Self-Archiving Mandates

---

SUMMARY: The actual impact of Open Access (OA) self-archiving on research, researchers, their institutions, their funders, and the tax-paying public (which has all already been shown to be highly positive) must be clearly separated from any hypothetical impact it might have on publishers (whether commercial or scholarly-society publishers). Researchers do not conduct research -- nor does the tax-paying public fund research -- for the benefit of publishers. The sole point at issue concerning the FRPAA is whether or not self-archiving should be mandated. The two concrete questions that researchers, their institutions and funders need to put to themselves regarding any "special relationship" with scholarly society publishers are therefore:
        (a) Would (or should) researchers, their institutions and their funders knowingly choose to subsidise their scholarly societies with their own actual lost research impact in order to immunise those scholarly societies from any hypothetical risk of lost subscription revenue?
        (b) If, contrary to all evidence to date, self-archiving were indeed destined one day to cause publisher revenue losses -- or even to force a shift to the open-access publishing cost-recovery model (with author-institutions paying the publication costs for their own institution's research output out of their own windfall savings from the cancellation of their former costs as user-institutions, buying in the published output of other institutions) -- is the prevention of that hypothetical outcome something that researchers, their institutions and their funders would (or should) knowingly choose to subsidise with their own actual lost research impact?
        A dissenting minority of 10 US provosts opposes the FRPAA Self-Archiving Mandate (vs. 125 in favor) on the grounds of risk to scholarly society subscription revenue. There is obviously a biomedical publisher lobby behind some or all of the 10 dissenting voices; the arguments are old ones, already rebutted many times:
        (1) The hypothesis that mandated self-archiving will force a shift to the OA publishing cost-recovery model is pure speculation at this time, with no evidence in its support, and evidence from both the American Physical Society and IOPP contradicting it.
        (2) But even if the hypothesis were ever to come to pass, it would not mean "diminishing funds available for research to benefit the public good".
        (3) To force a shift to the OA publishing cost-recovery model, there would first have to be substantial revenue losses for publishers, from institutional subscription cancellations.
        (4) But for every penny of revenue lost by publishers in the form of institutional subscription cancellations, there has to be a penny saved by institutions, in the form of windfall savings.
        (5) Hence if publisher revenue losses were ever to force a shift to the OA cost-recovery model, the institutions would have a large annual pot of windfall savings on which to draw to pay for their own outgoing publication costs.
        (6) Hence there would be nothing at all "requiring authors to pay for their publications through their Federal grants, diminishing funds available for research to benefit the public good."
        (7) It is only now -- when there are neither any institutional subscription cancellation pressures, nor any institutional subscription windfall savings -- that it looks as if paying OA publishing costs would require diverting money from research.
        (8) Hence it is both self-serving and self-contradictory to invoke both the "damage" hypothesis and the "research fund diversion" hypothesis against the FRPAA in the same breath: If the hypothetical "damage" is the hypothetical subscription revenue loss, then that is also the diversion: no need to poach hypothetical research funds.
        (9) The rationale for the FRPAA self-archiving mandate, however, has nothing to do with speculative hypothesizing about journal economics but with demonstrable facts about maximizing research usage and impact.

---

Peter Suber has already provided an excellent critique of the letter from a dissenting minority of 10 US provosts who oppose the proposed FRPAA Self-Archiving Mandate (versus the other 125 provosts who support it):

But one can never say enough in support of a good thing (and against a bad one!), so what follows below is a detailed, systematic, point-by-point critique of the dissenting provosts' position and the arguments adduced in its support.

It is fairly obvious that there is a biomedical publisher lobby behind some or all of these dissenting voices, since the statements of these 10 dissenting provosts (several of them members of executive committees of the American Physiological Society!) are almost a verbatim echo of the very same points that the publishing lobby has been making over and over, in trying to oppose self-archiving mandates worldwide [1, 2, 3, 4, 5], hence the accompanying press release from the American Physiological Society ["APhS"](not to be confused with the American Physical Society!) from which more will be heard below too [6, 7, 8].

