elevator on the first try, evolutionary psychologists know how to get their names in the newspaper, lawyers know you should never never never talk to the cops, and statisticians know about uncertainty. Of that, I’m sure.

108 Comments

  1. MattW says:

    I majored in economics in college, and the way I came to look at the rationality assumption is kind of like the US economy at full employment. At some point in US history anywhere between 3 and 7 percent unemployment rates have been called full employment, and the US economy was on the PPF curve. I think people are on the rationality frontier most of the time for most things. Of course if someone looks around they will always be able to find a little excess irrationality or unemployed resources.

  2. guest says:

    I worked in a top 3 American econ department pre-grad school, and have worked with many economists as a statistician. I think the answer is simple. Economics as practiced and taught, at least in this country, attracts a disproportionate number of arrogant, self-centered people. Obviously this isn't true of all, but the distribution of these traits among economists is very skewed towards arrogance and self-centeredness as compared to the population at large.

    Add to that all the training in fancy models based upon laughably unrealistic, but strongly ideologically defended assumptions of "rationality" (not to mention perfect information and all the rest) that lead many students to be overly sure of the conclusions derived from these laughable assumptions, and it seems entirely unsurprising to me that so many economists find logically inconsistent different ways of denigrating those who disagree with them as either irrational, having no true understanding of the world (you can't work with PDEs, you just wouldn't understand), or being illogically less self-centered than said economist and too naive to realize that most other people agree that the best way to be is to be arrogant and selfish like a disproportionate number of economists.

    • guest says:

      This isn't just true of economists.  This is true of ANY graduate program.  I did a Ph.D. in math at a top five institution and those guys were just as arrogant; the worst arrogance being displayed by the professors.  The only difference between economics and math is that mathematicians don't talk about things that affect everyone, so no one gets pissed at them for being assholes.

      • Guest says:

        Arrogant, self-centered: Instead of this being true of any graduate program, I would posit that this is positively correlated with the status of the PhD program. Higher status departments, on average, attract folks (faculty and students) who think more highly of themselves.

      • Guest says:

        Arrogant, self-centered: Instead of this being true of any graduate program, I would posit that this is positively correlated with the status of the PhD program. Higher status departments, on average, attract folks (faculty and students) who think more highly of themselves.

  3. sol says:

    I think you're confusing "efficiency" and "rationality". Both have precise definitions in economics and Levitt and other economists usually use them in a consistent way. People are considered "rational" if their pairwise preferences are logically consistent with one another (eg. a rational person cannot prefer apples to oranges to bananas to apples). An economic system is "efficient" if resources in that system are allocated such that total value in the system is maximized (eg. it would inefficient to give your soccer-loving child and baseball-loving child a baseball and soccer ball, respectively, for xmas). Levitt's quote is describing how economists evaluate systems based on efficiency rather than some other moral system (eg. a religious right and wrong). But this has nothing to do with rationality. I'm pretty sure that Levitt would tell you that a cultured and religious person is rational, so long as the choices they make in life are consistent with one another. Levitt saying economists prefer to measure systems in terms of efficiency rather than morality is like saying I prefer to measure goodness of fit with the mean-cubed error rather than mean-squared error. It's just a different metric.

    I think the contradiction you're setting up only works if you interpret "rationality" to mean something different than how it's used by economists (a fair mistake, given that they picked a common word for a fairly technical condition).

    • Sol

      I agree that efficiency != rationality, but I don't think this distinction resolves the puzzle of arguments 1 and 2 above.

      • Eugene Yun says:

        Andrew – Argument 1 ("people are rational") is a commonly uses and accepted assumption in economics. Argument 2 ("people are irrational and need economists") is not much less commonly used and accepted – at least as practiced by more thoughtful economists. Much good economics discusses an different version Argument 2, which we can call Argument 2A = "systems and/or institutions are inefficient and need economists". Within the inefficient systems, the people are behaving rationally. Therefore, what needs to be improved are the systems and institutions. Economists rarely try to 'improve' people. That job is better left to other professionals (teachers? psychologists? spiritual leaders? self-help gurus?…).Economic engineering is difficult and economist often get it wrong, but the best economics usually comes in the form of attempts to improve systems and institutions. 

      • Eugene Yun says:

        Andrew – Argument 1 (“people are rational”) is a commonly uses and accepted assumption in economics. Argument 2 (“people are irrational and need economists”) is not much less commonly used and accepted – at least as practiced by more thoughtful economists. Much good economics discusses an different version Argument 2, which we can call Argument 2A = “systems and/or institutions are inefficient and need economists”. Within the inefficient systems, the people are behaving rationally. Therefore, what needs to be improved are the systems and institutions. Economists rarely try to ‘improve’ people. That job is better left to other professionals (teachers? psychologists? spiritual leaders? self-help gurus?…).Economic engineering is difficult and economist often get it wrong, but the best economics usually comes in the form of attempts to improve systems and institutions. 

    • Jesse says:

      It is a matter of degree.  Efficiency, to the extent that they describe it in terms of self-sufficiency, presumes rationality. 

      Efficiency exists only in simple models, and idealizations.     You know, the sort of things that Glenn Beck can remember.

  4. Manuel says:

    I am an economist, but not a very successful or orthodox one (so maybe I don't qualify as 'true' economist). In my humble opinion, the key to solve this apparent paradox is that economists perceive themselves as 'experts' (in an elitist interpretation of the word). Thus, they apply the 'economic view' to everything except themselves. As experts, they are not a subject of inquiry. That's what is behind the perceived aloofness of economists. (Of course, it is very likely that aloof individuals self-select themselves to become economists).

  5. Manuel says:

    I am an economist, but not a very successful or orthodox one (so maybe I don’t qualify as ‘true’ economist). In my humble opinion, the key to solve this apparent paradox is that economists perceive themselves as ‘experts’ (in an elitist interpretation of the word). Thus, they apply the ‘economic view’ to everything except themselves. As experts, they are not a subject of inquiry. That’s what is behind the perceived aloofness of economists. (Of course, it is very likely that aloof individuals self-select themselves to become economists).

  6. Eamon says:

    My feeling is that number 1 is for thinking about market outcomes and number 2 is for thinking about government policy.

  7. Paul R. Pudaite says:

    Perhaps there's some conflation between the existence of an equilibrium and its achievement?

  8. Anonymous says:

    I don't really see a contradiction between 1 and 2. Number 1 is at an individual level showing that actions taken by people/entities that may seem irrational or confusing, actually have a logic to them. These are stories that buttress the argument that people are fundamentally rational. Number 2 (or at least the examples you have given for number 2) is about systems that have evolved or been created which are irrational or which promote inefficient outcomes. Although, individuals may be rational, there is no guarantee that a system created by groups of individuals will be rational. It also may not be obvious just how or why a complicated system is generating inefficient results. So peeling back layers of complexity and trying to understand the incentives that are motivating behavior seems valuable and in no way contradicts Number 1.

