Scott Sumner  

Behavioral economists aren't as smart as they think

China is rapidly shifting to a... High Profits as an Incentive t...

I recently did a post criticizing economists who second guess the behavior of others. This new post gives me another opportunity:

Thaler asserts that good economics needs good psychology, instead of its tradition of inventing bad psychology. He makes short work of various defenses of the rational maximizer model, including "if you raise the stakes, people will get it right," which studies contradict. Some examples of irrationality:
Markets cater to consumer biases rather than correct them. Extended warranties generate $27 billion/year in revenues, even though economists and Consumer Reports agree that one should never buy them.

Financial intermediation is 9% of GDP, despite the ease (and almost certainly higher returns for most investors from) index funds.

A closed-end mutual fund investing in the Caribbean, not including Cuba, had the ticker name CUBA. Long trading significantly below net asset value (itself an irrationality), its name made the shares spike when President Obama announced diplomatic relations with Cuba, and only slowly retreated over the next few months to net asset value.

Housing prices have historically averaged 20 times rents. In the 2000's they ballooned far above that ratio. A bubble, or what?

Economists' theory of labor markets says that workers earn their marginal product. But when workers shift from high-wage firms to low-wage ones, or vice versa, their compensation falls or rises according to the success of their firm. Janitors make far more at Goldman-Sachs than at Wal-Mart. Is that because they are more productive?

That's your evidence for behavioral economics? Seriously? I can refute three of these right off the top of my head, and I've never even done any research in this area. Imagine what a Gary Becker could have done with this list.

1. Suppose people have some sort of loss aversion. Their anger about being ripped off on a product is far higher than their happiness when things go well. Then buying the product warranty will buy some peace of mind. I don't buy them, but that's because I'm unusual. I don't have the normal amount of loss aversion. But for normal people it's perfectly rational, i.e. consistent with utility maximization.

2. In the 21st century, long-term real (and nominal) interest rates have fallen to a new normal that is lower that in the 20th century, due to lots of factors such as demographics. So asset prices should be higher, relative to rents. Again, perfectly consistent with rationality.

3. Of course janitors for Goldman Sachs are more productive than janitors for Walmart, how could anyone think otherwise? Suppose you are having a big meeting on a billion dollar deal, and your client walks into the bathroom. The toilet is plugged and overflowing. Gross! How does that affect her impression of your company? Indeed even the Goldman Sachs employees themselves care far more about cleanliness (in dollar terms) than Walmart employees. Not because they have better taste, but because they are richer.

These critiques of the rational choice model aren't just wrong; they are easily refuted with 5 minutes of thought. If that's the sort of "critique" that has come out of 30 years of behavioral economics, then I'm glad I never went into the field. And again, these refutations are just off the top of my head.

PS. I admit the CUBA one looks fishy, but even there it's possible that they thought the liberalization would provide the fund with new and lucrative markets to enter (Cuba is huge), and that given the fund had expertise in the Caribbean, they would have an advantage over other existing funds. That doesn't explain the reversion in price, but prices change for lots of reasons. And I admit my explanation is a stretch, I'm just saying it's always possible we are missing some angle. Or maybe it was indeed a random error, as they suggested, but with no important implications for anything else.

And what percent of GDP do the behaviorists believe should be in finance? So that one is hard to evaluate. The claim about index funds may be true, but it has no policy implications. Because investment information is partly a public good, if you adjust policy to push people into index funds, there'd be too little market research.

PPS. Sorry if the tone here was too negative, but if your entire career is built around the idea that most people are stupid, and you're one of the select few to see through their stupidity, then you better be pretty sure that you have your facts right, or else have a thick skin.

HT: Tyler Cowen

Comments and Sharing

COMMENTS (30 to date)
Jon Murphy writes:

Excellent post! We certainly see the danger of nudging

A study of the cognitive biases of people who study cognitive biases may be in order.

Ieiuus writes:

[Comment removed pending confirmation of email address. Email the to request restoring this comment. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

Kevin Erdmann writes:

Analysis that seems to support EMH could just be curve fitting. On the other hand, behavioral finance is like curve mis-fitting. Confirmation bias for EMH comes from trying too hard. Confirmation bias for behavioral finance comes from not trying at all.