(The attempted opposition to self-archiving mandates has already proven unsuccessful in the UK, where four of eight RCUK research funding councils have already mandated self-archiving, beginning October 1 2006.)

But never mind, we will take the points made in both the American Physiological Society press release and the letter from the ten provosts at face value:

On Fri, 22 Sep 2006, Martin Frank (American Physiological Society ["APhS"]) wrote:

SENIOR ACADEMIC OFFICERS EXPRESS THEIR CONCERN ABOUT S.2695, THE "FEDERAL RESEARCH PUBLIC ACCESS ACT OF 2006"

(APhS): (Bethesda, MD) - September 22, 2006 - Senior academic officers from 10 institutions issued a letter to Senators John Cornyn (TX) and Joseph Lieberman (CT) expressing their concerns about the provisions of S.2695, the "Federal Research Public Access Act of 2006." These institutions, which collectively make nearly $3 billion in annual research investments, expressed their concerns that mandating a six-month public release of journal articles would negatively impact the academic community and the publishers that disseminate their work.

(If the $3 billion dollar figure is pertinent at all, then the first thing to call to mind is the more than $30 billion dollars in annual research investments of the 125 institutions that had expressed exactly the opposite concern...)

But let us look at this more closely, for on the face of it, the effect of making research Open Access has already been demonstrated to have a highly positive impact on its impact (sic) -- i.e., the degree to which it is accessed, used, and cited.

So now let us hear more about the alleged downside of this -- but let us be very careful to separate its actual impact on the academic community (which we already know to be positive) from its hypothetical impact on (some) publishers.

For, lest we forget it: researchers do not conduct research -- nor does the tax-paying public fund research -- for the benefit of publishers [9, 10, 11, 12].

(APhS): In signing the letter in opposition to S.2695, Dr. Robert Rich, Senior Vice President and Dean, University of Alabama at Birmingham School of Medicine, expressed his concern that "the legislation would damage the special relationship between scholarly societies and academic communities who work in partnership to ensure that these communities are sustained and extended, science is advanced, research meets the highest standards, and patient care is enhanced with accurate and timely information."

The sole point at issue with the FRPAA is whether or not self-archiving should be mandated. All evidence so far shows that self-archiving enhances research impact. No evidence so far shows that self-archiving reduces scholarly society publisher revenues, and this has been explicitly confirmed by the two scholarly society publishers whose published contents have been self-archived the longest and the most, the American Physical Society and the Institute of Physics Publishing: They both report that they have detected no subscription losses as a consequence of self-archiving.

Nevertheless, some scholarly society publishers fear, despite the absence of any actual evidence, that self-archiving will cause "damage" if mandated by the FRPAA (and by other research funders and institutions worldwide). No one knows whether or when these fears of damage will actually come true, but let us agree that there is a non-zero risk for publisher subscription revenues here, and that it is definitely not outside the bounds of either logic or likelihood that universal availability of authors' final drafts could eventually generate cancellation pressure on subscriptions. Yet what needs to be done by the academic community is to weigh this hypothetical risk to publishers' subscription revenues against the demonstrated benefits to the research impact of researchers, their institutions, their funders, and the tax-paying public that funds them.

Hence there are two very concrete questions that researchers and their institutions and funders need to put to themselves regarding this "special relationship" with scholarly society publishers:

(i)Would (or should) researchers, their institutions and their funders knowingly choose to subsidise their scholarly societies with their own actual lost research impact in order to immunise those scholarly societies from any hypothetical risk of lost subscription revenue?

(ii) If, contrary to all evidence to date, self-archiving were indeed destined one day to cause publisher revenue losses -- or even to force a shift to the open-access publishing cost-recovery model (with author-institutions paying the publication costs for their own institution's research output out of their own windfall savings from the cancellation of their former costs as user-institutions, buying in the published output of other institutions) -- is the prevention ofthat hypothetical outcome something that researchers, their institutions and their funders would (or should) knowingly choose to subsidise with their own actual lost research impact?