    • Tbwhite:

      The trouble with stories 1 and 2 is that they're opposite, but either can be used at will in just about any example.

      Here are some examples:

      1. As an economist, I can assure you that "Moneyball"-type stories about sports are bogus. Baseball managers want to win, and there's no way that a good strategy will sit around for so long like the proverbial $20 on the sidewalk. If you look carefully at the record of Billy Beane, he just got lucky,
      2, "Moneyball" is a great illustration of how economics-style reasoning can solve problems. By thinking like an economist, Billy Beane realized efficiencies that other managers couldn't see.

      1. As an economist, I can assure you that there's nothing special about Wal-Mart. If Wal-Mart hadn't come along, some other company would've followed the same rational strategy.
      2. Wal-Mart succeeded by following the principles of economics etc. etc.

      1. It seems weird that they charge so much for candy at the movie, but as an economist, I will assume this is rational and then figure out why it's actually a good business decision despite how it might seem.
      2. It's irrational that movie theaters charge so much for candy. If only they would think rationally like economists, they'd make some money.

      1. Person X or Business Y seems to be doing something irrational. But as economists, we respect people's decisions. Actually, that wacky-seeming decision is perfectly rational once you look at the big picture.
      2. The reason why Person X or Business Y is losing money is that they are not rationally adapting to circumstances. If only they would look at their situation in a cool way, as an economist, they could do much better.

      Etc.

      • Anonymous says:

        So, of the four examples you just gave, the Moneyball example seems to be the most well-developed in your mind, so if you'll humor me I'll pick on that one.

        A foundational problem in economics is something called the "agency" problem. What it basically says is people respond to incentives! I'm kidding. What it basically says is that individuals within an organization can be optimizing their personal outcomes in such a way that the organization suffers.

        The chief lesson in Moneyball, although it's been about 5 years since I've read it, is that most members of baseball scouting organizations were not judged on the long-term win/loss record of the team, but rather by the short-term defensibility of their selections. That led them to the herd mentality, the "you're selling jeans" problem, the over-reliance on flashy skills like defense over unflashy skills like OBP, and so on. There is no tension between the "people act rationally" argument and the "there was a free $20 on the ground and only Billy Beane saw it" argument. No one picked up the $20 because there were negative consequences to stepping out of their column. Billy Beane used a risky move which was optimum for his particular place, time, attitude, and skill-set, and which happened to pay off for him. It was also better for the organization.

        Basically, your main point about "argument 2" being at odds with the notion of rational behavior is just, not true. There is literally no axiom of microeconomics which claims an organization will act efficiently by default — and more importantly — that the person in charge of structuring the organization has his personal incentives aligned with those of the organization. 

        That's why it's pretty easy for Levitt and Dubner to speak — as unmotivated (except for their book deals) third parties — about the various ways in which individual behavior is sub-optimal by an organizational standard.

        In other words:

        1. People behave rationally, if you just think about it
        2. People need third party observers, with their lack of misaligned incentives, to show them how to be more efficient. And economists like to write books about that kind of stuff.

        Paradox resolved!

        • Acslater:

          I agree with everything you wrote.  However, it's still true that in any given example a pop-economist gets to pick story 1 or story 2.  I don't know enough about baseball to know which of the two stories is more appropriate in the Moneyball case–I'm guessing it's some combination of the two.  Either story 1 or story 2 could make sense–that's one reason pop-economics is so appealing–but it's a little disturbing to me how confidently a pop-economist will state one of the stories with complete confidence, seemingly not recognizing that the other story could have been picked instead.

          • J. says:

            Andrew:

            In my opinion, what makes these two statements potentially reconcilable is the difference between "equilibrium" behaviour and "out-of-equilibrium" behaviour.

            1) states that in equilibrium, irrationalities cannot persist, whereas 2) seems to me to refer to how equilibrium arises (learning/evolution/market forces drive irrational behaviour out of the market).

            Whether these statements are true is a different question, but they do not seem inconsistent to me, and the Moneyball example clearly illustrates why: Billy Beane exploited out-of-equilibrium behaviour to make a profit (principle 2)), but if he had not someone else would have done so, and once the equilibrium is reached, all remaining actors act rationally (principle 1)).

            • J:

              I agree completely.  Again, my problem is not with the use of arguments 1 or 2 but rather with what seems to me as the arbitrariness of the choice, accompanied by blithe certainty in its correctness.  This looks more to me like ideology than science.

      • Anonymous says:

        I think maybe my disagreement is with your description of story #2. I think it is more accurate to say that story #2 is not that people are irrational, but rather groups, organizations or systems can be irrational, and an economist can help diagnose the cause of the irrational behavior. Again, the difference is that story #1 is about individual behavior, story #2 is about collective behavior. What makes these stories compelling is that groups of individuals acting rationally can cause collective irrational behavior. I guess it feels more like two sides of a coin, rather than a genuine conflict.

        To use the Moneyball example story #1 would be something like: Why is Billy Beane building a beer league softball team instead of a major league baseball team ? When Billy Beane signed a light hitting catcher and made him a first baseman it seemed irrational, explaining that he did it because he thought OBP was undervalued would make it an example of story #1.

        An example of story #2 would be if after his Beane's success with the A's someone wrote an article asking why isn't there more innovation in professional sports ? Then you could make an argument that the unique level of public scrutiny of virtually every decision in professional sports leads to a herd mentality. That yes, while managers want to win, they also want to keep their jobs. Given that few win, you'll do what you can to keep your job as long as possible when not winning, and that means avoiding controversial decisions that make you a lightning rod for public criticism.

      • Jb_harshaw says:

        BINGO.

        Use #1 and #2 selectively as you have demonstrated and you have an "unfalsifiable" set of tools with which to "prove your expertise."

  9. Radford Neal says:

    I think the idea is that (1) applies when people are "acting naturally", and perhaps don't even know why they do what they do, whereas (2) applies when people are trying to apply some highfalutin intellectual framework, and perhaps don't actually even believe what they are saying.  (Economic theory, of course, does not count as such a highfalutin framework.)
    One can resolve the two by observing that it can be rational (1) for people to say stupid things (2), since it may, for instance, help them get laid.  But presumably this works less well if your friends are economists.

  10. Realist Writer says:

    "Statisticians are special because, deep in our bones, we know about uncertainty."

    Can you really say that if you can quantify how likely you are for your estimates to be wrong?

    • anon says:

      As long as you are appropriately uncertain about the uncertainty

    • Jerzy says:

      Isn't this (quantifying how good or bad your estimates are) exactly the "value added" by statisticians?

