Early career researchers are often reported to express the view that they face a dilemma. Submit to – and hopefully publish in – an open access journal, with possibly a relatively low impact factor, or in a traditional, pay-walled journal with a relatively high impact factor.
Given the large number of traditional pay-walled journals with low, or no, impact factors, I find this not the most credible argument. And even for ‘glam’ journals there are now good open access alternatives.
And yet, there are moments when I understand researchers when they are having to decide where to submit their papers. Do they choose an older subscription-supported journal, or a younger APC-supported open access journal? In the latter case, they’ll have to find the funds to pay the Article Processing Charge; in the former, they don’t, since subscriptions are paid out of the library budget. It does make a difference to a researcher's perception. Even though in many cases it is the funder who provides the money for the APCs, the researchers are aware of the cost and part of the decisions they take are financial/economic ones, even if sometimes subconsciously. They are not confronted with financial/economic decisions if they submit to a paywalled journal. Convenience may set in, perhaps in the form of a certain laziness, and a decision to stick with the old hassle-free subscription journals is easily taken.
It may happen here and there, but what I have not seen is attempts by the library community to confront researchers with the cost of paywalled journals. I'm not talking about the subscription price, but about the cost to the system of a single paper published in such a journal. It is a significant cost. For subscription journals published by the major publishers, this is on average in excess of $5000 (there are differences depending on the publisher), and for the ‘glam’ journals presumably more, much more (Phil Campbell, editor-in-chief of Nature, estimated costs of $30,000–40,000 per paper in 2013. That’s costs to the publisher; costs to the system will be higher, as they include profits.)
Now imagine that universities, perhaps via their libraries, take care of any payment to publishers, be they subscription charges or APCs, and then reclaim a per-article fee from their grants whenever researchers publish their articles. The amounts for APCs identical to the amount charged by the open access journal in question, of course; the amounts for articles in subscription journals on the basis of the average per-article revenue of the publisher of those journals. (These amounts may be reasonable estimates, I imagine, as they will seldom be known in detail.) The amounts thus reclaimed for articles in subscription journals could then be used for the journal acquisitions budget.
I have no illusion that this would solve all the problems of the cost of scientific publication, but it will increase general awareness of the true cost of publishing in subscription journals, and may help to level the playing field, to use an old cliché, between open access and pay-walled literature in the mind of scientists at the point when they decide where to publish their papers.