Re: cost of peer review and electronic distribution of scholarly journals

[Stevan Harnad <harnad@ecs.soton.ac.uk>]

         ** Apologies for Cross-Posting **

On Thu, 22 May 2008, N. Miradon wrote:

> The current issue of Nature has correspondence from Dr 
> Raghavendra Gadagkar. The abstract of his letter (available at 
> [1]) compares and contrasts 'publish for free and pay to read' 
> with 'pay to publish and read for free'. To read the letter in 
> full will cost you USD 18.
>
> N Miradon
>
> [1] http://www.nature.com/nature/journal/v453/n7194/full/453450c.html
> Nature 453, 450 (22 May 2008) | doi:10.1038/453450c; Published online 21 May
> 2008
Here is the part you can read for free:

Open-access more harm than good in developing world
         Raghavendra Gadagkar
         Centre for Ecological Sciences,
         Indian Institute of Science,
         Bangalore 560012, India

"The traditional 'publish for free and pay to read' business 
model adopted by publishers of academic journals can lead to 
disparity in access to scholarly literature, exacerbated by 
rising journal costs and shrinking library budgets. However, 
although the 'pay to publish and read for free' business model of 
open-access publishing has helped to create a level playing field 
for readers, it does more harm than good in the developing 
world..."

It is easy to guess what else the letter says: That at the prices 
currently charged by those Gold OA publishers that charge for 
Gold OA publishing today, it is unaffordable to most researchers 
as well as to their institutions and funders in India and 
elsewhere in the Developing World.

This is a valid concern, even in view of the usual reply (which 
is that many Gold OA journals do not charge a fee, and exceptions 
are made by those that do charge a fee, for those who cannot 
afford to pay it). The concern is that current Gold OA fees would 
not scale equitably if they became universal.

However, the overall concern is misplaced. The implication is 
that whereas the user-access-denial arising from the the 
unaffordability of subscription fees (user-institution pays) is 
bad, the author-publication-denial arising from the 
unaffordability of Gold OA publishing fees (author-institution 
pays) would be worse.

But this leaves out Green OA self-archiving, and the Green OA 
self-archiving mandates that are now growing worldwide.

Not only does Green OA cost next to nothing to provide, but once 
it becomes universal, if it ever does go on to generate universal 
subscription cancellations too -- making the subscription model 
of publishing cost recovery unsustainable -- universal Green OA 
will also by the very same token generate the release of the 
annual user-institution cancellation fees to pay the costs of 
publishing on the Gold OA (author-institution pays) cost-recovery 
model. 
http://www.publications.parliament.uk/pa/cm200304/cmselect/cmsctech/399/399w 
e152.htm

The natural question to ask next is whether user-institution 
costs and author-institution costs will balance out, or will 
those institutions that used more research than they provided 
benefit and those institutions that provided more research than 
they used lose out?

This would be a reasonable question to ask (and has been asked 
before) 
http://www.google.com/search?num=100&hl=en&q=+site%3Alistserver.sigmaxi.org+ 
amsci+%22net+provider%22&btnG=Search -- except that it is a 
fundamental mistake to assume that the *costs* of publishing 
would remain the same under the conditions of universal Green OA.

It is far more realistic to expect that if and when journals 
(both their print editions and their online PDF editions) are no 
longer in demand -- because users are all instead using the 
authors' OA postprints, self-archived in their IRs -- that 
journals will convert to Gold OA not under the current terms of 
Gold OA (where journals still provide most of the products and 
services of conventional journal publishing, apart from the print 
edition), but under substantially scaled-down terms.

     Harnad, S. (2007) The Green Road to Open Access: A Leveraged
     Transition. In: The Culture of Periodicals from the Perspective of
     the Electronic Age. L'Harmattan, pp. 99-105.
     http://eprints.ecs.soton.ac.uk/15753/

In particular, all the current costs of providing both the print 
edition and the PDF edition, as well as all current costs of 
access-provision and archiving will vanish (for the publisher), 
because they have been off-loaded onto the the distributed 
network of Green OA IRs, each hosting their own peer-reviewed, 
published postprints. The only service the peer-reviewed journal 
publisher will need to provide is peer review itself.

That is why Richard Poynder's recent query (about the true cost 
of peer review alone) is a relevant one.

As I have said many times before, based on my own experience of 
editing a peer-reviewed journal for a quarter century, as well as 
the estimates that can be made from the costs of Gold OA journals 
*that provide only peer review and nothing else today*, the costs 
per paper of peer review alone will be so much lower than the 
costs per paper of conventional journal publishing today, or even 
the costs per paper of most Gold OA publishing today, that the 
problem of the possibility of imbalance between net 
user-institution costs and net author-institution costs will 
vanish, just as the the subscription model vanished.

Alma Swan has forwarded the link to a JISC-funded study of such 
questions being conducted by John Houghton (Australia) and 
Charles Oppenheim (UK) (in the context of UK research, where 
there are, I assure you, author-institutions that are every bit 
as worried about current Gold OA publishing fees as Developing 
World institutions are) and RIN has released a study:

http://www.jisc.ac.uk/whatwedo/programmes/programme_rep_pres/economicsscholarlypublishing.aspx

Alma also forwarded this study, by RIN:

http://www.rin.ac.uk/costs-funding-flows

Peter Suber has pointed to Fytton Rowland's 2002 estimates of the 
cost of peer review alone:

     Rowland, F. The Peer Review Process. Learned Publishing, 15(4) 247-58.
     http://www.jisc.ac.uk/uploaded_documents/rowland.pdf

Peter writes:

     "Rowland does a literature survey to determine the costs of peer
     review (see Section 5).  He concludes (Section 7) that it's about
     $200 per submitted paper, or $400 per published paper at a journal
     with a rejection rate of 50%.

     "I'm not in a position to vouch for the results, but it's the only
     paper I've seen trying to answer this narrow question.

     "Note that the paper came out in 2002 and doesn't reflect the latest
     generation of journal management software.  This matters because
     steadily improving software (including open-source software) is
     steadily taking over the clerical chores of facilitating peer review,
     and thereby reducing its costs."

I would add that even at $400 per paper, that would make peer 
review alone cost only 10% of the average price of $4000 that 
Andrew Odlyzko estimated was being paid per article in 1997 
(i.e., the total collective contribution summed across 
subscribing institutions) and less than a third of most Gold OA 
publishing fees per article today.

     Odlyzko, A. (1997) The economics of Electronic Journals.
     First Monday 2(8)
     http://www.firstmonday.org/Issues/issue2_8/odlyzko/index.html

Stevan Harnad