May 2018

Sun Mon Tue Wed Thu Fri Sat
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31    
Blog powered by Typepad

« Non-Tenure Track Professors are VERY Important and Will Be More Important in the Future, But Prepare Yourself So You are Never One of Them | Main | New Working Papers on the Current Recession »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Some silly commentaries.

In the article:

"In mainstream modern economic theory, a great deal of attention is paid to how players learn their way to “equilibrium” and what kind of equilibrium might result."

O really? I think that there is a very large lag between the cutting edge research in economic theory and textbook writing since in the most important microeconomic textbook of your time we only have about 1.5% of the text about disequilibrium adjustment and the other 98.5% about equilibrium analysis.

"It requires that beliefs be correct about the things that players actually see – the consequences of the offer they actually make – but not that they have correct beliefs about things they do not see – the consequences of offers that they do not make."

So the game theorists finally have discovered the concept of errors of overpessimism. How much years before they envision the concept of alertness?

I found the paper disappointing in that Levine did not discuss at all invariance.

Tversky's great challenge was a thoughtful attack on the principal of invariance.

Invariance is the principle which allows the mathematization of decisions.

Levine simply ignores this challenge and potters on about prospect theory.

Nice choice. An interesting paper, which references interesting papers, which are now added to the list.

I dislike the word "certainty" as it is used in statistics.

I have met people who are each absolutely certain that the other is wrong, and other uncertain people who are usually right. How certain I am about something is more influenced by what I ate for breakfast this morning than some probability calculation. Moreover, I don't want to be certain about anything, because while certainty does not guarantee that I am right, it is a significant barrier to learning that I am wrong. And in that case, why should I pursue more certainty? Uncertainty equals neither indecision or falsity.

It all seems to have rather ugly roots in justificationist philosophy, and in an attempt to control what ideas it is just for people to believe, and how strongly they are to believe them.

Although I haven't read this paper, I have read a ton of pro-behavioral stuff. The root of the problem when this literature is applied to paternalistic legislation is that, if you take the findings seriously, there are so many factors moving in different directions that it is hard to make definite statements. Most of the behavioral literature takes one bias at a time and comes up with a critique of a standard result accordingly. Throw in a few biases and you don't know what is going on.

Another problem is that some behavioralists think economics is about individual behavior. In reality it is about market behavior. One has to deal with the filtering mechanisms of markets. Often "irrationalities" are filtered out by the market.

Glen Whitman and I have two papers coming out at the end of the year on behavioral paternalism: One in the Arizona Law Review and the other in the Brigham Young Law Review. I'll put the final versions up on my bepress site as soon as the editors finish "perfecting" our papers.


John List's work is great on this among contemporary economists.


Yes, List is excellent.


This is a very poor paper in many ways. I have done a pretty full list over on Marginal Revolution, but a few points. Yes, standard theory works fine for some things, such as Vernon Smith's double auction stuff. And the law of large numbers holds for voting results. But we have just seen the law of large numbers fail miserably in the recent financial crisis. What is herding, and why did the pricing for CDSs by the linear bivariate Gaussian copula break down? (hint: law of large numbers failed).

Levine claims he and Fudenburg invented self-fulflilling equilibria. Baloney. Azariadis, Guesnerie, Mordechai Kurz, and a rather long list of others ahead of them on that one, but this paper looks like one of those by Krugman ignoring everybody before him.

Plott and Zeiler provide a good critique of some framing effects in some lab experiments using undergrads and coffee cups. I know forthcoming papers that point out that they have themselves engaged in framing effects. The main evidence for WTA-WTP discrepancies comes from massive amounts of field studies. P and Z do not begin to refute those.

He poses prospect theory as the only alternative to expected utility. There are several others. Simply incompetent presentation of the literature. This guy is paid how much at Wash U?

He poses quantal response equilibria as a possible explanation for some results. I have news for you folks. One can explain almost anything with a properly calibrated QRE. Does that prove rationality or disprove irrationality?
Sorry, but I do not think so.

Levine simply ignores some of the more important findings that are now widely accepted, such as hyperbolic discounting (although Rubinstein has made a critique, not very convincing). There is now lots of neuro stuff supporting the hd effect, with essentially the short run discount rates being driven by lower brain areas that are associated with "fight or flight" and immediate rewards or punishments, with having those effects not necessarily being "irrational" from some long-run evolutionary perspective, but definitely irrational from the standard Chicago perspective which requires time consistency.

Levine's effort to make impulsiveness "rational" by noting that Elliott Spitzer and Larry Craig apparently planned for their shenanigans simply shows that their shenanigans were not impulsive. Pathetic.


Clearly there is a vein of behavioral paternalism that shades into authoritarianism. But there are plenty of behavioralists who are libertarians and mostly free market. Vernon Smith is the supreme example, with his experiments showing the functioning of double auction markets usually being viewed as very pro-free market. At the same time, he is one of the main leaders of the experimental research showing the profoundly ubiqutous and deeply rooted nature of speculative bubbles, which, I remind, are exactly the sort of collective herding behavior that overthrows the law of large numbers that you cite.