      Andrew, I remember reading your comments on the 2009 "What Is Statistics?" paper in the American Statistician.
      You and Madigan stated that statisticians traditionally focus on the error distribution, but it sounds like you think we ought to focus more on modeling the mean. But isn't the error distribution where statisticians really "add value" to the scientific process?
      Other scientists can do a decent job of finding apparent signals/trends on their own, but the field of statistics specializes in telling apart apparent signal from mere noise, or at least helping you not put too much faith in something that might be noise.

      • Jerzy:

        The distinction between mean and variance is somewhat arbitrary.  Consider Red State, Blue State, in which our key finding was that income is a better predictor of voting in poor states than in rich states.  You could call this the deterministic part of the model– the coefficient of an interaction in a linear regression–or you can say that what we really were doing was finding structure in the error model.

  11. […] have a fixation on rationality, and this article is a partial expression of that.  My answer to the article is the people are limitedly rational.  That’s vague, and […]

  12. […] How to differentiate an economist from almost anyone else in society — Gelman […]

  13. […] The great rationality paradox of […]

  14. Lu_si_na says:

    I think Sol's comment nailed it. Economists believe everybody is rational, according to their definition of rational (which is not how most people define it, and I don't even agree with the economist premise, but I digress). They also believe people are sometimes inefficient, and that efficiency is always the best thing. Other people believe that efficiency isn't always the best. They aren't being irrational, according to the economist's definition of irrational. They are simply inferior in the economist's eyes.

    Oster made a mess in that quote, but to be fair, I don't think "everybody is fundamentally alike" is meant to imply that everybody has the same preferences or opinions.

    • Lu:

      I don't think Oster was trying to imply that everybody has the same preferences or opinions, I think she was trying to say that the intellectual framework used by economists is different (and better) than that used by other social researchers.  This fits with Levitt's claim that economists think in a fearless way that makes ordinary people uncomfortable.  (This last bit captures the tension between economist-as-populist–unlike everybody else, economists trust ordinary people to make rational decisions–and economist-as-elitist:  unlike everybody else, economists trust ordinary people to make rational decisions.)

      That's fine with me–I have no problem being a populist in some contexts and an elitist in others–but I'm not at all happy with Oster's formulation of "circumstances, not culture."  When applied to the purported differences between economists and everybody else, we're led to the conclusion that "everybody else" gets things wrong because of circumstances, not culture.  I assume the idea is that sociologists, for example, get rewarded not for getting things right but for publishing in the American Journal of Sociology, which in turn is presumably controlled by a cartel of sociology professors who want to publish B.S. (because, perhaps, they are rationally wanting to preserve the value of the crap that they published themselves in the journal in previous years).  So you can break down the "cultural" explanation into a network of "circumstantial" explanations.  I think some economists would love this story–hey, it's an explanation for why sociology is so bad!–while others would distrust it as an implausible gravity-defying story of thwarted rationality.  (It's like the qwerty story:  either a delightful story of how differences can persist because of incentives that point in the wrong direction, or a myth that ignores people's agency and willingness to find solutions to their problems.  In either story, people are locally rational–I think that's the point made by some commenters above–but the stories point in opposite directions.)

      OK, back to the "economists are different from everybody else" story.  We have a story of why sociologists are so deluded (they're stuck in a self-reinforcing network in which the local motivations are to bloviate), and we can make a story of why public-health researchers are so deluded (that's an easy one:  they're paid by governments and nonprofit organizations, so they are not subject to market discipline), and we can make a story of why anthropologists are so deluded (see "sociologists" above, only more so).  But why can Oster be so sure that the economists' perspective (as she sees it) is correct?

      This is not a trivial "gotcha" on my part, nor is it an accusation of hypocrisy.  If you really think different groups are different, and that these differences arise from positive and negative incentives, I'd think you'd want to look carefully at your own incentives to figure out how to get closer to the truth.  It's my impression that sociologists do this:  that, to the extent they believe that group differences can be explained by "culture" (i.e., those interlocking sets of expectations, pressures, and incentives), sociologists do try to xamine their own culture and reflect upon how their work fits in to it.

      P.S.  I'm not trying to pick on Oster here.  She's a young scholar and is excited about her work, and she said some ill-considered things in an interview.  I'm sure that a good interviewer could extract lots of silly and even self-contradictory statements out of me, too.  I'm only banging on this example because I think (and commenters above seem to agree with me) that the view she expresses is widespread among economists and in pop-economics.

      • Jerzy says:

         I can see a charitable "weak form" interpretation of economists' Arguments 1 and 2:
        1) "People are rational" == "It's possible to model people as rational (when such models are appropriate)"
        2) "Economists are rational, others aren't" == "Economic models assume a particular type of rationality, while models in other fields simply use different assumptions, whether or not we all fundamentally believe these assumptions on a personal level"
        No contradictions, no problems. I hope this is what Levitt and Oster's quotes meant.

        But I agree that Levitt and Oster here, at first glance, *sound* like the sensationalist "strong form":
        1) "People REALLY are rational and these models are true"
        2) "Economists actually BELIEVE that people are rational, and they are correct; people in other fields fully BELIEVE other models but they're wrong"
        These are nonsense and lead to the contradictions you mention.

        I wish pop economists would state things better and stop implying the strong form. But… maybe they have rational incentives to be sensationalists :P

  15. […] (Economix) • “The easiest ways to differentiate an economist from anyone else in society” (Andrew Gelman) • China to Dive for Buried Treasures (WSJ) • Phone hacking crisis shows News Corp is no […]

  16. Pete says:

    The point of being open-minded is that both arguments are considered given the circumstances and scope of a topic. Who cares what argument they use – they are essentially hypotheses, right? What matters is the evidence they bring to bear to support the argument.

    Isn't this simply another version of the old joke about wanting a one-armed economist? Defining the circumstances where people are rational and where they aren't is the objective.

  17. […] Gelman discusses a “puzzle that’s been bugging [him] for a while“: Pop economists (or, at least, pop micro-economists) are often making one of two […]

  18. valter says:

    "The trick is knowing whether you’re gonna get 1 or 2 above. They’re complete opposites!"

    Yes, but it is not a trick. And you are not going to know it until you evaluate the arguments in the specific case you are looking at.

    The point is that in many cases (maybe even "most cases", as economists tend to assume) people have good reasons to behave the way they do and it may take some economic and statistical ingenuity to discover what those reasons are – that's #1 above. 

    But in other cases people simply make mistakes – they may have factually wrong beliefs, reason incorrectly or not at all, cannot resist temptations that they will later regret, etc. If the #1 approach above fails to find a reasonable explanation, then it seems reasonable to try approach #2 – until someone gets a bright idea that can explain the apparently "incorrect" behavior using #1. Hence the appearance of flip-flopping. 

    As you wrote: "The two different arguments get pulled out as necessary." (where Levitt would emphasize the "necessary").

    Economists simply tend to stick with approach #1 more than other social scientists. I am not sure whether economists, when in #2 mode, are more interventionists than other social scientist.