(APhS): Rich also expressed concern that "S.2695 would divert scarce Federal dollars away from research in order to provide a service already provided to the public by society publishers."

It is already a speculative hypothesis that self-archiving would damage subscription revenue; but it is wildly counterfactual to say that FRPAA is about diverting funds away from research in order to pay publishers!

FRPAA is not a mandate to convert to OA publishing: it is a mandate to convert to author self-archiving. FRPAA says nothing about diverting funding from research to publication.

The hypothetical long-term sequel of mandated self-archiving -- a conversion from institutional subscription-charges to institutional publication charges -- is merely speculation, in the absence of any objective supporting evidence (and in the face of counter-evidence). But even if the hypothesis were ever to prove true, it would not mean diverting a penny of research money from research funding to publication costs! It would mean redirecting the very same money that institutions currently spend on subscription charges toward paying instead for publication charges.

(APhS): The nonprofit publishers comprising the DC Principles Coalition are among those who are able to provide public access to literature either immediately or within months of publication without government mandate through corporate and academic subscriptions.

This is playing loose with the words "public access." The purpose of an Open Access self-archiving mandate is to provide access to all those would-be users who do not have subscription access, today. Voluntary provision of OA by publishers is of course very welcome, but it is far too few and far between. And the research impact loss problem is now, and urgent.

It would be absurd for the research community to continue sustaining needless annual impact losses in order to wait passively for publisher voluntarism to decide whether and when to remedy them. If publisher voluntarism were indeed inclined to put an end to those impact losses, surely publishers would not be lobbying against the FRPAA self-archiving mandate: they would be supporting it: The distributed author mandate would be saving them the trouble of having to provide OA themselves, from their own resources!

But it is clearly for access-denial, not access-provision, that (some) publishers are lobbying here. Let there be no doubt about that, and that the voluntarism of the DC Principles Coalition is far too little, too late.

(APhS): According to Martin Frank, Ph.D., Executive Director of the American Physiological Society (APS) and a member of the Coalition, "a six-month release mandate may force some journals to shift to a publication model requiring authors to pay for their publications through their Federal grants, diminishing funds available for research to benefit the public good."

Martin Frank, Ph.D.
Executive Director, American Physiological Society
http://www.the-aps.org

(1) As noted, this hypothesis -- that mandated self-archiving will cause a conversion to the OA publishing cost-recovery model -- is pure speculation at this time, with no evidence in its support, and the prominent evidence from both the American Physical Society and IOPP contradicting it.

(2) But let us suppose, for the sake of argument, that the hypothesis should one day come to pass: does this mean "diminishing funds available for research to benefit the public good"?

(3) What Martin Frank seems to be forgetting in his calculations is that in order to force a shift to the OA publishing cost-recovery model, there first have to be substantial revenue losses for publishers, from institutional subscription cancellations.

(4) But for every single penny of revenue lost by publishers in the form of institutional subscription cancellations, there has to be a penny saved by institutions, in the form of windfall savings.

(5) Hence if publisher revenue losses were ever indeed to force a shift to the OA cost-recovery model, the institutions would have a large annual pot of windfall savings from incoming subscription cancellations upon which to draw, in order to begin paying instead for their own outgoing publication costs.

(6) Hence there would be nothing at all "requiring authors to pay for their publications through their Federal grants, diminishing funds available for research to benefit the public good."

(7) The only thing that would have happened would be the augmentation of the public good derived from research, by maximising its access, usage and impact -- whether or not the hypothetical shift in publish models came to pass.

(8) It is only now -- when there are neither any institutional subscription cancellation pressures, nor any institutional subscription windfall savings -- that it looks as if paying OA publishing costs would require diverting money form research.

(9) Hence it is both self-serving and self-contradictory to float both the "damage" hypothesis and the "research fund diversion" hypothesis in the same breath: If the "damage" is subscription revenue loss, then that is also the diversion: no need to poach research funds!

