You Get What You Pay For

Be careful about provoking scientists.  They can be persistent.

The issue I’m writing about today was first raised a quarter century ago when Henry Herman (“Heinz”) Barschall published two papers in Physics Today about the cost of academic journals.  He performed analyses of various physics journals at the time, based on some widely used metrics that attempt to quantify the “value” of a journal in terms of its reach and citation frequency.  His analyses showed that according to his criteria, physics journals published by the American Institute of Physics (“AIP”) and by the American Physical Society (“APS”), two not-for-profit academic societies devoted to physics in the United States, were more cost effective than other journals, particularly journals produced by commercial publishers.  Copies of his articles can be found here and here.

His articles prompted a series of litigations in the United States and Europe alleging that they amounted to little more than unfair advertising to promote the academic-society journals that Barschall was associated with.  At the time, he was an editor of Physical Review C, a publication of the APS, and the magazine Physics Today where he published his findings was a publication of the AIP.  (In the interest of disclosure, I was also previously an editor of an APS publication and also recently published a paper in Physics Today.  I have also published papers in commercial journals.)  The academic societies ultimately prevailed in the litigations.  This was relatively easily accomplished in the United States because of the strong First Amendment protection afforded to publication of truthful information, but was also accomplished in Europe after addressing the stronger laws that exist there to prevent comparisons of inequivalent products in advertising.  A summary of information related to the litigations, including trial transcripts and judicial opinions can be found here.

The economics of academic-journal publication are unique, and various journals have at times experimented with different models in order to address the issues that are particular to the publication of academic journals.  They have as their primary function to disseminate research results and to do so with a measure of reliability that is obtained by their use of anonymous peer review.  Despite their importance in generating a research record and in providing information of use to policymakers and others, such journals tend to have a small number of subscribers but relatively large production costs.  Consider, for instance, that Physical Review B, the journal I used to work for and one of the journals that Barschall identified as most cost-effective, currently charges $10,430 / year for an online-only subscription to the largest research institutions, and charges more if print versions are desired.

While commercial journals have generally relied upon subscription fees to pay their costs, academic-society journals have been more likely to keep subscription fees lower by imposing “page charges” that require the authors to pay a portion of the publication costs from their research budgets.  The potential impact on their research budgets has sometimes affected researcher decisions about where to submit their papers.  More recent variations on economic models distinguish among subscribers, charging the highest subscription rates to large institutions where many researchers will benefit from the subscription, and charging lower subscription rates to small institutions and to individuals.  All of these economic models have needed to compete more recently with the growing practice of posting research papers in central online archives, without fee to either researcher or reader.  The reason that traditional journals still exist despite the presence of these online archives is that they provide an imprimatur of quality derived from their use of formalized peer-review considerations that are still relied on by funding bodies and other government agencies.

As reported this weekend in The Economist, Cambridge University mathematician Timothy Gowers wrote a blog post last month that outlined his reasons for boycotting research journals published by the commercial publisher Elsevier.  A copy of his post can be read here.  The post has prompted more than 2700 researchers to sign an online pledge to boycott Elsevier journals by refusing to submit their work to them for consideration, by refusing to referee for them, and by refusing to act as editors for them.  My objective is not to express an opinion whether such a boycott is wise or unwise.  The fact is that journals compete for material to publish and for subscriptions.  While they accordingly attempt to promote distinctions that make them more valuable than other journals, they are also affected by the way their markets respond to those distinctions.

What is instead most interesting to me in considering the legal aspects of this loose boycott is the extent to which it is driven by a general support among commercial publishers for the Research Works Act, a bill that was introduced in the U.S. Congress in December and that would limit “open access” to papers developed from federally funded research.  The Act is similar to acts that have been proposed in 2008 and 2009.  Specifically, it would prevent federal agencies from depositing papers into a publicly accessible online database if they have received “any value-added contribution, including peer review or editing” from a private publisher, even if the research was funded by the public.

The logic against the Act is compelling:  why should the public have to pay twice, once to fund the research itself and again to be able to read the results of the research it paid for?  But this logic is very similar to the arguments raised decades ago against the Bayh-Dole Act, which allowed for patent rights to be vested in researchers who developed inventions with public funds.  The Bayh-Dole Act also requires the public to pay twice, once to fund the research itself and again in the form of higher prices for products that are covered by patents.  Yet by all objective measures, the Bayh-Dole Act has been a resounding success, leading to the commercialization of technology in a far more aggressive way than had been the case when the public was protected from such double payments — and this commercialization has been of enormous social benefit to the public.

The Bayh-Dole Act has been successful because it provides an incentive for the commercialization of inventions that was simply not there before.  This experience should not be too quickly dismissed.  In the abstract, the Research Works Act is probably a good idea because it provides an incentive for publishers to provide quality-control mechanisms in the form of peer review that allows papers to be distinguished from the mass of material now published on the Internet without quality control.  This is worth paying for.  But the Act also has to be considered not in the abstract, but instead in the context of what not-for-profit academic societies are already doing in providing that quality control.

The ultimate question really is:  Are commercial publishers providing a service that needs to be protected so vitally that it makes sense to subject the public to double payment?  The history of the Bayh-Dole Act proves that the intuitive answer of “no” is not necessarily the right one.  But what the commercial publishers still need to prove in convincing the scientific community that the answer is “yes” is, in some respects, no different than the issue Heinz Barschall raised 25 years ago.