    I also think that Levitt is correct in saying that economists are relatively more open to consider "repugnant transactions" – possibly to the point of disregarding valid non-efficiency reasons for their repugnance. I suspect that is because economists are used to defend the profit motive for various actions against many who find such motive itself repugnant – hence develop some repugnance for repugnance arguments.

    (btw, I am an economist, but not a strong fan of freakonomics)

    • Pete, Valter:

      Yes, it's fine for economists (and others) to employ different arguments in different settings.  The problem I have is that when they apply argument 1, they seem to be so sure that argument 2 is irrelevant, and the reverse when they apply argument 2.  And in practice I am not at all convinced they are making a good choice–or, more to the point, if a choice is needed at all.

      Also, the two arguments have strong and contradictory rhetorical baggage.  Argument 1 goes with the populist story that economists are different from other social researchers in that they respect individuals' choices.  Argument 2 goes with the elitist story that economists are different from other social researchers in that they are willing to look at difficult decisions in the face.

      • Pete says:

        Being only a dabbler in economics, I actually find it refreshing that economists consider all possibilities. I don't think it is random, and I generally don't think it is particularly emotional (except for one economist at the NYTimes… ;-)). The truth is never as pretty as the assertions, and usually includes some pieces of both sides to an argument. Having the strongest arguments for or against a particular situation is of much more benefit than the veiled objectivity of the media, anyway.

        I don't think it is a bad thing to assume rationality until you can come up with situations that appear irrational, and then explore them. i think that is the normal order of things with economists.Even the explanations are illuminating.

        I must be lucky as a dabbler, because my experience with economists is that they are typically more willing to explore the possibility that they are wrong… if only to describe the circumstances under which they are right ;-)

        • Pete:

          I like a lot of the work of economists and I agree that it's good for a researcher (economist or otherwise) to consider multiple explanations.

          My problem is with much of pop-economics in which explanation 1 or 2 is taken is an obvious sure thing:  Either we get the claim that it's axiomatic that people are rational and respond to incentives or we get the claim that it's axiomatic that people are emotional and need the help of an economist to solve their problems.

          • jerzysblog says:

            So would you be happier with pop economists if they stated 1 and 2 more subtly, AND/OR always used both?

            "1) WITHIN ECONOMIC MODELS, it's axiomatic that people are rational and respond to incentives.
            2) IN REALITY, we recognize that people don't always use this approach to thinking about their behavior, so sometimes it can be useful to ask an economist for advice.
            In today's report on [topic X], we'll tell both story 1, to explain how people's standard behavior might be modeled as rational, and story 2, to show how taking an economist's approach [has made / could make] the situation even better.
            Or we'll explain why 1's good enough that 2's unnecessary.
            Or we'll point out weaknesses in 1 in this situation and why 2 is needed instead."

            If pop economists actually said it this way, would that resolve your annoyance?

  19. engineer27 says:

    There are two kinds of people: those who divide people into two categories.

  20. Jacob Weaver says:

    After studying both economics and statistics, I am uncertain whether I was rational!

  21. Kh13542 says:

    "I thought, that Levitt is identifying economists as rational and ordinary people as irrational."

    Nope.  That isn't what he's doing. You are conflating "efficient" and "rational" – either an intellectual error or a rhetorical trick.  Levitt says that people have reasons for finding things repugnant which are efficient.  If you asked, I'd bet Levitt would also tell you he finds some things repugnant which are efficient.  Efficiency is far from the only virtue, but it is the one that economic analysis is able to handle.  Other virtues fall under "preferences" and are taken as given most of the time.  People like Levitt like to examine preferences, finding that some are vulnerable to economic analysis, but that's a different issue. 

    You have misspoken in claiming Levitt is identifying "ordinary" people as irrational.

    • Kh:

      You're picking up on half of the pop-economists' argument.  Indeed, the Levitt quote in my blog above fits in perfectly with story 2:  that, by virtue of their inclination and ability to find efficiencies, economists can solve problems that other people can't.  This story is consistent with the idea of an organization hiring an economist to develop a new business plan.  This is economics as operations research.

      But pop-economics also has story 1, as exemplified by the Oster quote:  that economists view ordinary people as rational optimizers (or some close approximation to that ideal).  This doesn't fit in so well with story 2.

      As various commenters noted above, it's fine for story 1 to apply in some settings and story 2 to apply in others.  My problem is that when pop-economists are telling story 1, they seem to forget story 2; and when they're telling story 2, they seem to forget story 1.

      Just for example, an example that relates to my own research, Levitt has repeatedly written that it's irrational and indeed silly to vote, that the only reason to do it is for such reasons external to politics as "your wife will love you more if you do it."  That's an explanation of type 1.  That's fine–I have my own "story 1" reason why I disagree, but that's a research disagreement (I'm guessing that if Levitt were to dispute my reasoning on that one, he'd argue that I set the parameter alpha too high in my model).  My problem is that when he makes his claim, he forgets about story 2.  For example, here's a story 2 argument that I find somewhat compelling:  millions of people do vote, maybe we should respect their preferences and model this as rational behavior.  Not only do people vote, they seem to think carefully about how to vote.  (And to simply model this as simply "it's fun" or a taste for voting . . . that's just tautological.  The value of "story 2" explanations is that they go beyond tautology and the imputation of taste and take the assumption of rationality and purposive decision-making more seriously.) 

      Anyway, my point here is that once story 1 came out, there was no serious consideration of story 2.  Conversely, when story 1 comes out, story 1 gets forgotten.  So when a too-clever-by-half economist tells us why it's actually rational to buy lottery tickets or explains some seemingly-irrational behavior, then efficiency is no part of the argument.  We're left with the ludicrous (to me) conclusion that it's irrational to vote but rational to buy lottery tickets.

      P.S.  Thanks to you (and others) for the thoughtful comments, which among other things help me more clarify my not-fully-thought-out ideas.

  22. Anonymous says:

    I think the main distinction here is between "rational" and "cognitively impaired."  The economic mindset is, as you point out, that actors are rational.  Taking that rationality seriously takes all the type 1 stuff.  But one can be rational and still be cognitively impaired — you know what you want to do, but the means by which you do it are inefficient, or, in the extreme, actually counterproductive.  That's they type 2 stuff.  Now, how do pop economists (or even the rest of us) confidently assert that a given situation is really a Type 2 situation?  Well, you can't.  You can assert it,but the history of economics is really the history of people showing that what looks like type 2 behavior is really type 1 behavior that the first guy just missed.

    An interesting example is option pricing.  People priced options in markets for years before Black and Scholes showed that a fairly simple mathematical model captured a lot of the basics of what prices emerged in these markets — eve though none of the players in these markets used these models — at least until Black and Scholes wrote their models.  But there are ways in which actual markets diverge frorm the predictions of these models.  Whether these differences are corrected by a richer model, or by admitting that humans have flawed cognitive capabilities to maximize is a debatre which rages to this day.