To Pete and Mario,

List's work is not a defense of rationality. It is a defense of field experiments over lab experiments. I remind that at least in the case of the WTA-WTP stuff, it is the field experiments that support the violation of standard theory in contrast with the lab experiments of Plott and Zeiler.

Altogether, this is a profoundly poor paper in many ways, just pathetic.


Not sure that we are not talking past one another. Neither Mario nor I are looking at List's work to address the "rationality" hypothesis, but instead the "filter mechanism" of the market economy. Similar to Vernon Smith's Nobel, the question of individual rationality is more "contextual" than either mainstream microeconomics which often attributes to individuals hyper-rationality and quasi-omniscience or modern behavioral economics which sees any deviation from the ideal model of abstract rationality as an anomalous behavior.

It is an important matter of the history of ideas to point out that Menger, Mises and Hayek all warned against the hyper-rational model. As Jaffe pointed out in his important paper on Menger, Jevons and Walras, to Menger man was not a lightening calculator of pleasure and pain, but caught between alluring hopes and haunting fears. Mises argued that homoeconomicus actually precluded the development of praxeology, and Hayek contrasted his notion of human rationality (evolutionary rationality) with the rationalistic notion of homoeconomicus.

So it is not that List is defending the "rationality hypothesis", but instead how the filter of the market does the job. In other words, it is not behavioral assumptions that is doing the hard work, it is the institutional conditions.

One last reference, Hayek in Economics and Knowledge and later in his essay in response to Langer, Lerner, etc. "The competitive solution", the key idea as far as I am concerned is that the marginal conditions are by-products of the market, not an assumption going into the analysis.

Peter B.,

No disagreement with this last comment. However, it looks like Pete L. might be in danger of smoking too many cigarettes in the office of Bob Lucas. If he thinks this ridiculous piece by Levine is worthy, I would not be surprised if he starts telling us that the law of large numbers means that the rational representative agent model is all we need, even though I suspect that he knows better, :-).

Hi Barkley,

What I like about this paper is that it questions some behavioral findings in the same way that behavioral findings question certain aspects of the traditional rational actor model.

Sometimes one gets the impression from talking to those influenced by the behavioral literature that rational choice has been overturned by the coffee cup experiment (so to speak). I took Levine's over-arching point to be: "no, it hasn't."

In my mind, the only people who should be surprised by most behavioral results are people who, prior to behavioral experiments, took the traditional rational actor *model* as a literal description of reality in every case. I'm not one of these people, so I'm not surprised that people don't always behave in accordance with this model in lab settings. I consider the model to be just that--a model. In this sense, I've never quite gotten what all the hub-bub with behavioral economics is about.

But by the same token, demonstrating that the model isn't a literal description of reality in every case and then concluding from this that economics is "broken" seems like quite a jump to me. As a few people in this thread have already pointed out, many of the most important 'results' of economics don't require homo economicus in any event. But I've had some conversations with folks that lead me to believe that they think this is what behavioral findings show--that economics and consequently its most important results are broken. To the extent that Levine is urging us to at least reconsider this notion, I think the paper is useful.

Pete L.,

Certainly there are people who do not think homo economicus holds individually, but who argue that for many situations markets work as if they did, with Vernon Smith being an example already mentioned, and with whose views I am sure you are well acquainted (and he is a fan of Hayek), even as he is aware that there are situations where the failure of homo economicus can matter, as with speculative bubbles.

As I think you know I see lots of papers on all this stuff, and a nice summary of the evidence would be that there are indeed individuals who do follow a pretty narrow version of homo economicus (good chess players can even do many levels of backward induction, so bring on Tyler Cowen and Ken Rogoff, although very few others go more than a step or two into the future). However, we also know now that they do not necessarily dominate markets as some old theorems used to argue. Noise can dominate, with herding leading to disasters, and as we have been reminded, this can be very important.

I would agree that some of the rhetoric of behavioralists has been overblown. For sure it is not "economics" that is "broken," only a certain version of economics, and it is indeed broken (and I think that even Lucas and Becker recognize this, at least in private, or so I have been told).

Regarding Plott and Zeiler, their results are well done and important. But I fear they have oversold them, and Levine has bought into this overselling.

As for his paper, my opinion of it holds. Maybe it is worthwhile having somebody poke at overblown rhetoric by behavioralists, but his paper is so poorly done as to be an embarrassment. I could have listed even more problems with it, but will not bore everybody to death here.

Oh, a bottom line on models assuming homo economicus. I have always thought that they are useful to study and think about, if as nothing more than benchmarks of a sort. The problem has always been when people start going around saying this is how the world works, or the world works as if these models are true (for reasons such as law of large numbers or whatever), and it cannot be denied that the profession has been dominated by such arguments for decades, and certain areas of it continue so to be, even if this is weakening before our eyes now.

The comments to this entry are closed.

Our Books