Re: Calculating the Cost per Article in the Current Subscription Model

[heatherm@eln.bc.ca]

On Tue,  4 Jan 2005 20:20:12 EST Phil Davis wrote:

> Firstly, this very conservative model predicts that most ARL
> institutions would pay more in a producer-pays model than in a
> subscription model.

Phil, with all due respect, I think the prediction is not supported by
available data.

I would agree that this is a fiscally conservative model - to me, a
"worst-case" scenario. The cost per article is exaggerated.  Even taking
the Wellcome Trust figures and multiplying by 1.6 as Sally suggests -
thanks Sally! - gives a range of $1,640 - $2,400 for author-charges
journals. This accounts for 35% overhead and 15% profit.  Whether 35%
overhead and 15% profit are reasonable expectations is another question,
of course.

Both figures are lower than the low end on your chart. The actual author
payments being charged by publishers are even lower than these estimates,
however.  (BMC $525, PLoS $1,500, Optics Express $450 under 6 pages, $800
over 8 pages). Estimates by some STM publishers of what they would need to
charge are also significantly less - e.g. Peter Banks of Diabetes /
Diabetes Care estimates about $1,700 for publication costs without print;
another STM publisher privately quoted to me about $750 as necessary.

The focus to date has been on STM publishing.  It seems quite possible
that social sciences and humanities would have lower costs still, because
these areas have never enjoyed the high revenues that journals in the STM
area have.

This model also does not account for other very likely funding sources,
such as funding agencies covering publication charges, and departmental
funds (used by some libraries to supplement budgets to purchase
subscriptions at present).

As for the ISI data, if they do not cover all the journals, is it possible
that they cover the more prolific STM journals - the ones with hundreds of
articles per yet?  In this case, a relatively small proportion of the
journals could account for a disproportionate percentage of the total
articles.

In sum, you have reached a conclusion (that ARL libraries would pay more
under a production-costs-based OA system), which, in my opinion, is not
supported by the available data.

This is not to say that anyone has proved that such a system is
necessarily less expensive, only that more information and analysis - and
possibly further developments - are needed before any conclusions can be
reached.

Unless, of course, we can agree that your data show that publication
charges of less than $500 would result in cost savings as well as
access/impact advantages for ALL the ARL libraries, and prioritize any
publishers that can produce quality journal articles at this rate for
experiments with shifting from funding subscriptions to funding OA?

best, Heather Morrison