  23. Robi Ragan says:

    I agree that economists often play this sort of mental shell game with "rationality" when talking about policy. However, I think your enumerated points do not really describe that behavior. To my reading, the two are in no way in conflict and are simply a reflection of the (new) institutionalist view.

    1) Is the idea that agents choose strategies based upon with their knowledge of their current fitness landscape (institutions broadly defined and other agents' actions), in a way consistent with optimizing or satisficing behavior. The outcomes we see emerge from this system, often look strange until one looks under the hood at the underlying "incentives".
    2) Is the idea that if we modify institutions, agents will choose different strategies, and we will get different outcomes.

  24. […] 3. How popular economics treats rationality. […]

  25. Alezanderclark says:

    i would encourage reading closely B. Bernanke's  Ec 101 textbbok, as i did recently. it is incredibly didactic, a true manifesto of a politcal theory. Since he and Greenspam are in power for some 25 years now, there really is no point in discussing economics – it is enforced by him and the Bureaus.
    Consider that most developed countries subsidize almost everything, yet Bernanke informs us that all subsidies are destructive intrusions in the marketplace, leading to pervasive unintended consequences.   Yet he and the government are themselves the subsidizer of the financial world, to prevent "failure"!  
     

    He also uses metrics such as core  inflation , ignoring energy, real estate prices, cost of medical service or insurance, etc.   Yet  these exempt sectors of the economy are conveniently heavily subsidized by tax loopholes,  free mortgage money from the government, non-profit status of medical schools, reasearch facilities and hospitals.   this goes on and on.   So it appears that in all areas, government has skewed the numbers systemically in such a way that there is no science of Economics.  

  26. Gary Hewitt says:

    This discussion is reminding me of Yoram Bauman's (Standup Economist) schtick when he goes over Mankiw's ten principles of economics – a couple of them are reduced to "people are stupid" and "people aren't that stupid" (and macroeconomics is reduced to "blah blah blah" …)

  27. Michael Bishop says:

    Yes, we can imagine an economist making either of two types of arguments which are in tension with each-other.  The important question is whether they do a good job of deciding which is more appropriate.  We need to move this discussion to a place where people on different sides make contrasting predictions.

  28. TGGP says:

    Both Alex Tabarrok's post on Oster and Esquire's article are no longer at their original urls. I searches Esquire for Oster and found this: http://www.esquire.com/features/ESQ1206BOWEN_206_

  29. m s says:

    The economic rationality of #s 1 and 2 can be explained by the idea that economics is really a "science" of rationalization. 

    #1 Immeasurable explanations of signaling, imperfect information, personal utility functions, group game theories etc. can be added or subtracted at will to rationalize public behavior.  Similarly, more immeasurable concepts can be assumed to rationalize #2, the economist's own ideas about how the public should behave — And since it incorporates more information, (albeit immeasurable assumptions about the motivating psychology) #2 concludes surely this more rationalized rationalization is superior to the rationalization observed in #1.

    Of course, it's not as sexy to be in the same room as rationalization, so why not conflate rationality with the ability of economic theory to rationalize behavior whenever convenient…

  30. Dave says:

    "1. People are rational and respond to incentives. Behavior that
    looks irrational is actually completely rational once you think like an
    economist.2. People are irrational and they need economists, with their open minds, to show them how to be rational and efficient."

    You should read economists like Schelling and Herbert Simon (more of a polymath than pure economist, but he did win the Econ Nobel…) if you want a more nuanced view.  Simon spoke of bounded rationality and satificing, while Schelling described addiction as a struggle between the best interests of the addict in the short term vs long term.  They make Gary Becker and co look like an intellectual infants.

  31. Pietro Poggi-Corradi says:

    "The key, I believe, is that “rationality” is a good thing. We all like to associate with good things, right?"

    This is a type #1 argument, I think.

  32. guest says:

    I think this article conflates making choices with analyzing the reasons why people make choices.  One could resolve the purported inconsistency this article discusses by noting that all of the 1 circumstances involve people making choices themselves, while all of the 2 circumstances involve economists discussing how other people consciously analyze either their own decisions or other decision makers as a group. 

    The inconsistency here only would exist if pop economists believed that everyone consciously computed the rational choice and then made it, not if they only believed there is a rational reason to make the choice that people make.

    I should note that I believe it doesn't matter who is 'right' because it's possible to rationally justify nearly anything, as the sophists pointed out.  Pop economics is a thought experiment to help see things from a different perspective, and is useful in that way, just as pop psychology is.  Those that believe in it as the 'truth' are certainly amusing, but a bit of a straw man. 

  33. Anonymous says:

    I think the following is a consistent set of views that resolves this dilemma:

    1.  People ought to always behave rationally. 
    2.  People sometimes behave
    rationally. 
    3.  Rational behavior theory assumes that people always behave
    rationally, in order to make predictions. 
    4.  Rational behavior theory makes useful
    and reliable predictions. 
    5.  Useful and reliable predictions help people
    behave rationally. 
    6.  Therefore, people ought to use rational behavior theory to guide their behavior.

    If rational behavior theory is treated as truth, then 3 contradicts 2, and 1 becomes meaningless.  So you're correct to reject rational behavior theory as dogma.  But a theory's value should be measured by its utility, rather than its truth.  This theory–though obviously false–has considerable utility, and is thus pretty valuable.

    When economists "assume everyone is fundamentally alike" they don't actually think it's true; they just assume it for purposes of making predictions.  Economists' minds are "skewed in just such a way that instead of thinking about whether
    something is right or wrong, they think about it in terms of whether
    it’s efficient".

    • CM says:

      How do you figure that theories assuming rationality have considerable utility? What is the evidence for this claim? Revealed preference in that the theories are used by economists?

      2. Is not really true, because rationality is defined across an individual's set of preferences, not just the particular preferences a researcher/economist observes. We know (almost all) people systematically violate rationality assumptions in certain situations, therefore it is incorrect to describe their behavior as rational, even if we don't go to the trouble of observing the irrational/inconsistent behavior.

      3. Is true and is the reason 4. fails.

      5. and 6. are then irrelevant.

      • Anonymous says:

        Sure, you can look at the entirety of economic / business literature–theory of supply and demand, pricing theory, game theory, comparative advantage, compensating differentials, etc.–and decide it doesn't make any useful predictions.  I'm not particularly interested in debating the empirical claim that economics makes useful predictions, since it's unrelated to the original post.

        All I take Andrew Gelman to be claiming is that there's an inconsistency between assuming rational behavior and giving counsel.  I think that this framework unravels that inconsistency.