Unlearning writes:

I am forced to ask what exactly could refute your idea of rationality. Mathematically, you can put virtually anything into a utility function, but that doesn't make for a meaningful definition of rationality; it's just a tautology.

Behavioural economics does more to show how obscene economists' idea of 'rationality' is than to show that people are 'stupid'.

And I admit my explanation is a stretch, I'm just saying it's always possible we are missing some angle.

Yes, it's always possible we're missing some angle for any observation, dataset, experiment or case study. If you consider this justification for dismissing said observation, then you can dismiss virtually any observation you don't like.

Or maybe it was indeed a random error, as they suggested, but with no important implications for anything else.

Nobody suggested it was a random error. The suggestion is quite clearly that it is a systemic error, based on investors mistakenly believing the fund has something to do with Cuba. This is quite similar to another case, where investors invest in Deutsche Bank because they think it's the central bank of Germany:

If you can find a decent explanation for these things that saves the EMH, then fine. But this is just handwaving and speculation.

Ray Tseng writes:

You should give Thaler a little more credit. He has written voluminously about some of your responses.

Specifically, about interindustry wage differentials, Thaler notes many empirical anomalies that cast doubt on a compensating for unseen abilities explanation. There is a chapter in The Winner's Curse on these findings.

Here's some commentary from page 49...

"Now consider the case of two large firms with plants located in the same community. Both firms have clerical staffs that perform virtually identical services within the firms. Firm H is in a high-wage industry and pays its clerical staff WH, while firm L is in a low-wage industry and pays its clerical staff only WL

Ray Tseng writes:

Suppose that Firm H decides to save money by cutting the wage of its clerical workers to WL. Is this action profitable? That depends on the reaction of the clerical workers. If the workers think of their old wage as a fair one they may resist the wage cut in various ways that can be summarized as saying they become less cooperative. The reduction in worker cooperation could easily offset any gains from reducing the wage bill... To sum up, I find the pattern of industry wages difficult to understand unless we assume that firms pay attention to perceived equity in setting wages, an assumption that only an economist would find controversial"

John Hall writes:

I would expect that janitors at Goldman Sachs work for an outside contractor.

Daniel Kuehn writes:

Unlearning -
I don't think it's really an issue here of what will refute Scott, it's that Thaler goes way farther than he has warrant to.

I think behavioral economics is important but, as Scott points out, not as some kind of replacement for a bankrupt neoclassical theory. I don't think, for example, the marginal productivity theory of wages is unassailable. Rules of thumb almost certainly matter, as do concepts of fairness, perhaps some backward-looking behavior in expectation formation, and any number of other things that go into wage formation. To the extent that behavioral economics can provide some of this realism I think that's great. But Scott's point, I think (which I agree with) is that this is just simply a bad example that doesn't prove anything that Thaler purports it proves.

Behavioral economics is not the new paradigm. At best its a sign of scientific crisis, but it might not even be that.

Brett writes:

I don't think your response to #1 really refutes it. That's the whole point - that rationality is bounded by built-in psychological and behavioral stuff. In other words, people aren't full rational-maximizers because they place a disproportionate weight on losing stuff versus gaining new things.

Scott Sumner writes:

Thanks Jon.

Joseph and Kevin, I agree.

Unlearning, I don't see how this addresses my point. We have a list of items that supposedly refutes "rationality" and I can immediately refute three of the five without even doing any research on my own. I grant you that the CUBA example is quite possibly an example of investor ignorance. But even he doesn't view it as persistent, as the price reverted.

Regarding your "tautology" claim, read the earlier post I linked to.

Ray, First of all I was addressing the specific claims made in this post. Pointing to other possible examples doesn't undercut the fact that this particular blogger thought there was no explanation for janitors making more money when they shift to another company with higher wages.

Second, my argument applies to the clerical staff at Goldman and Sachs vs. Walmart, just as well as the janitors. Yes, I'm aware that the behaviorists have lots of other examples, and I've seen some of them. A few are probably correct--people aren't perfect (Trump proves that). But this sort of lazy post only undercuts their argument.

