The Occasional Pamphlet

Recently, the representative of a major scientific journal publisher expressed to me the sentiment that the position that Harvard faculty have taken through our open-access policies — setting the default for rights retention to retain rights by default rather than to eschew rights by default — is in some sense unfair to subscription-based journals that require embargoes, that we are favoring one scholarly publishing business model over another and setting up an unlevel playing field.

By way of background, the Harvard open-access policies specify that faculty authors grant a nonexclusive license to the university to distribute our articles, which can be waived for any reason at the sole discretion of the authors. The license applies immediately upon copyright vesting in the article, and thus predates any transfer of copyright to a publisher. If a publisher has a policy that is inconsistent with this license — for instance in requiring that no distributions occur until expiration of an embargo period — then it must either make an exception for an article falling under the OA policy or get the author to obtain a waiver of the license. Open-access journal publishers, who do not mandate embargoes on distribution, will not need to engage their authors in obtaining waivers. The publisher in question thought that this difference was an unfairness toward embargo-carrying subscription-based publishers because it favored open-access publishers.

Of course, as a private company, Harvard is well within its rights to set up its policies to favor whatever it wants, even if third-parties are disadvantaged. For instance, Harvard has policies that disallow faculty performing research for funders that have certain kinds of policies, even though that disadvantages such potential funders. But I for one am not interested in unfairly advantaging one business model for scholarly publishing over another, and in any case, the argument that the Harvard policy does so is fallacious.

Before describing why, let me mention again that there is a different issue that I will not address here, that immediate open access of author’s manuscripts endangers the subscription-based business model. That is a separate issue, with a set of standard arguments and counterarguments, which have convinced me that the only empirical evidence is to the contrary, though others may not be so persuaded.  But for the purpose of this discussion, let us stipulate (counterfactually to my mind) that wholesale adoption of the Harvard policy would in fact prove detrimental to the business of subscription-based journals at least a little bit.  Does that make it intrinsically unfair?

This other argument is different and seems to rely on an assumption that the Harvard policy is an undue intervention, whereas without the policy, there is no intervention. It seems to go something like this: In designing a system of rights allocation between author and publisher, the approach in which all rights are transferred to the publisher with the publisher selecting some rights to provide back to the author is the privileged position, and any other arrangement involves some kind of intervention. But what makes that the privileged (nonintervening) position? Why isn’t the privileged position an approach in which all rights are transferred to the publisher subject to a nonexclusive limited noncommercial license to the author, with a publisher’s requirement for exclusive transfer of all copyright being the intervention? It is true that one was the status quo ante for some time, but that is a historical contingency, and gives it no privileged position as consisting of the appropriate default position. There is in fact no privileged position one way or the other.  All choices of design involve intervention.  There are just a large variety of possibilities, which the actions of various stakeholders choose among.

Another way of thinking about the issue is this: The faculty must make a decision about what portion of their rights they retain by default: none, a few, a lot, this subset, that subset.  A decision is inevitable; there is no option of making no decision.  The status quo ante made a decision, namely that no rights were retained by default.  The Harvard policy makes a different decision, that a nonexclusive limited license is retained by default; some other policy might state that exclusive rights are retained by default.  But one way or another, a decision must be made.  The faculty are obviously inclined to make decisions that benefit them as they perceive it, and have done so.  The change is not from making no decision to making a decision, it is from making one decision to making another that the faculty perceives as preferable.

Sunstein and Thaler make this argument in detail in their paper “Libertarian Paternalism is Not an Oxymoron” (a precursor to their popular book Nudge).  They point out that in a wide variety of cases, decisions as to a default rule are inevitable, so that changing the default does not constitute a new intervention.

They use as an example a company cafeteria presenting a variety of foods.  “Suppose that the director of the cafeteria notices that  customers have a tendency to choose more of the items that are presented earlier in the line. How should the director decide in what order to present the items?” If he places the desserts first, the employees will eat more desserts, which will have an aggregate negative effect on the health of the employees, to the detriment of the company. If he places the desserts last, they will eat fewer and be healthier, perhaps to the detriment of the dessert supplier. If the cafeteria changes its placement of desserts from the front of the line to the back of the line, the dessert supplier may rightly claim that the change will hurt its business, but it can’t claim that this is an undo intervention in the freedom of choice of the employees.  The original cafeteria layout was not privileged; the company has to place the desserts somewhere.  Why is the front of the line the privileged position just because of historical familiarity?  It isn’t.

We all make decisions about policies, defaults, and the like.  So long as the designs of the systems do not preclude options (and the Harvard open-access policy certainly does not), they protect the liberties of the participants.  The particular design may advantage or disadvantage some participants.  I do not think there is any evidence that the Harvard policy, even extrapolated to all universities, disadvantages subscription-based publishers.  But even if it were to, that does not make it intrinsically unfair relative to a given alternative.

I appreciate that some publishers may not like the waiver process, and I do not begrudge them that view or think it is unreasonable.  But preferring X to Y doesn’t mean that Y is unfair.