Scott Sumner  

Congratulations to Richard Thaler

Hume on Pessimistic Bias... Henderson on Thaler's Nobel...

As a University of Chicago alum, it's always nice to see the UC pick up another Nobel Prize in economics.

Economists are often accused of engaging in empty theorizing, but Thaler has developed useful ideas, such as methods by which the public can be "nudged" into boosting their saving rate. Some of these have been successfully implemented.

Alex Tabarrok and Tyler Cowen have nice summaries of some of his contributions, and this comment by Tyler brought a smile to my face:

Perhaps unknown to many, Thaler's most heavily cited piece is on whether the stock market overreacts. He says yes this is possible for psychological reasons, and this article also uncovered some of the key evidence in favor of the now-vanquished "January effect" in stock returns, namely that for a while the market did very very well in the month of January. (Once discovered the effect went away.)
Went away? Well that's one way of putting it. Imagine a study in psychology, perhaps one of those where they prime people with shocking pictures, and see how it affects their behavior. Then imagine a number of follow up studies that fail to replicate the original study. (This inability to replicate happens all the time in the social sciences.) Would we say the effect "went away"? So why do we show such respect to the study of anomalies in finance?

Behavioral economics is not my area, so perhaps someone can help me with the following remark by Tyler:

Very lately Thaler on Twitter has been making some critical remarks about price gouging, suggesting we also must take into account what customers perceive as fair.
I had thought that the point of behavioral finance was to try to identify areas where the publics' intuition is in some sense "wrong" and design public policies to nudge them in the right direction. In that case, shouldn't we be encouraging price gouging, not discouraging it? Especially with public policy. Perhaps President Trump could "name and shame" companies that run out of essential supplies because of an unwillingness to price gouge. (Yes, not likely, I'm just trying to show the logical implications of behavioral economics.)

Note that this is a separate point from whether firms themselves ought to take into account public perceptions---I agree that there are cases where firms may decide it's wise to avoid price gouging for reasons of public relations. But that's different from endorsing laws that restrict price gouging. Don Boudreaux and David Henderson recently made related criticisms.

Screen Shot 2017-10-09 at 12.17.09 PM.png

COMMENTS (13 to date)
John Thacker writes:
So why do we show such respect to the study of anomalies in finance?

The same respect is shown for research on betting line anomalies and sports strategies. I would say that it is because there are market makers who can read the papers and adjust the market returns or betting strategies. Indeed, part of the reason for the EMH is that people are searching for inefficiencies and arbitrage to eliminate them.

It seems less plausible that ordinary people are reading about priming research and adjusting their behavior to eliminate it. It's possible that people do respond to research in that way, but it would require nearly everyone to adjust their own behavior to eliminate the effect. By contrast, in finance or betting a small number of sharps can eliminate the excess returns even without most people being familiar with the research.

Emerich writes:

As usual, you hit it on the nose. Nudging should indeed be deployed to encourage gouging. Why do I doubt it ever will be, or even that Thaler and his co-nudgeniks will acknowledge that it should be?

Vera te Velde writes:

2nd half of your post is a most excellent point! And one that needs to be made much more often and louder. On the other hand (not referring specifically price gouging) I think behavioral sciences are shifting focus from "look at all these ways people are irrational" to "why things that seem irrational on the face of it are actually not" so a similar goal of correcting behavioral actors may miss the mark in other areas.

As for why we say "the effect disappeared" instead of "the effect didn't replicate"; as John said above, it's of course because we are more Bayesian than our p-values. I have a strong prior that identified free lunches will be taken advantage of.

On many situations, blackboard economic reasoning flows from behavioral anomalies to a bona fides normative argument for government intervention in order to correct distorted market outcomes. In practice, however, governments will often act to make the distortions even worse as political actors are subject to the same problems.

People misjudge gouging, therefore there is a bona fides economic reason for the state to intervene and promote price gouging. We observe the opposite.

People are loath to take a pay cut, leading to the zero bound on wages, which can have awful macro consequences. Therefore, there is a bona fides economic reason for the state to intervene and promote flexibility of wages. We observe the opposite.

People overvalue owning property vs renting. Therefore, the state should make it less appealing to own and more appealing to rent. We observe the opposite.

I could go on.

Scott Sumner writes:

Everyone, Good points.

Just to be clear, I am suggesting there never was a January effect---it was an artifact of data mining.

Nathan Smith writes:

"Just to be clear, I am suggesting there never was a January effect---it was an artifact of data mining."

Fair enough, but you asked, "So why do we show such respect to the study of anomalies in finance?" There's a very good reason why we shouldn't *expect* an anomaly in asset prices to persist, once it has been noticed, and therefore, why the failure of an anomaly to persist after it has been noticed is not evidence that it wasn't really there in the first place.

Mark writes:


"Just to be clear, I am suggesting there never was a January effect---it was an artifact of data mining."

I actually initially misinterpreted what you were saying in what I think is an interesting way relating to behavioral finance and the rationality of markets.

That is: how does the discovery of a pattern of irrationality in the behavior of markets impact the behavior of markets? What I thought you were suggesting that, once made aware of the January effect, investors self-corrected.

If this is the case, one could argue (this seems relevant to EMH) that it is impossible for markets to behave in a persistently irrational manner that is observable, for once it is observed that investors are behaving irrationally, they will take notice and modify their behavior.

Scott Sumner writes:

Nathan and Mark, Yes, that's possible. But why assume there was ever any effect to be corrected?

Why not simply assume the first study was wrong?

Matthew Waters writes:

First of all, on the EMH generally, there is also the general outperformance of value stocks, 1987 crash, etc. If the EMH is to be falsifiable in any way, there has to be a decent amount of information falsifying or at least significantly limiting it.

On the January effect itself, it was found in 15 of 16 countries. Interestingly, the January effect was even a thing in Japan, which does not have capital gains and loss offsets.

Based on international evidence, the stock returns tend to coincide with semiannual bonuses. There is also a smaller outsize return in the middle of the year.

It's not clear if the effect was profitable for a large investor to act against, due to transaction costs. As Thaler says, that doesn't mean the effect should happen. At least some of the buyers in January had money to buy in December, for example. Market makers could have also held on to a higher inventory of stocks in December. With all players being rational, the January effect shouldn't have happened but it did.

The January effect also "went away" as bid-ask prices went down and there was more capital and leverage available to hedge funds such as Renaissance Technologies. The January effect also subsided as it became more well-known.

In any case, I don't see how the effect could be casually dismissed as spurious data mining. It happened in 15 out of 16 countries and the effect had common-sense reasons (tax loss realization and spending bonus on investments).

Matthew Waters writes:

Econlog doesn't like long URL's. Here's the link to Thaler's paper.


[Actually, it’s shortened URLs we don’t like. We prefer when people can see where they are going to go when they click the link. I’ve changed your shortened URL to the full URL and made it into a link for you. —Econlib Ed.]

Mark writes:


If markets correct for discovered patterns of irrational behavior, then no such discovery of irrational market behavior could be consistently replicated, as once it's first demonstrated, it will disappear. This may make it more difficult to conclude what the simplest explanation is for such a discovery.

Kevin writes:

I thought we had learned that pretty much all priming was not real and that was not replicated because it was not real and just appealed to our ridiculous biases.

Nudging on the other hand...

Andy writes:

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

Comments for this entry have been closed
Return to top