Issued on September 22, 2006, the letter reads:

Dear Senators Cornyn and Lieberman:

The undersigned senior academic officers write to express our concerns about S.2695, the "Federal Research Public Access Act of 2006."

We agree that the broadest dissemination of scientific literature is good for research.

If this were indeed clearly thought-through and sincerely meant, we could stop right here. Because "the broadest dissemination of scientific literature" is Open Access, and the FRPAA self-archiving mandate will provide Open Access.

Letter: However, mandating a six-month public release of journal articles would have negative unintended consequences for the academic community.

Are these hypothetical negative unintended consequences negative for the academic community (i.e., researchers and their institutions and users), or for the publishing community? (The two are completely conflated in what follows below.)

Letter: The free posting of unedited author manuscripts by government agencies [1] threatens the integrity of the scientific record, [2] potentially undermines the publisher peer review process, and [3] is not a smart use of funds that could be better used for research.

(0) The posting is by authors, not by government agencies: the FRPAA proposal is for the government agencies to mandate that the authors post the manuscripts.

(1) How does posting a free final, peer-reviewed, accepted draft of the author's paper for those would-be users webwide who cannot afford access to the publisher's version of record "threaten the integrity of the scientific record"?

(2) What the FRPAA proposes to require to post is the author's final, peer-reviewed draft, accepted for publication; the only thing it might be missing is some copy-editing: How does that "undermine the publisher peer review process"?

(3) And is it, then, smarter use of funds to subsidise copy-editing with lost research impact?

If copy-editing is such an important added-value, what are publishers worried about? The subscription smart-money will then just keep on paying for it, since that added-value is missing from the author's peer-reviewed final draft, which is merely a supplement for those who cannot afford the publisher's official copy-edited version of record, online or on paper.

(But please let's leave the "peer review process" out of this, because it is not even at issue. The peers review for free, as a service to both authors and publishers; their services are not what the subscription money is being spent on.)

Letter: Scientific publishers, in collaboration with academic institutions, scientists, and libraries, have been at the forefront of innovations that have improved and continue to improve access to research information. As a result, more scientific papers are now available to more people than at any time in history.

Absolutely true, and commendable, but irrelevant. Because it is not enough. Substantial amounts of potential research impact are still being needlessly lost, cumulatively, in an online age when this loss can easily be prevented, once and for all, at long last.

The supplementary self-archived author's draft is for all those would-be users whose institutions cannot afford the above-mentioned "innovations and improvements." For without the author's self-archived version, they have no access at all.

For an estimate of how many users are being denied access -- and hence how many authors are being denied impact -- simply look at the studies thay show the degree to which self-archiving enhances article usage and impact.

Letter: Even when federal funds support the research reported in journal articles, these funds do not cover the costs associated with turning raw data into archived scientific manuscripts. The cost of peer review, copy editing, formatting, printing, online publication, search engine development, and permanent archiving ranges from $2,500 - $10,000 per article.

Even without challenging those figures, one can point out that all those costs are currently being paid, in full, by subscriptions, with no evidence that self-archiving reduces those subscriptions. If and when self-archiving should ever reduce those subscriptions enough to require another way of meeting those costs, the costs will be met out of the windfall subscription savings.

But for now, this is mere speculation. The only thing that is not speculation is the demonstrated benefits of OA self-archiving to research and researchers, in enhancing research usage and impact.

Nor -- as long as we are speculating -- is it at all clear that if self-archiving were indeed ever to induce subscription cancellations, that "the cost of peer review, copy editing, formatting, printing, online publication, search engine development, and permanent archiving [$2,500 - $10,000 per article]" would all prove irreducible:

It is not only hypothetically possible but quite likely that the cost of implementing peer review [since the peers review for free] could turn out to be the only essential remaining publication cost, and that is only about $500 per article.

The value of copy editing remains to be determined empirically, but formatting, printing, online publication, search engine development, and permanent archiving look very much like the kinds of things that could readily be offloaded onto authors and their institutions, with their distributed network of OA IRs and their distributed and hence much diluted costs per article, nowhere near the current $2,500 - $10,000 figure.

