On Academic Publishers

This seems like a strange system to me. Publishers, despite assertions to the contrary, add little value to academic research. They run websites that host journal articles and make them searchable. The most complex and costly part of these websites are the security features that keep people from accessing articles. Serving static files (pdf files mostly) with an attractive front-end that makes them searchable is trivial from a technical perspective and could cost little.3 Printing, binding, and mailing print copies of each issue has obvious costs, but printing is unnecessary.

Publishers don't usually develop and maintain the software that authors use to submit journal articles and track them through the peer-review process (though they license it) and, in any case, it isn't clear this software is value added. There are also a number of free (as in speech and beer) solutions to this.

Publishers also do typesetting and copy-editing. The typesetting is often nice, but could be done easily by anyone with a knowledge of TeX. There are also a variety of ways to automate transalation of documents in various forms (docx, tex, md, etc) to whatever form desired (e.g., pandoc), and, in any case, journals are generally restrictive about the formats you can submit your article in, which is frustrating, since the article may very well be rejected. Publishers copyedit accepted articles, but again, they don't always do a great job (I see typos in published articles all the time).

For-profit academic publishers are a relatively new phenomena.4 The first academic journals were published by non-profit academic organizations (such as the Royal Society, not for-profit publishers). Since academic publications in the form of scientific journals have existed since at least the 17th century, profiting on the publication process itself is a relatively new phenomena. Because of the internet, the costs of disseminanting information should be lower than when scientific journals first appeared. For example, the Journal of Machine Learning Research, which is open-access, has no publication fees, and has a print edition, has a per-article cost of around $6.50. Political science journals could be (and should be) run similarly.5

I don't think that journals themselves will go away since having a research quality hueristic is useful, despite the fact that the process is capricious. Whatever happens, we should own our own work. The new open-access political science journal Research and Politics is a good start. To get rid of the for-profit publishers each discipline has to shift prestige to open-access publications (which are generally new). This is a collective action problem that will be difficult to solve because prestige is sticky, and is important to hiring and tenure decisions. Those in a position to take risks by submitting things to open-access journals that might be published in (currently) more prestiguous journals run by for-profit publishers should do so. The more high-impact research that is published in open-access journals the easier it will be to convince less secure researchers to submit things there.


  1. Penn State's library devoted 5% of its budget to journal subscriptions in 1997 and 53% in 2008 according to this report

  2. This isn't always the case. Wiley charges around $700 a year for an instituional online-only subscription to the American Journal of Political Science. Cambridge doesn't publish the cost of a subscription to the American Political Science Review

  3. It should be noted that almost all journal websites are ugly. In my experience their search is also quite bad. 

  4. According to Wikipedia the trend began in the 1960s and 1970s. 

  5. This is in stark contrast to the reported $526 average per-page cost of an article published in a flagship journal in the social sciences or humanities (according to the Chronicle of Higher Ed).