Skip to content

Arrow’s theorem update

Someone pointed me to this letter to Bruno Frey from the editor of the Journal of Economic Perspectives. (Background here, also more here from Olaf Storbeck.) The journal editor was upset about Frey’s self-plagiarism, and Frey responded with an apology:

It was a grave mistake on our part for which we deeply apologize. It should never have happened. This is deplorable. . . . Please be assured that we take all precautions and measures that this unfortunate event does not happen again, with any journal.

What I wonder is: How “deplorable” does Frey really think this is? You don’t publish a paper in 5 different places by accident! Is Frey saying that he knew this was deplorable back then and he did it anyway, based on calculation balancing the gains from multiple publications vs. the potential losses if he got caught? Or is he saying that the conduct is deplorable, but he didn’t realize it was deplorable when he did it?

My guess is that Frey does not actually think the repeated publication was deplorable; he’s just saying this because he’s broken the rules and he has to apologize. Just like if I’m stopped by a cop for running a red light, I’d apologize till my lips turned blue even if I didn’t think I did anything wrong.

In the self-plagiarism case, I’m guessing Frey’s thoughts are a mixture of:

1. There’s nothing really wrong with publishing the same material in different places. Five journals represent five different audiences, and if it’s good work, why not try to reach more people? The only trouble is that nowadays nobody reads journal articles anyway. Journal publication is little more than a stamp of approval. And what’s the point of getting five stamps when one will do?

2. Everybody does it. It’s like doping in the Tour de France: if you don’t bend the rules, you’ll fall behind.

What do I think? It looks like Frey and his collaborators broke the rules, got caught, and are paying the price. The research in question doesn’t seem like anything special; my guess is that all these journals accepted the paper because (a) it was written in an “economics” style and seemed professional, (b) the topic of the Titanic seemed like a fun diversion, and (c) they weren’t aware of the many papers that had already been written on the topic. The Frey et al. paper is ok, worth publishing once, perhaps, but it’s not really good enough to be published in five different places. The econ journals should reserve this treatment for the very best research.


  1. zbicyclist says:

    “Frey and his collaborators broke the rules, got caught, and are paying the price.”

    What price? Are they on probation, and therefore their entire department is unable to publish in a “bowl quality” journal for 3 years? Are they restricted in the number of grad students they can recruit? Are they prohibited from appearing on TV?

    It’s not much of a crime, but they’re not paying much of a price, either.

    • Andrew says:

      The benefit to publishing your paper in a top journal is that you get a reputation as a top researcher. The penalty here is that Frey is now the butt of a joke, and he has lost much of his reputation. To me, the punishment and the wages of the crime are in the same coin.

  2. Anon says:

    This may well be the main point “Journal publication is little more than a stamp of approval”

    Seems so for some journal editors – hey someone has to keep the barbarians out ;-)

    But publication counts presume independence – right?

    Might be neat to study citation patterns for these multiple published papers – before and after the whistle is blown.

  3. A. Tsai says:

    A related phenomenon is publishing the same result not in English language journals in different fields, but in different language journals in the same field. For example, Chinese academics do this all the time:

  4. A. Tsai says:

    “The only trouble is that nowadays nobody reads journal articles anyway. Journal publication is little more than a stamp of approval. And what’s the point of getting five stamps when one will do?”

    Is that the case in all fields? Different fields have different models of publication. In economics (and sociology?), for example, the bulk of the reputation-building occurs in the brown-bag lunch conferences. You spend 3 years presenting at seminars like these around the country, revising your paper, posting it on your web site, etc. By the time you submit it for publication in a top journal like JPE or AER, getting the actual publication is more or less an afterthought.

    Contrast that with medicine, where journals make money from drug ads, and there is a much greater emphasis on scoops and getting picked up by the media. Journal editors don’t like it if you present the findings at more than one conference, and for major publications being presented at scientific meetings they may even want to coordinate publication with the conference presentation. Here the publication is most certainly not an afterthought.

  5. afinetheorem says:

    German-speaking countries rank professors/schools using a system that is not too far from publication counting, which probably explains why he did it. Also, we now know of other cases of self-plagiarism involving this author. A letter like this from Autor in a flagship journal is a major punishment indeed.

    What still is unresolved is the earlier article not by Frey et al (nor cited by them) on actions under stress on the Titanic. Basically, the *very first* article you’ll find if you search Google Scholar for something like “class” and “Titanic” or “social class” and “Titanic” is an article very similar to that by Frey written in 1986. Even just searching “Titanic” restricted to social science articles, it’s the 12th entry! It’s possible this was just missed, but the author of the 1986 article is a professor in Brisbane, Australia, the same city as Frey’s two coauthors work in. That may just be a coincidence, but that seems like quite a coincidence indeed.

  6. Berry says:

    In the last days, a number of people have collected evidence that suggests that the titanic story is not a singular event but rather a common practice by Frey et al. A number of articles with quite similar content have been published without cross-referencing (see I never thought a person of Frey’s standing needs this. Moreover, it casts doubt on the sincerity of his apologies.

    • Andrew says:


      As noted above, I have no reason to think that Frey’s apology is sincere. But is sincerity required in an apology? I think all that is required is that you say the words.

  7. Hans says:

    Hi Andrew, please have a look at these two papers in JPE and AER by Bruno Frey and Felix Oberholzer-Gee (at Harvard):

    They are almost identical (but rephrased) and do not cross-cite

  8. […] works, Statistical Modeling, Causal Inference, and Social Science, musing about the honesty of an apology offered by an author who plagiarized himself, Computational Complexity gives examples for how to get ideas for things to […]

  9. Hans says:

    Please see this new article on many other instances of self-plagiarism by Frey:

  10. error says:

    So Andrew, did you make a mistake in your first blog on this issue? In the first blog you ridiculed the affair and said Frey did not do anything wrong.

    • Andrew says:


      I don’t think I ridiculed the affair but, yes, I made a mistake in saying that I didn’t think Frey did anything wrong. I was sympathetic to Frey’s research findings and I’d also met him personally—just once, but he seemed nice and also very intelligent and alert—so I was inclined to be supportive on the grounds that it is important to reach different audiences. So I put in that last sentence (“P.S. Just to clarify: No, I don’t think Bruno Frey did anything wrong”) so my post would not be seen as an attack on Frey. But, really, who was I to judge based on that scant evidence whether he did anything wrong? And, if he did do something wrong (as it appears he did), that doesn’t invalidate all his work, it just means that one has to view his apparent productivity in a different light.

      In retrospect, it would’ve been better for me to simply say: (1) I don’t think self-plagiarism is as bad as actual plagiarism (it’s the difference between taking drugs to win a race, vs. hurting someone else to make them lose a race—Lance Armstrong vs. Tonya Harding, if you will—and (2) I didn’t have enough information to judge whether Frey just went overboard in publishing similar papers in similar places, or whether he was systematically breaking the rules.