"The True Cost of the Essentials (Implementing Peer Review)" (Feb 2000)
"Hypothetical Sequel"

The above is of course all just speculation too, but hypothetical speculation invariably breeds counterspeculation. The only certainty here is that mandated OA self-archiving will be highly beneficial to research usage and impact, as has already been repeatedly demonstrated empirically.

Letter: At present, publishers cover these publication costs through the sale of subscriptions. A Federal policy mandating public access after six months would threaten the financial viability of many of these journals through the loss of subscription revenues, forcing them to identify other means to cover costs.

First, to repeat: There is no evidence to date that this hypothesis is correct, even in fields that have been self-archiving at 100% for years now.

But should the hypothesis ever prove true, then, yes, it will be necessary to "identify other means to cover costs" (whatever those irreducible costs turn out to be). And the other means of covering those costs is already obvious: Author-institution payment of publishing costs out of institutions' own windfall subscription savings.

Even without cost-cutting and new efficiencies (such as phasing out the paper edition and relying on the worldwide network of OA IRs to provide access, hence leaving only the cost of implementing the peer review service), a forced shift to the OA publishing cost-recovery model after 100% OA had already been reached via mandated self-archiving would merely mean that the money that currently changes hands between institutions and publishers in the form of subscription costs would instead change hands between institutions and publishers in the form of publishing costs.

Note, though, that no matter how shrilly one raises the volume on the hypothesizing and counter-hypothesizing, it is still merely a speculation that mandated self-archiving will force a shift in publishing models. The only objective certainty is that mandated self-archiving will greatly benefit research impact.

Letter: One such means is to shift the costs to the scientists/authors. This is the business model currently used by the Public Library of Science, for example, which recently increased fees to $2,500 per manuscript. These fees either come from [1] the author's Federal research grant -- thereby decreasing the amount available for research -- or [2] from the university, which could ultimately lead to higher institutional costs than those needed for journal subscriptions.

PLoS and other OA publishers today are struggling to make ends meet in a world where spontaneous OA self-archiving is still only hovering at 15%, and 100% of institutional journal budgets are still tied up in covering subscription costs. Hence these brave new OA publishers need to find other sources to pay their OA publishing costs.

But on the above hypothetical scenario, a forced shift to OA publishing caused by mandated self-archiving would result from institutional subscription cancellations! Hence the institutions could then use their own windfall savings to pay the irreducible costs in another way: via the OA publishing model. No need to poach from either research funding (1) or other institutional resources (2).

Letter: In fact, some studies have already shown that research intensive universities would have to pay considerably more to gain access to the same amount of research under an author-pays model than a subscription model.

This is compounding speculation with speculation, since no one knows what the true costs would turn out to be, under pressure from subscription revenue declining to unsustainable levels because of institutional cancellations.

There is clearly more than enough money in the system already to sustain publication today. Research-intensive universities are also subscription-intensive universities, so one would have to see just what assumptions are being made by studies that claim that these universities would be worse off if there were ever a transition to OA publishing.

The only thing that is sure is that all research institutions would be far better off in terms of their own research impact (and access). The rest is all speculation, assumptions, and guesstimation.

Letter: Mandating free dissemination of scientific manuscripts within six months would significantly limit the ability of non-profit and commercial publishers to cover the upfront reviewing, editing, and production costs of creating these manuscripts. Some journals would simply cease to exist. Others would be much less able to support innovation in scientific publishing and archiving. Ultimately, this could lead to a system in which NIH and other federal agencies must sustain a significant portion of the research publishing enterprise, maintaining 100+ years of archival journals, as well as producing new research articles.

Not only is this merely a shriller version of the speculative scenario already mooted above, at a still higher volume, but it throws in a nonsensical and irrelevant alarum about legacy archiving, something that is not even at issue in the FRPAA self-archiving mandate, which only covers prospective author self-archiving, not retrospective journal archiving. (Let the journals hand over their legacy archives, and I'm sure the research and funder community will know what to do about them: don't fret about the cost...)

