I’ve recently had a number of conversations about the value and incentives of publishing in the top 5 economics journals. Why do people try so hard to publish in AER/QJE/JPE/ReStud/Econometrica? Why do departments and the profession reward them for it? For a fun article on Top5itis, click here. For an excellent discussion on the issues by top economists, click here (video of a 2017 AEA session). For a discussion of the information problem that “top” journals help solve, keep reading.

Here’s my theory: the only reason to care about journals in the internet age is because they act as a signal of a paper’s quality (this view is also echoed by one of the economists in the aforementioned video).  This is a blog, so don’t ask me for a formal model, but here’s an informal proof by counterexample.

Suppose we lived in a world where the quality of a paper, however you want to define “quality” were perfectly observable. Why would we ever care if a paper published in journal A or journal B? One reason is if journal A were read more widely, in which case your articles would have a bigger impact there. But in a world where paper quality is observable, why would journal A be read more widely?

I can think of three reasons. All of them assume that it’s hard to read individual articles from many journals, so you pick your journals first and your articles to read second. Without this assumption, I struggle to think of any reason to care about journals. So here they are:

  1. Historical lock-in from a time when quality wasn’t observable.
  2. Some sort of market power by the publisher of journal A that results in journal A being foisted upon all of us.
  3. Better turnaround times, editors, or other features of the journal publication process that attract more high-quality papers to journal A.

Clearly, (a) and (b) are bad, while having more competition on the speed and quality of the publication process would be a welcome thing. Also, pretty much every part of the review process is easily replicable except editor quality, so in equilibrium we might see journals improve a lot on all dimensions except this one.

Of course, paper quality is not observable, so the main point of this discussion is to highlight how valuable such quality measures would be. Departments could fret a lot less about how many “A-level” publications you have and so could you. Starting a new journal would likely become a lot easier. World hunger would be history [cross out]. Frankly, you wouldn’t even have to publish your paper!

But, alas, quality is difficult to observe, unless a paper happens to be in your field and you have read it carefully. Rating papers by the number of cites is problematic for at least two reasons. First, it preserves journals’ advantage: I’m 95% confident that there’s a causal effect of publishing in a top 5 on citation counts. Second, it would take a long time for paper quality to be revealed, at which point we’re all dead.

Could a well-designed rating system work to reveal paper quality and move us away from caring about journals? I think so – look at all the people who are providing public goods on Quora, Wikipedia, Amazon, etc. There’s already a recommendation systemalong these lines for biology and medical publications, though it’s hard to judge how well it works. Because stakes in academia are so low, there’s reasons to think that letting ratings be a free-for-all might go terribly wrong, but we’ll never know until we try! In the meantime, I’ve gotta go and try again for that top 5 publication.