Gold Conversion: A Prisoners' Dilemma?

---

SUMMARY: Given the undeniable, irreversible and growing clamour for Open Access (OA) worldwide, journal publishers face two Prisoners' Dilemmas. (1) The first concerns whether to continue business as usual, to mounting opprobrium from the academic community as well as the tax-paying public, or to convert directly to Gold OA now, at the risk that institutional subscriptions at current prices for incoming journals may not transmute stably into institutional "memberships" for outgoing article publication costs at the same institutional price. If publishers convert from institutional subscriptions to institutional Gold OA "memberships" today, they counter the opprobrium and lock in current subscription rates for a year (or whatever duration-deal is agreed with institutions), but they risk institutional memberships defecting after the duration elapses, with cost-recovery fragmented to an anarchic individual author/article level that may not be enough to make ends meet. (2) The second Prisoners' Dilemma facing publishers is that if they instead counter the opprobrium by converting to Green OA now (as 62% of them already have done), Green OA Self-Archiving Mandates may still force their conversion to Gold eventually, but because access-provision and archiving (and their costs) will by then be performed by the distributed network of mandated Green OA Institutional Repositories, the revenues (and expenses) of journal publishing then may be reduced from what they are now. (Perhaps this can all be integrated into just a single Prisoners' Dilemma -- or perhaps it is not a Prisoners' Dilemma at all: just the optimal and inevitable outcome of the powerful new potential unleashed by the online medium for the communication of peer-reviewed scholarly and scientific research.)

---

Although I no longer write much about it -- because there are strong reasons for according priority to Green OA Self-Archiving first, and I am ever fretful about doing anything that might instead help get us bogged down, yet again, in passive, pre-emptive speculation rather than practical action -- I too expect and welcome an eventual transition to Gold OA journal publishing, and have done so from the very beginning.

The question, of course, is how we get there from here. My own expectation (based on much-rehearsed reasons and supporting evidence) is that it will be the eventual cancellation pressure from mandated Green OA that both forces and funds the transition to Gold OA, with the institutional cancellation savings paying the institutional Gold OA publication fees. But this scenario is predicated on two necessary prior conditions: (a) universal Green OA and (b) universal journal cancellations.

This scenario for converting to Gold OA does not work if it is not universal; in particular, it cannot unfold "gradually" and piecemeal, either journal by journal or institution by institution. The three reasons for this are that (1) the true, fair costs of Gold OA publishing are not known at this time, (2) nor is the money available to pay for them, (3) nor (and this is perhaps the most important) would publishers be willing to downsize to those asymptotic reduced costs at this time of their own accord. Only (a) the cancellation pressure from universal Green OA, together with (b) the distributed infrastructure provided by universal Green OA -- allowing the functions (and costs) of access-provision and archiving to be offloaded from journal publishers and libraries onto the distributed network of Green OA Institutional Repositories -- will suffice to force both the downsizing and the transition, while at the same time freeing the funds to pay for it.

(My profound ambivalence about again raising this speculative hypothesis concerning the future of journal publishing at this time is that it risks delaying universal Green OA, by increasing publisher resistance to the Green OA mandates that are needed to bring OA about. Yet I keep having to resurrect the hypothesis now and again, as a counter-hypothesis, to answer equally speculative hypotheses about a direct transition from non-OA to Gold OA, neglecting the nonhypothetical, tried, tested, demonstrated and hence feasible, intermediate step of universal mandated Green OA, which is, apart from all else, an end in itself, being eo ipso 100% OA.)

The trouble with the "flip-over" hypothesis (the aggregator's-eye view proposed by then-CEO of Ingenta, Mark Rowse in 2003 -- see Peter Suber's recent summary) is the same as the trouble with the "institutional membership" strategy of BioMed Central as well as the "hybrid Gold" option offered by a number of publishers today (the author/institution can choose either conventional, no-fee non-OA publishing or fee-based OA publishing, paid for per individual article published): The reality is that today most of the potential institutional funds for paying for Gold OA (whatever the price) are still committed to paying for institutional journal subscriptions. Although the idea of locking this all in at current subscription rates, using the very same money, and just "flipping" -- from institutions as users, buying-in journals (i.e., annual collections of articles published by other institutions), to institutions as providers, paying-out for publishing their own individual articles -- sounds appealing (especially to an aggregator, and as long as we forget for the moment that the current subscription prices and publishing costs are arbitrary and inflated, not reflecting the substantial economies to be made from distributing the access-provision and archiving load across the network of Green OA institutional repositories), there is a logical problem inherent in the minutiae of this flip that make it into something of an Escher drawing:

An institution can commit in advance to paying for the buy-in of a certain yearly collection of journals for its users. But can it commit in advance to publishing, in any particular journal, a certain yearly number of articles by its authors? Are even the prior years' publication figures for that journal from that institution a valid predictor of what will be submitted by that institution to that journal the following year? And can a peer-reviewed journal commit in advance to accepting a certain yearly quota of papers from a given institution? (Is it not the referees who must decide, article by article, journal by journal?)

Is it not more likely that the yearly institutional quota of articles published in any particular journal will vary substantially from year to year, and from institution to institution? And is it not the author who must decide, in each case, where he wishes to submit his article (and for the referees to decide whether they will accept it)?

The equation does balance out, even at current prices, if the "flip" is universal. But as long as it is instead piecemeal and local to a journal or institution, it contains certain internal contradictions. While there is no universal OA, individual institutions will still need subscription access to the individual journals their users require. (This is equally true if the subscription access is transfered from the journal level to the individual article level, through "pay-per-view.") As long as an institution is paying for those annual institutional incoming content access-fees, that money is not available to pay for outgoing article publication-fees.

