Do publishers simply waive their exclusive rights without any resistance?

Question

Assumption based upon my experience from CS: Typically, the business model of academic publishers is to archive and disseminate papers whose authors have transferred exclusive publication rights to the publisher. Sometimes, there are limited provisions such as authors being allowed to host a copy on their personal website. In general, though, the revenue of publishers is based upon providing paid access to the papers.

This is reflected by the choice authors often face upon publication:

• Transfer exclusive publication rights of the paper, pay nothing.
• Retain publication rights so as to provide open access to the paper, pay a (usually substantial) fee¹.

Now, the fee in the latter case is obviously the way publishers make up for their loss in revenue by not being able to restrict access to the paper to paying subscribers. This is discussed e.g. here and here.

Point of this question: Recently, the University of California has announced a "Presidential Open Access Policy" (other announcement, other article). In short, it sounds like all university employees are asked to transfer rights to their papers to the university prior to transferring rights to any commercial publisher. Furthermore, the announcements make it sound as if the consequence were that publishers simply do not receive exclusive publication rights and the university provides open access to the papers in addition to the publishers, all else apparently equal.

Now, to me, this raises a huge question mark and some speculation. In one question, I am asking here: How does this actually work? But to make things clearer and more explicit, I'm going to list the various facets of this big question one by one:

• To my knowledge, universities weren't the obstacle so far when it came to publishing as open access. Publishers were, because they wanted exclusive rights or a substantial fee, as described above. How is it that a university policy can change this?
• Do publishers readily provide a modified license transfer agreement because of such a policy? One that does not insist on exclusive rights? Otherwise, granting rights to the university first, and then signing an exclusive copyright transfer agreement sounds like the author is on their way to committing copyright infringement (or at least a breach of contract) themselves, rather than a witty way around the publishers' restrictions.
• Is this actually a legal loophole that allows authors to circumvent the publishers' wishes? If so, it would seem quite an obvious loophole ("Don't want to transfer exclusive rights? Easy. Just grant someone else rights before."); why isn't everyone doing it? Is the University of California administration just the first to have the idea for some reason?
• Do publishers just comply with this because the University of California is a comparably "big player" (as the press release states, "the UC system is responsible for over 2% of the world’s total research publications")? Even if so, what do publishers gain from complying? One might say they're not getting less, they're getting absolutely nothing.
  • And even if so, how would UC threaten them? Would they threaten to use only their in-house publisher, which offers free open access hosting in accordance with the policy?
    • Does the in-house publisher actually offer free open access hosting? It seems like UC Press relies on its sales just like commercial publishers.
  • If so, would the in-house publisher offer that hosting also to non-UC employees? Otherwise, it would seem unlikely that any journal or conference would switch to that in-house publisher, hence such a boycot would amount to banning oneself from participating in research for at least some months (if not years), until the boycot has an effect.
• Is the University of California simply testing its limits here, as a part of an ongoing legal battle of varying degrees of copyright enforcement by publishers and varying degrees of copyright infringement by authors and their institutions?
• Does the policy actually mean that all university departments are obliged to pay the open access publication fee to the publishers (i.e. requests by employees to get reimbursed for such a fee cannot be dismissed)? I was told (and hence cannot provide an accessible reference) for wealthy universities like the University of California, the open access publication fees in the lower thousands of dollars per paper might be negligible (?)
  • The latter speculation might not apply completely. Other wealthy universities are having issues to pay for subscriptions, and in numbers, these issues seem to be well in a range where one might get by paying all those open access fees, as well. But ... is this maybe a bureaucracy thing? Library subscriptions have to be paid from some (relatively limited) permanent account, whereas open access fees would apply upon publication and can thus be paid from (possibly much more generous) research grants?
• Or of course: Am I missing anything crucial about the announcement?

¹: For instance, ACM charges (for non-members) USD 900 per conference paper and USD 1,700 per journal paper.

Answer 1

To my knowledge, universities weren't the obstacle so far when it came to publishing as open access. Publishers were, because they wanted exclusive rights or a substantial fee, as described above. How is it that a university policy can change this?

Many publishers already allow self-archiving, partly as a result of pressure from authors who want to freely allow others to access their research. University open access policies and open access policies of funding agencies (like the NIH) are another way to exert pressure on publishers to allow self-archiving, deposit in institutional repositories, and other forms of open access.

---

Do publishers readily provide a modified license transfer agreement because of such a policy?

Some do. From the Columbia FAQ:

For example, the Columbia Libraries have identified the top 20 journals in which articles by faculty and staff at Lamont-Doherty Earth Observatory have appeared most often in recent years. Of those 20 journals, 19 of them include in their standard agreement a provision allowing deposit of some version of the article with a university repository.

For publishers whose copyright transfer agreement conflicts with the institution's policy, authors may include an addendum with their signed copyright agreement.

If a publisher is unwilling to accept the OA terms, authors can opt out of the institutional policy. According to Oregon State University:

A small number of publishers require that authors at institutions with open access policies receive a waiver from the policy. The following publishers have informed the library that they require a waiver: ARRS, Imprint Academic, Institute of Physics, JTE Multimedia. AAAS and Nature Publishing Group also routinely require waivers from university Open Access policies.