  34. Points 1 and 2 are not complete opposites. You can actually put them together to build a more unified theory: 1) people are rational in how they respond to stimuli, and 2) People who attempt to produce an outcome based on how those rational people will respond to stimuli will often create inefficient mechanisms to achieve the results. This article is basically claiming that simply being a rational individual is the same as being a person who builds predictive models. You can be rational and not build behavioral models, you can be rational and build terrible behavioral models, or you can be rational and build good behavioral models. The economists above are just saying that they are in the latter category (because that is their training), AND that we are all rational. This article's premise a false dichotomy. There is no contradiction between points 1 and 2.

  35. Dan says:

    I do believe you are stating #2 incorrectly. It should read something like "Economists are good at looking at problems in different ways and often provide new information that can show that some behavior that people rationally engage isn't really rational when taking the new information into account." Levitt probably wouldn't accept your statement, but would probably accept mine. Countless things that people used to do were rational based on the information available, but were no longer rational when new information rolled in.

  36. […] Jeremy, check out Andrew Gelman’s hard questions for pop-economics: I think I’m starting to resolve a puzzle that’s been bugging me for […]

  37. Curmodgeon says:

    "Pop economists" are rational and respond to incentives. Behavior that looks irrational is actually completely rational once you think like an economist.

    Celebrity is a bigger concern for 'pop' economists than a traditional research economist (notwithstanding the need to publish), and there's certainly anecdotal evidence that making blanket, controversial statements tends to increase celebrity more than making consistent ones.  Endorsing one argument or the other, as situations permit, can be seen as rational if — as the agencies considerations noted by other posters persist — the pop economists are more concerned about advancing their own welfare rather than economic practice.  As such, you can argue that Levitt is demonstrating economic rationality by not applying rationality (coherence) to his statements.

    ===============

    I disagree that economic rationality (coherent preferences) is, by its nature, "good." The trick is that neo-classical microeconomics is a descriptive discipline, unless you make the subjective assumption that utility maximization is good (and even if you do so, you need to make further subjective assumptions — time discounting, social preferences, etc.) . As such, economists may show people how to be 'rational and efficient', but it's debatable whether they make people 'better.'

  38. […] Maybe not . . . I think I’m starting to resolve a puzzle that’s been bugging me for […]

  39. Anonymous says:

    It's really pretty easy to square this.

    Economists talking about rationality require the concept of rational ignorance.  A rational person shouldn't care about knowing many things, and would even have many opinions about things they knew little about.

    The claim is that a person behaves generally rationally, given constraints.  That doesn't mean they are correct propositionally about anything…but that they will generally behave in their own self-interest. 

    Making decisions about other-people is NOT generally something that impacts your rational behavior, and it is rational in most cases to be ignorant about that.  Mostly, economists want folks to be less ignorant about the impact of policies they advocate…because there are real costs.  In the aggregate, better understanding might promote less bad policies.  In the individual case, there are no/few costs to bad opinions, and so it is meta-rational to hold them.

  40. guest says:

    I think both arguments start from the assumption that the relevant actors are rational. The second provides a caveat of cases when that assumption fails.

    I think in most cases, with respect to the first argument, when an economist distinguishes herself as more open minded or rational than someone else, that someone else is an outside observer, not the agent. The agent remains rational. In argument number two, its mitigating circumstances that somehow disrupt the rationality of the agent. A rational observer in this case could potentially remove those obstacles.

  41. Kaw says:

    The most important difference between economists and other social scientists is that other social scientists read economists before deciding they are wrong, whereas economists don't read other social scientists before deciding they are wrong.

    This quote; "anthropologists, sociologists, and public-health officials . . . believe
    that cultural differences–differences in how entire groups of people
    think and act–account for broader social and regional trends" is stunning not only for its arrogance but for its ignorance.

  42. Lord says:

    There is some legalism here, deciding on the result and creating the arguments to support it.  Rationality is tossed about like an objective fact when it only exists within the context of facts, principles, and theories of how to arrive at it.  The assumption of rationality allows for globally self consistent solutions that would not exist in its absence.  The world becomes much messier and little can be concluded without it.  Economists want conclusions, too often their own conclusions.  As a result rational means they have a theory about why that is true and irrational means why it shouldn't be.  Irrational arguments are always suspect since the results don't accord with reality but much of reality does not conform to what economists conception of it should be.  Dan Ariely demonstrates humans are deeply flawed in rationality in some specific situations and some have suggested nudges to shift people towards rationality.  Others consider this taking advantage of peoples irrationalities which advertising and marketing do so well.  In the end, economists just aren't interested in irrationality, only in making the world consistent with their conception of rationality.  

  43. Ragnar says:

    Statisticians don't have an innate sense for uncertainty, but they do have the status to make incredible mistakes out of their conviction that they understand uncertainty better than non-statisticians. I'm glad Nassim Taleb illuminated that for me.

    • Ragnar says:

      Could it be that people are rational in the aggregate with individuals largely unaware or indifferent of said rationality?

      No, can't be. Then we can't slam economists.

    • David Shor says:

      Eh. Every statistician is generally fairly familiar with Taleb's points, and has been for decades. Kurtosis generally isn't actually very important outside of finance. 

  44. Rtnlty says:

    Start with the following simple assumption that fits the data perfectly well.
    Economists are a collective moron who has not defined any problems leave alone solve them.

    Fan of Freakonomics – really?! Stab me in eye with a knife. For heaven's sake fan boy go pick up a story book for a change.

  45. […] door to link to individual posts.  A couple of days ago he put up a terrific post called “One of the easiest ways to differentiate an economist from almost anyone else in society.”  He wonders why it is that so many economists can simultaneously believe these two things: […]

  46. Jesse says:

    Excellent expose.

    But there is a simpler explanation. 

    Like some other professions, the economics profession is particularly attractive to those inclined to sell themselves and their integrity, and sometimes shockingly so.   So we might add Economists to the modern pantheon of Politicians, Lawyers and Whores.

  47. matt says:

    nowhere near all economists believe that all individuals are rational (p.s. when people say this, they really mean that agents either satisfy the vNM, Savage, or Anscombe-Aumann axioms since nearly any application of rationality these days is non-deterministic). decision theorists would be out of a job if people were fully satisfied with any of those treatments. Read Allais (1953); Ellsberg (1961); Dekel (1986); Schmeidler (1989); Gilboa & Schmiedler (1989); Gul (1991); Dekel, Lipman, & Rustichini (2001); and Gul & Pesendorfer (2001) for example.

  48. Mark says:

    The easiest answer is that economists (of  all varieties including "pop" "armchair" "amateur" and "professional") often aren't very good at what we should be doing — helping interpret human behavior in various situations.  The situations include factors that influence preference, including cultural norms, which may create a separate set of "rational" behaviors.  The question is not whether incentives matter — because they do — but what creates the various incentives people operate under and how do the effects of those preference creating factors interact.