The world is a much more complicated place than many economists seem to imagine it to be. The truth is that we don't know why lots of people do lots of different things. But that doesn't make them irrational, it makes us ignorant.

John. Then use a different profession as an example.

Daniel, Good comment.

Brett, Why is that not "rational?" Who's to say what people should care about? Read the post I link to about Christmas presents---are those also not rational? Since when does neoclassical econ assume people don't have emotions like anger? If it doesn't fit the mathematical models we've come up with, then create new models that better describe reality. Maybe people care more about appliance breaking risk than losing 50 bucks at a casino risk. Anything wrong with those preferences?

Don Boudreaux writes:

The late, great Stanley Lebergott would have agreed with Scott. In his 1993 volume, Pursuing Happiness, Lebergott - then discussing the assertions of Veblen, J.K. Galbraith, and E.J. Mishan - described these sorts of intellectual criticisms of observed consumer choices as "Petulant differences in taste masquerading as high analysis" [p. 26].

Steve J writes:

"Anything wrong with those preferences?"

Yes I think we should say many utility functions are wrong. Fat people should eat less, poor people should not play the lottery, most people should not smoke, etc. If I could change my own utility function I would. But I am too lazy ;-)

Mark writes:

I get the impression that most of behavioral economists' observations of irrational consumers could be easily recharacterized in terms of unaccounted for utility. The joy people derive from gambling, for example, could be readily described by their von Neumann-Morgestern utility curve. Now, one might argue that gamblers have the 'wrong' von Neumann Morgensern curve, but then what is the 'right' (rational) one? Absolute risk neutrality? Are people then irrational to be (most of the time) somewhat risk averse? If not, then until behavioral economists can say what the optimal risk aversion level is, they are largely just pretending to scientifically prove that chocolate ice cream is objectively better than vanilla.

Never mind that that the supposed policy relevance of behavioral economics is usually rendered moot by the fact that, little as many people may know about what's good for them, the average policymaker knows even less about it. Even if we accept that a rational consumer is a rare thing, a rational public official is an even rarer thing.

Anonymous writes:

@Steve J

If you take that argument - roughly, that people's utility functions should include as few conflicting wants as possible - to its logical conclusion, you end up with the ideal utility function being to want nothing at all. After all, the problem with fat people liking to eat and smokers liking to smoke is that satisfying their demand for short term pleasure reduces their ability to satisfy their demand to stay alive. The problem with poor people playing the lottery is that satisfying their demand to do so makes it slightly harder to satisfy their other demands. But the same applies to any other demand they have. It seems equally valid to say that poor people shouldn't want to do anything but to play the lottery, or fat people and smokers shouldn't want to stay alive.

Ultimately I think this is a kind of pointless argument, because the whole point of economics involves maximising satisfaction in light of conflicting demands; assuming away the problem is cheating.

Jack PQ writes:

A theory should be more than a collection of loosely related anomalies. Neoclassical theory is impressive not because it explains everything--it does not--but because it explains a great deal using very little. It is parsimonious, and fairly universal.

Behavioral economics--my professors in grad school insisted on calling it "psychology & economics"--has identified many anomalies, some explained fairly easily, others not. But their answer, which is to have an ad hoc theory that takes a different form for each setting and situation, is not a good alternative.

I liked Dr. Arnold Kling's analogy from several years ago, he said behavioral economics was like the rookie of the year who impresses many and shows great promise. Rookie of the year does not mean he (or she) is automatically headed for the Hall of Fame. There is a long road ahead. But I'll be watching out for him (or her).

Steve J writes:


When people act in ways that go against their stated interests that appears irrational to me. Possibly it indicates that we don't know what our own interests are. For me the extent that people act rationally should influence the extent we trust in free markets. Not pointless.

Roger McKinney writes:

Behavioral econ seem to fixate on the very short run while the EMH looks only at the very long run. We need a medium term view like Austrian econ.

Nathan W writes:

Behavioural economics is ridiculously new as a field and poorly funded to be highly critical.