MIT keeps a list of publishers' responses to the policy, including whether or not authors must submit an addendum and whether or not authors must opt out of the institutional open access policy. The University of California has a list of the number of waivers requested by UC authors for various publishers between August 2013, when the UC-wide policy was announced, and August 2015.

---

Is this actually a legal loophole that allows authors to circumvent the publishers' wishes?

The University of California is not the first to enact an open access policy for faculty. See for example MIT open access policy, Oregon State open access policy, Duke open access policy, etc. To the best of my knowledge, these have not yet been tested in court. However, there does seem to be a legal basis for the idea that a license granted to the university will persist even if an author later signs a conflicting copyright transfer agreement:

Section 205(e) of the Copyright Act provides that a prior nonexclusive license evidenced in a writing signed by the right holder prevails over a subsequent conflicting transfer of copyright ownership, so the answer appears to turn on whether permission mandates satisfy the requirements of § 205(e).

(Source: "Copyright and the Harvard Open Access Mandate", Eric Priest, Northwestern Journal of Technology and Intellectual Property, 2012)

It is feasible that

permission mandates satisfy the requirements of § 205(e) and establish the license’s priority over the subsequent transfer of copyright ownership largely because they fulfill the underlying purposes of § 205(e) by providing sufficient evidence and notice of the license to potential copyright transferees (typically academic publishers).

The legal rationale is beyond the scope of this answer, but can be found here.

Note that there are steps a university can take to make sure to satisfy these requirements, so that their license will be durable. For example, Harvard had all faculty sign a paper agreement to satisfy the requirements of § 205(e).

I would rather expect that granting a license to the university and transferring copyright later to a publisher would mean that the latter action is invalid, as the author has probably (?) forfeited the option to transfer copyright due to the previously granted license. (It's like selling something you do not own any more.)

Not really. As long as you hold the copyright, you can transfer it. If you grant someone a non-exclusive license, you still hold the copyright and are free to later transfer the copyright or grant an exclusive license to someone else. (An exclusive license is treated as a transfer of rights by the law.)

If you do so, the person with the non-exclusive license can still use it under the terms of that license (assuming a non-revocable license). The new copyright holder or person with the exclusive license can enforce its rights and prevent anyone except for the non-exclusive license holder from infringing.

You cannot, however, transfer copyright (or exclusive license) twice. That would be equivalent to selling something you don't own anymore.

This is just like the way in which an author can grant arXiv a non-exclusive license to distribute, and still transfer the copyright to a journal later. From the arXiv license FAQ:

However, granting rights for arXiv to distribute an article does not preclude later copyright assignment. Authors are thus free to publish submissions that already appear on arXiv. Authors may wish to inform the journal publisher that a prior non-exclusive license exists before transferring copyright or granting a publication license.

(Of course, a publisher is free to decide as a matter of policy that they're not interested in publishing content that has previously been licensed to someone else, in which case an author at an institution with an OA policy may take advantage of the waiver mentioned above.)

---

Does the policy actually mean that all university departments are obliged to pay the open access publication fee to the publishers

No. The author (or author's institution) does not pay open access fees to the publisher. The publisher is not making the article available as an open access article.

The University of California FAQ clarifies:

My publisher charges $____ for open access. Do I have to pay that to comply with UC’s OA policies?

No. The publisher charges those fees to fund open access publication of your article at the journal’s website, but there are two ways to make scholarship open: through publisher-hosted OA (which sometimes involves fees) and through self-archiving by an author. UC’s OA policies use the latter route, by reserving rights for authors to include the author’s version of their articles in an open access repository like eScholarship. There is no fee associated with this self-archiving function. Authors may choose to pursue paid, publisher-hosted OA for their own reasons, but that is not required or suggested by the UC OA policies.

---

What do publishers gain from complying? One might say they're not getting less, they're getting absolutely nothing.

It's not entirely clear yet what the effect of green OA is on publisher revenue. There is some evidence that the number of downloads of an article from the publisher's site is reduced, but no evidence of cancelled subscriptions. Some journals have found that certain kinds of OA policies (e.g. OA after a short embargo) increases subscriptions and submissions. You can find more information here.

Answer 2

I will quote from the "Senate" policy, which now extends to non-faculty.

Each Faculty member grants to the University of California a nonexclusive, irrevocable, worldwide license to exercise any and all rights under copyright relating to each of his or her scholarly articles, in any medium, and to authorize others to do the same, for the purpose of making their articles widely and freely available in an open access repository. Any other systematic uses of the licensed articles by the University of California must be approved by the Academic Senate. This policy does not transfer copyright ownership, which remains with Faculty authors under existing University of California policy.

They do not transfer rights to the university, they simply allow the university to distribute copies for free (and only for free). This effectively forces all academic articles to be openly accessible. I expect that as a consequence of this policy, and as such policies spread, publishers will have to modify their copyright policies and drop the unconditional transfer of rights (which a number have done). In disciplines other than CS, licensing rather than transfer may be more common.