    You have to understand that because individually and in the aggregate people generally respond to "normal" economic incentives to at least some degree in the fashion that "normal" economics would predict, statistical analysis will almost always show that traditional economic incentives matter.   This fact creates a great deal of prestige for a "science" that seeks rigor through increasing the use of mathematics and statistics to create "precision" in supposed explanations or predictions.  Thus, rather than simply say, as economists should, "___ types of incentives play a role in ___ decision-making", they seek explanations that fit with the economic man — or assume you just can't really understand the behavior from an "economic" (rational) point of view.  Further, through use of more sophisticated statistical techniques, economists can seek to "control" for factors which often cannot be controlled for because of the fluid, interactive and multiple bases for human behavior and often seek unattainable precision in finding answers.

    Thus the "economic" view can ignores much of the reality of human behavior: the role of internal dramas, cultural norms and difficult to isolate or define value choices.  It can also misunderstand the nature of logics (the plural is intentional) and systemic decision-making.  Economics as a discipline seeks to examine one realm of human behavior (or more correctly, originally sought to examine one realm of human behavior) and now is attempting to universalize its methods as the most "scientifically" based among the social sciences.  It uses the patina of rigor created through mathematical explanations of behavior which have been proven effective — to a certain degree — in the physical sciences where "truth" is often seen to reside.   But I strongly suspect that the search for the type of scientific truth or even something remotely similar in the realm of human activity beyond what one must see as general observation and ballpark estimates is unrealistic and unfortunate (see Karl Popper's "The Poverty of Historicism" for a partial explanation of why this will always be the case). 

    This does not imply that economic analysis and statistical analysis is not useful.   But we needs to understand and internalize more clearly the limitations of our tools and our explanations for human behavior and not assume that we can explain everything quantitatively — at least not with "scientific" precision.  We need, moreover, to admit to ourselves that "rational" choices exist within the realm of cultural norms and systemic thinking that do not fit with traditional economic man. 

  49. I think the argument boils down to:

    1. People are "rational" in that they will with perfect information choose the option that they like better (optimize their utility)
    2. People are "irrational" in that they don't have perfect information in all cases if they have not thought through all costs of their decisions (opportunity cost, externalities, etc.) which economists have been trained to do. This leads to non-optimal solutions (from an economists perspective) that economists love to attempt to fix.

    • Dan says:

      David has it right, here. There is no inconsistency there.

      • David, Dan:

        As noted above, my problem with some pop-economics is not with the use of arguments 1 or 2 but rather with what seems to me as the arbitrariness of the choice, accompanied by blithe certainty in its correctness.  This looks more to me like ideology than science.

    • Lord says:

      And if they don't arrive at the economists conclusion it is because they are uninformed, and if they do they are informed.  It couldn't possibly be that it is the economist that is uninformed or misinformed or swayed by biases, or forbid, data.  No, that couldn't be it at all.  

    • Anonymous says:

       > 1. People are "rational" in that they will with perfect information
      > choose the option that they like better (optimize their utility)

      Personally, I will choose:
      1) Chocolate ice cream over strawberry
      2) Strawberry ice cream over vanilla
      3) Vanilla ice cream over chocolate.

      I don't think I am unusual in that, and no one has ever successfully used a "money pump" to empty my wallet or that of any person I know who makes similar choices.  Such very common human behavior, while possible to work around with very sophisticated math, renders the concept of everyday perfect rationality somewhat dubious.

      Cranky

  50. Greg says:

    You guys are overthinking the whole thing. It's a confusion of terms.

    My decisions need not make sense (point 2), but they are always rational (point 1). The problem is that "rationality" was more or less normative until economists got their hands on it. And then economists, being human and having normative opinions, freely use the term "rational" nontechnically, often referring to decisions made by rational actors (remember *all* actors are necessarily rational) that *appear* irrational from a third-party perspective due to incomplete information and calculation costs. You can even argue that efficiency is really a measure of normative societal rationality. Now it has at least three meanings, depending on context. And it's confusing the hell out of everyone, even the people using it.

    Positively rational decisions need not be normatively rational. And we ought to be mindful and explicit about which "rational" we're talking about, or better yet avoid the term entirely except where explicitly technical.

  51. Joshua Lyle says:

    Alternate:
    #1: Most people are pretty rational in "near" mode.
    #2: Economists are pretty rational in "far" mode.

    Or vice-versa. Or something.

  52. Eric Wilhelm says:

    …and you can differentiate a statistician from everyone else because they know when to backtrack (see: post starting at about "OK, fine. I wouldn't take these quotations too seriously…") under the protecting veil of "uncertainty". You forget that hard scientists have the luxury of comparing "alternate universes" (see: control groups). Economists don't. There is only one universe to "observe" that marches on… 

    • ezrazuckerman says:

      Excellent post.  This is not just an issue for pop economics though.  As Arrow (1974, p. 16) wrote, “an
      economist by training thinks of himself as the guardian of rationality, the ascriber of rationality to others, and the prescriber of rationality to
      the social world.”  This quote includes both 1 and 2 ("ascriber" and "prescriber"), and the tension is resolved– but not really– in exactly the way that Andrew suggests– i.e., that the economist is always on the side of (in fact, the "guardian" of) rationality.  And where Arrow is deft in how he handles this tension, the field of financial economics has been… er.. gauche.  All the big shots who preach/insist that the market is efficient such that there are no arbitrage opportunities ("ascriber"), have tried to make fortunes as advisors and/or as principles in funds that engage in the very kinds of arbitrage (indirect "prescriber"– i.e., they are smarter than the market, and are apparently necessary to make it efficient) that should not be possible if markets were efficient (sometimes they have been successful and and sometimes they have been spectacular failures; see Long-Term Capital Management). 

    • Eric:

      Astronomers are hard scientists and their data are all observational.

    • Eric:

      Astronomers are hard scientists and their data are all observational.

    • P.S.  I don't consider my "OK, fine" etc. statement to be backtracking.  I consider it to be politeness.

  53. Eli says:

    Arguments 1 and 2 aren't opposites. Echoing Greg above, I'd say argument 1 is a descriptive claim, whereas argument 2 is normative.  
    1) If you're asking for an explanation for a person's behavior, an economist will look at the incentives and preferences in place and explain why, from a certain point of view, said behavior makes sense.  So people are at least a little rational.
    2) If you then ask whether that person's behavior is optimal, an economist will very often give reasons why they are miscalculating and could do better.  So people are clearly not Perfectly rational.
    Economics is clearly useful in a limited way for answering those kinds of questions.  Whether economists have earned their high opinion of themselves is a totally separate question, though the fact that they allow such confusion and equivocation with respect to concepts as important as normative vs. descriptive Rationality should maybe give them pause.

    • Lord Liege says:

      But if there were more rational approaches they would already have been found and adopted, people being rational, so evidence these approaches are not used is proof they are not more rational.  No $20 bills lying on sidewalks.  Economists don't even believe themselves.   