On the coverage of the prospective costs if/when subscriptions should become unsustainable, the obvious answer remains the same: institutions will cover those costs out of their own subscription savings.

Letter: As a member of the Senate Budget Committee, you are certainly sensitive to the various forces that shape and reshape the Federal budget from year to year. Recently, for example, we learned that the Biomolecular Interaction Network Database -- the world's largest free repository for proteomic data -- lost its funding and curtailed its curation efforts. As leaders in our respective academic institutions, we are profoundly concerned that one unintended consequence of S. 2695 would be to put both our current research publications and our research archives in jeopardy.

The FRPAA self-archiving mandate devolves on the distributed network of institutional repositories (IRs) of all the US research institutions. In the unlikely event that someone would ever want to pull the plug on a major central repository such as PubMed Central, the primary research providers, the universities themselves, are certainly likely to become more and more reliant on the IRs, rather than less and less, in our distributed online, networked age. And the costs of creating and maintaining individual OA IRs are so risibly low that it is absurd even to discuss them.

So the obvious and optimal prophylaxis against any risk of central funding loss for central archives is for researchers to do all their self-archiving locally, in their own institution's IR, and let central collections be harvests from those IRs, rather a locus for direct central self-archiving.

"Central vs. Distributed Archives" (began Jun 1999)
"PubMed and self-archiving" (began Aug 2003)
"Central versus institutional self-archiving" (began Nov 2003)
"Harold Varmus: 'Self-Archiving is Not Open Access'" (began June 2006)

(APhS) Given the widespread access to the scientific literature that already exists and the negative unintended consequences this bill will have on the academic community, we urge you to reconsider whether S.2695 is needed.

Translation: Based on our evidence-free and counter-evidential speculations about risk to publishers, we urge you to renounce the demonstrated benefits to research, researchers, and the tax-paying public that funds them.

Thank you for considering our request.

Sincerely yours,

Robert R. Rich, MD, Senior Vice President and Dean, University of Alabama at Birmingham School of Medicine, Birmingham, AL
Richard P. Saller, Ph.D., Provost, University of Chicago, Chicago, IL
John R. Sladek, Jr., Ph.D., Vice Chancellor, Research, University of Colorado Health Sciences Center, Denver, CO
Kenneth L. Barker, Ph.D., Provost and Vice President for Research, SUNY Upstate Medical University, Syracuse, NY
Mary J.C. Hendrix, Ph.D., President & Scientific Director, Children's Memorial Research Center, Northwestern University Feinberg School of Medicine, Chicago, IL
Bruce A. Holm, Ph.D., Senior Vice Provost, SUNY at Buffalo and Executive Director, NYS Center of Excellence in Bioinformatics & Life Sciences, Buffalo, NY
Leonard R. Johnson, Ph.D., Vice Chancellor for Research, University of Tennessee Health Science Center, Memphis, TN
Barbara A. Horwitz, Ph.D., Vice Provost-Academic Personnel, University of California, Davis, CA
Richard J. Traystman, Ph.D., Associate Vice President for Research, Planning, and Development, Associate Dean for Basic Science Research, Oregon Health and Sciences University, Portland, OR
David E. Millhorn, Ph.D., Vice President, Office of Research and Economic Development, University of Tennessee System, Knoxville, TN

About the DC Principles Coalition for Free Access

The DC Principles for Free Access to Science Coalition represents more than 75 of the nation's leading nonprofit medical and scientific societies and publishers. The not-for-profit publishers are committed to working in partnership with scholarly communities such as libraries to ensure that these communities are sustained, science is advanced, research meets the highest standards, and patient care is enhanced with accurate and timely information.

As noted above, it would be absurd for the research community to keep sustaining its needless annual impact losses and just sit waiting passively for publisher voluntarism to decide whether and when to remedy them: The voluntarism of the DC Principles Coalition is far too little, too late.

Stevan Harnad
American Scientist Open Access Forum