Information quality is generally very poor for practical purposes, and I have very little certainty of whether the most important economic decisions I make are good decisions.

Kurtis Fechtmeyer writes:

Behavioral economists find economic irrationality when people are thinking for themselves, but somehow assume cognitive biases disappear when bureaucrats sit in wood-paneled offices as arbiters of other people's interests.

Charlie writes:

The way Thaler tells this CUBA story is really disingenuous. The Herzfeld Caribbean Basin CUBA Fund is expressly trying to make money off of Cuba and political developments that would change US diplomatic relations with Cuba.

Thaler tells the story like it's just some random mistake like CUBA is some acronym, and I was pleased to have Unlearning confirm this interpretation, "Nobody suggested it was a random error. The suggestion is quite clearly that it is a systemic error, based on investors mistakenly believing the fund has something to do with Cuba."

The facts:

1. The fund is trying to make money off of Cuba. Here is the 2013 report to shareholders.

Statement of purpose:
The Herzfeld Caribbean Basin Fund seeks long-term capital appreciation through investment in companies which the Advisor believes are poised to benefit from economic, political, structural and technological developments in countries of the Caribbean Basin. U.S. relations with Cuba are particularly influential to the region and, therefore, we keep a close eye on any developments.

The first subsection is "Cuban Relations".

In the portfolio review, it describes current stock holdings and explains the fund is invested in them, because they believe they are positioned to benefit from a resumption of US and Cuban trade relations.

"We believe all are in a position to benefit from any relaxation of the U.S. embargo against Cuba and normalization of relations that would result in a boost to tourism. The Fund currently holds core positions in all three."

2. The managers suceeded in getting exposure to Cuba. That is, the assets they hold did great when the US announced they would begin relaxing the Cuban Embargo. The companies the fund held jumped 15-20% on announcement. This is the jump in the value of the assets that CUBA holds.

3. Since CUBA is a closed end fund, the value of its assets and the price it sells those assets for are disconnected. The price of CUBA's fund rose about 30%. The assets started selling at a premium. This is justifiably weird.

4. Why not just buy all the stocks CUBA sells individually? Well, why pay anyone to invest for you? Presumably, you think they will do a better job than you can. The possible answer here is that you think this fund, which has been following Cuba, the Caribbean, and thinking about making money off U.S. Cuban relations, could find a lot of good investments in the near future. With articles like, "Herzfeld Caribbean is Fully Prepared to Invest in Cuba" one could logically think that if investment opportunities were to happen quickly, the fund could pounce. If you wait until their quarterly disclosures and try to copy cat their strategy, it may be way too late, prices are too high by then.

5. Ex-post, we know they were probably wrong. Some bought the fund at too high a price (not that many, this is a tiny $30 million fund), so we can laugh at them and say how obvious it was. But then again, I first heard of Bitcoin at $50 and didn't buy any, and now it's at $373, so maybe I'm the idiot.

Still, I heard Thaler discuss this example on a Bloomberg masters in business interview, and the way he tells it is not even misleading, it's just wrong.

Alex Demitraszek writes:

This is what's wrong with economics today. Lumping all behavioral economists in with Thaler is as ridiculous as considering Sumner just a mainstream macroeconomist. Behavioral economics has become one of the biggest and most funded fields of economics, especially following the large amount of Nobel laureates that have come from the field. There are behavioral economists who tend to be all for paternalism, some that are middle of the ground and theoretical about it (of which Thaler is one) and then some Behavioral economists that don't even deal with paternalism. The funny thing is that Most people generally consider George Mason to be the headquarters of Austrian economics, yet it ranks highest in Behavioral, decision theory and Experimental Economics because of the work of Vernon Smith and Kevin McCabe. If behavioral economics was as baseless as Sumner describes, then surely it wouldn't be able to pass the market test of academic publishing. Yet all of the deep insights that have been published In behavioral economics, mathematical psychology,the behavioral aspects of public choice, and prospect and utility theory are mainstays in the top ranked journals and have won Nobels. There are even Austrian behavioral economists (albeit much like standard austrians their work doesn't pass the market test of publication as often). Lots of behavioral economic insights have been used and applied with great success, especially in the developing world (Esther Duflo and Bill Easterly's work is testament to this.)

