I’m going to disregard the usual speculation about what type-setter and editorial assistant salaries are, and how much distribution infrastructure costs because these are all tied in to the true costs of publishing from a publisher’s perspective and not what I’m interested in. Instead, I’m going to use figures from my employer, the University of St Andrews, to crudely examine what this very small market can bear open access articles to cost.
The first assumption I make here is that journal subscriptions and gold open access journal publication costs should be drawn from the same pool of money. That is, they are university outgoings that support publishers, thereby funding the publication of university-based researchers’ work.
The second assumption, which almost immediately serves to highlight how useless this back-of-the-envelope calculation is, is that we no longer need to subscribe to paywalled journals and can therefore channel all funds that we would have spent on this into open access publishing. For argument’s sake, let’s suppose that the UK government has negotiated a nationwide subscription to all journals with all closed-access publishers for the 2014/2015 academic year. This leaves the University of St Andrews Library with journal subscription money that it needs to spend in order to continue its current funding allocation. Naturally, it ploughs all of this into open access publishing costs.
Once comfortable with these assumptions, we can fairly easily estimate how much a university like mine could afford to pay for each article published, if every single output was a gold open access article such.
Total St Andrews University spending on journal subscriptions per year:
According to the library’s 2011/2012 annual report: £2.11m
According to a tweet from the @StAndrewsUniLib twitter account: ~£1.7m
Given that the higher value also included spending on databases and e-resources, I’ll go with the £1.7m/year estimate.
Total number of publications by St Andrews University researchers per year:
We have a PURE research information system on which all researchers are meant to report all of our publications. Accordng to a tweet from @JackieProven at the University of St Andrews Library:
over 2000 publications/yr, about 1200 are articles and around half of those will have StA corresponding author
We can therefore assume 600 publications/year.
Open access publication costs which could be absorbed in this hypothetical situation:
£1,700,000/600 = £2,833
This value is higher than I was expecting it to be, and suggests that for even a small institution like the University of St Andrews, article processing charges (APC) in gold open access journals aren’t too far off the mark. According to PeerJ’s roundup, even PLOS Biology’s steep APC of $2900 is considerably less than what St Andrews could bear in this highly unrealistic situation.
Of course, there are quite a few caveats that sit on top of this hypothetical estimate and its assumptions:
1) I may well be underestimating the number of publication outputs from the University’s researchers. This would push the per-article cost the library could afford to pay down.
2) Larger universities would have a greater number of researchers and therefore publications. The increase in the denominator would be offset by an increase in the numerator-larger universities have medical schools and law schools which St Andrews does not-but I have no idea what effect this would have on the per-article cost these better endowed libraries could afford to pay.
3) The ecosystem would change. Gold open access journals have higher publication rates than paywalled journals. If more articles were published, this would also push the per-article cost the library could absorb down.
4) This estimate makes no consideration of the open access publication option in closed access journals. This publication option, as well as being more expensive than the gold open access offered in open access only journals allows traditional publishers to milk the cow at both ends (subscription costs AND APCs) and I imagine library administrators would struggle to justify supporting this from the same fund as that used to pay journal subscriptions.
I’ve been meaning to do this calculation for a few months and am grateful to the staff at the University of St Andrews Library for providing me with these figures. I’m interested in what others make of this, and would be keen to hear your thoughts in the comments below.