If an individual journal agrees to make all of an institution's outgoing articles OA in exchange for the current subscription fee, that's fine -- so far that's still just a bonus for renaming the "institutional subscription fee" an "institutional publication fee." The institution continues to get access to all the incoming articles in that journal, and, in addition, its own outgoing articles in that journal become OA: What subscribing institution would not happily agree to receiving that bonus as well, in exchange for merely rebaptizing its current "subscription charges" as "publication charges"?

But then (assuming this no-risk bonus is offered to all subscribing institutions rather than just one, and they all accept this renaming), the result would of course be that, next year, virtually all articles in that particular journal become Gold OA, for all institutions, whether or not they publish in or subscribe to that journal. So, the following year (or whenever the "membership" deal elapses), why bother to subscribe to that journal at all, especially for institutions that only publish the occasional article in it every few years?

In evolutionary biology, this is what is called an "evolutionarily unstable strategy". At the single-journal level, it is a recipe for inviting cancellations, soon. It does not scale, either across time, or across individual journals.

The same offer may sound less risky at the publisher "big-deal" level, in which it is a joint subscription to a whole fleet of journals that is at issue, rather than a single journal. But, first, if that is viable at all, it is only viable for publishers with fleets of journals. And even there, it is still the authors (not their institutions) who decide, individually, each year, in what journal they should publish. Libraries can consult annual user statistics to decide what journals to subscribe to next year, but it is not clear that this also translates coherently into author publication statistics. Again, libraries may be happy to take the Gold OA bonus in exchange for just renaming their fleet-subscription fees "publication fees" today, but what happens in subsequent years, when it is author statistics that are consulted on which fleets of publishing fee "memberships" to "renew"?

The system may stay stable for a while, if there is wholesale transition by most journals at a fleet level. In fact, initially, the ones most at risk for cancellation might then be the journals that do not offer the OA bonus in exchange for renaming their subscription fees publication fees; so this would in fact act to further universalize the transition to Gold (a good thing). But we should be clear on the fact that this exercise would have been a name-game, alongside a wholesale voluntary transition to Gold OA publishing on the part of publishers, with libraries ready to commit to pay for it at current rates, for now, as "membership fees."

(For the subscribing institution, the fee-based "product" was incoming journals or fleets of journals; but for the publishing institution, the fee-based "service" is based on individual outgoing articles, each in its own author's chosen journal. A "flip" here would be rather like all countries agreeing to pay McDonalds, Burger King, etc. a flat annual rate out of taxes for all the burgers their tax-payers eat annually, based on their running national averages for the latest N years: Fine for the fixed big-mac-eating tax-payer, perhaps, but not for the ones who never touch the stuff, or prefer more wholesome fare for their money. And that's without taking into account that this would also lock in current prices in a way that is impervious to supply and demand; or the possibility that it could prove a lot cheaper to produce burgers some other way, some day. McDonalds' promise to "pass on" any future economies to the consumer would sound pretty hollow in this captive-market "membership" arrangement.)

Nevertheless, I'd certainly be happy if this could all be agreed quickly and amicably, between publishers and institutional libraries: But can it? Or would publishers, in a kind of prisoner's dilemma, worry that institutions might then defect on some of their journals -- the ones they currently subscribe to and use, but in which their authors do not publish much? The prospect of such selective "cancellations" might well be enough to keep publishers from making the first move, preferring instead to stick with subscriptions and just offer hybrid OA (as many already do) as an option, at an extra institutional fee per article, with no risk to the publisher, rather than as an unconditional freebie in exchange for the current subscription fee (simply renamed), relying on faith that "memberships" will stay loyal in the long term even after everything becomes OA.

I can't second-guess the outcome of this prisoner's dilemma concerning voluntary publisher conversion to Gold OA, but I can already say confidently that the current option of hybrid Gold OA won't scale, because there isn't the extra money to pay the extra OA fees while the potential money for paying them is still paying for subscriptions. So hybrid Gold OA fees will remain just an occasional extra bonus to publishers (and an extra expense to institutions).

The one thing that just might encourage publishers to make the full transition to Gold OA voluntarily, however, is the worry that if they wait to make the transition under the anarchic pressure of Green OA self-archiving and self-archiving mandates at the article level, then the transition may indeed come with a forced downsizing and loss of income, as I have hypothesized, whereas if they convert voluntarily now, at the journal level, then they might hope to "lock in" current prices for a while longer yet. This is in fact a second prisoner's dilemma, and I certainly can't second-guess its outcome either, except to say that if it does drive the transition, then it will have been the prospect of Green OA mandates that induced the transition, rather than the actual practice of Green OA mandates -- but the cause will still have been the Green OA mandates!

What the research community must not do in the meanwhile, however, is to just sit passively, waiting to see whether or not the publisher and library community resolve their Prisoners' Dilemma(s) in favour of Gold OA. Rather than "waiting for Gold," I hope we will continue pushing full-speed for 100% Green OA by mandating it. That way we win, regardless of how the Prisoners' Dilemmas are resolved. The Gold OA dilemma, after all, is between the publishing community and the library community, whereas Green OA is entirely between the research community and itself.

Harnad, S. (2007) The Green Road to Open Access: A Leveraged Transition. In: Anna Gacs. The Culture of Periodicals from the Perspective of the Electronic Age. L'Harmattan. 99-106.

Stevan Harnad
American Scientist Open Access Forum