Scott Sumner writes:

Alex, I did not say behavioral economics was baseless. I was responding to someone who said it disproved the rationality assumption of neoclassical economics, and then proceeded to provide examples that did nothing of the kind. I said if this is the evidence for behavioral, it's pretty weak.

I happen to think the rationality assumption is quite useful. If there are good arguments that it is not, then I'd like to see them.

Federico writes:

On CUBA -- in my experience you can have a fund trade at a price different than the NAV. The NAV reflects the price on the books. If investors think the true price is actually different than what's on the book then the NAV can be justifiably different than the price of the fund. Now, of course, there is a way to take advantage of this but it's quite complex. Basically you have to buy a very large share of the fund and then sue the board of the fund so that you can takeover and sell the proceeds of the fund at the actual price (you do this if the NAV is above the price of the fund, if not you do the opposite which is more complicated). Btw, the other explanations cited about "skill" and being ready to "pounce" are also valid.

Jose Romeu Robazzi writes:

I did not read all the comments, but I don't think behavioral economics disproves rationality in economic decision making. It just tries to understand how people's biases impact their utility functions, which in turn become subjective. But every single individual making these decisions believe themselves to be rational ... and they are!

Arthur B. writes:

CUBA is probably not a good example, but TWTRQ is.

BG writes:

Prospect theory. Cumulative prospect theory. Why no mention of the actual frameworks for behavioral econ?

Note: rational choice != neoclassical economics. Neoclassical economics imposes additional assumptions on rational choice - transitivity of preferences, etc. I tend to view behavioral as challenges to the additional assumptions, instead offering its own assumptions (which won the Nobel).

Also, why not write an article for a top peer-reviewed journal rationalizing the problems? Surely that is how this "scholarship" thing is supposed to work.

J Mann writes:

Brett, if you call preferences irrational, then you haven't really invented anything, since the classical model includes preferences.

For example, suppose you discover that peanut butter and vegemite have about the same nutritional value, but that most Americans are willing to pay more for peanut butter than vegemite, and vice versa for people in Britain or Australia or wherever it is that vegemite comes from.

Would you say that it's "irrational" that most Americans would pay more for peanut butter than the nutritionally equivalent vegemite, and even if so, is that a criticism of textbook economics?

Arthur: I think that TWTRQ is a good example of people making mistakes (or maybe of a thinly traded stock being bid up by day traders who hoped to sell it to mistaken investors), but that's not a great story for behavioral economics either. Sometimes people drive past their exit on the highway, even when it would be more rational to take that exit. I think you can do a better job of modeling mistakes through some Bayesian analysis, search costs, and rational ignorance theory than with behavioral economics.

Yaj Reizarb writes:

Most, if not all of the behavorial economics enterprise seems to be motivated by a desire to demonstrate that the average persons values, preferences, and motivations fail to satisfy a set of optimality criteria that have been pre-defined to conflate the arbitrary preferences of left-progressive technocrats with "rationality." Put more simply, the point is to show that the people are too stupid to look after themselves and must become wards of the the likes of Thaler et al in order to live in ways that progressive technocrats "know" is best for them.

There's always been a similar ideological motivation behind the academic research on "happiness," in which the decreasing association between wealth and happiness over and above a given threshold are used to justify ever higher rates of taxation. "All that money they're working so hard to make isn't making them any happier, so taking most of it and giving it to those who make less will increase happiness." Here, as in the behavior econ literature, the key to achieving utopia lies in ceding power and authority to progressive technocrats.

Charlie writes:

"But when workers shift from high-wage firms to low-wage ones, or vice versa, their compensation falls or rises according to the success of their firm."

I thought of another example of this. Think of baseball players. A player on the Scraton Rail Riders often has a fairly modest income, but when that player moves on to the New York Yankees, he very quickly gets paid considerably more, even though it's the exact same player. How could that possibly be justified except that the Yankees are a higher status team than the Rail Riders? Clearly, something behavioral is going on.

Comments for this entry have been closed
Return to top