  54. […] Economists can’t have it both ways:  people can’t be rational while economists are.  (Andrew Gelman) […]

  55. Anonymous says:

    "One of the easiest ways to differentiate an economist from almost anyone else in society is to test them with repugnant ideas. Because economists, either by birth or by training, have their mind open, or skewed in just such a way that instead of thinking about whether something is right or wrong, they think about it in terms of whether it’s efficient, whether it makes sense. And many of the things that are most repugnant are the things which are indeed quite efficient, but for other reasons — subtle reasons, sometimes, reasons that are hard for people to understand — are completely and utterly unacceptable."-Steven LevittA good post, with a problem: this initial quotation doesn't make sense. Levitt makes an illogical shift midstream in the argument. (Assuming he is using the word "repugnant" in the only sense I understand it, to mean morally repugnant or "wrong", and assuming that "wrong" is something defined by a broad cross-section of the population.}So he starts by appearing to want to say that economists are unconventional and reason beyond normal  assumptions ("in terms of whether it's efficient.") Then he appears to realize that the logic of his argument demands that he acknowledge that a preference for efficiency makes economists consider anything, however immoral (kill everyone over 65 and lower tax rates by spending less on healthcare!). So he decides, "Whoops! Can't say that! Too many easy objections." So he decides that there is some third body of gnostic economic wisdom — neither efficiency not morality — that is revealed to economists but not ordinary folk. And this reveals to us that the contemplated action is ….. "completely and utterly unacceptable."In this case, Professor Levitt is deeply confused.

  56. Anonymous says:

    Suppose I am watching a chess game by a master. I don't understand a particular move the master makes, so I construct a model of optimal chess-playing. You could argue that masters make sub-optimal moves sometimes and so my model is wrong. But it probably gives a lot of insight into the game, and is a much better starting point for understanding chess master behavior than a model based on sub-optimal play.

    In the same way, suppose I walk into a movie theater and think that the price of popcorn is too high. Unless I'm an expert on the movie theater industry, chances are I'm wrong and the price is pretty close to what comes out of a model of rational agents picking popcorn prices. If you want to make a general theory of economic behavior, rationality is a useful assumption to have around.

    Suppose you're an economist and want to write a best-seller. One option is to take a lot of seemingly irrational behaviors and show how they come out of a model of rational decision-making. Freaky! In the same way, if you want to write a chess book you can take a lot of seemingly bad or inexplicable moves, and show how they come out of a model of optimal chess-playing.

  57. Anonymous says:

    One of the most perpetuated misconceptions about microeconomics – even among those who should know better – is that its results depend on people being rational ("economists think ordinary people are rational…"). The foundations of microeconomics depend on people having meaningful preferences and beliefs, and is agnostic as to whether those are rational or not. The "radical" next step (which can be weakened) is to require that if a person prefers A to B, and also prefers B to C, then that person prefers A to C. (The relation may or may not be strict. The theory is equivalent in either case if we allow the possibility of indifference.)
    The assumption that people act reasonably consistently with respect to their preferences and beliefs is a far cry from, and much weaker than, the claim that people are rational. It is hardly possible to analyze behavior that is attributed to conscious intent without the former, but nobody who functions in the real world could think the latter.
    Thus, for example, we can analyze the implications of an individual prefering not to increase the debt ceiling without determining if it is rational to prefer that. As a separate issue we can debate whether or not such a preference is rational, i.e. consistent with an individual's other preferences and beliefs, but a priori it is entirely possible that the preference could be part of such a consistent system, and that that consistent system might not yield the same result as your consistent system. A consistent system can completely ignore empirical evidence. A corollary of Goedel's incompleteness results is that, for any consistent countable set of axioms (beliefs and preferences), there are independent axioms for which either the axioms or their negations (but not both) combined with the original set yield a consistent system. There are questions that are formally decidable but too hard to solve – i.e. NP-complete. For these and other reasons it is generally beside the point to try to decide whether or not a given behavior is rational beyond the question of whether or not it is consistent. If a person's behavior is clearly inconsistent we generally say "You made a mistake" or "You are lying." If a person's behavior is inconsistent in a complex way we attribute it to bounded rationality.
    "Efficiency" to a microeconomist usually means "Pareto efficiency," but all good textbooks and instructors will accompany the definition of Pareto efficiency with a disclaimer that it is only one possible definition of efficiency. Many will also say that it seems to be a necessary condition for a reasonable concept of efficiency but that additional conditions might also be appropriate. (Pareto efficiency means that no one can be made better-off without hurting someone else. Thus, if for example an allocation is not Pareto efficient, it is possible to make at least one person better-off without hurting anyone else.)
    I think an important part of what you said here depends on its restriction to "pop" economists. I agree that a lot of what you read in mass media is so confused that it is next to useless. There are people who don't know what they are talking about trying to sound like they do (In the land of the blind the man with one eye is king). There are people who do know what they are talking about but have an agenda. There are people who know what they are talking about but dumb things down so much they sound like they don't. But I'm a microeconomist, I don't think people are entirely rational, and I also get useful (and empirically verifiable) results by analyzing models that assume people have consistent preferences.
    PhD Advanced Micro Theory, UC Berkeley

  58. ezra abrams says:

    an example of liberal econ arrogance..if you look at all the prominent liberal econ bloggers, Brad deLong, and Paul Krugman and R Reich, and so forth, what is striking is that each is in his own little tower, trumpeting his own little views; they don't organize and amplify their voices to make an impact.
    The only conclusion is that it is more important to them to feed their own egos, by maximizing the loudness of their own blogs, then to actually ahve an effect.
    I guess you could exempt PK, cause the Times amplifes his voice.

  59. […] Gelman makes the following observation: Pop economists (or, at least, pop micro-economists) are often making one of two […]

  60. […] more I think about this Andrew Gelman post, the more ridiculous it seems.  Gelman argues that economists, especially popular economists, use […]

  61. Frank says:

    What is "rationality" is just an adaptation to win arguments – rather than arive at truth? http://www.nytimes.com/2011/06/15/arts/people-arg

    I think the reason economists switch between the two arguments is that economic theories are better at predicting macro-level events, while they often fall apart at the micro-level. This is why behavioral econ has taken off within the last few years, in an effort to fill the gap.

  62. […] proverbial 2 cents about math competitions — including a grueling quote. But there was also Statistical Modeling, Causal Inference and Social Science picking up a piece about the Barber’… while Rhapsody in Numbers wondered about a young woman that wants to be the first on […]

  63. […] of modern finance and economics. one particular bugbear is the notion that people are rational. Here is an interesting blog saying that everyone can’t be […]

  64. […] Gelman recently wondered how so many economists can hold two seemingly contradictory beliefs: 1. People are rational and […]

  65. […] in economics and failures of imagination. Andrew Gelman hints at the edges of this problem in a blog post over the summer. Gelman is talking about “pop economists” like Steven Levitt but the […]