In his comment to my note on Marty Frank's NEJM editorial, Marcus points to what he sees as the major flaw in Marty's argument -- that publication shouldn't be seen as something separate from the overall research process. I think he's on the right track, although I'd give it a slightly different focus.
Marty questions the wisdom of diverting NIH funding from research grants to support the author-pays model of open access funding, and suggests the amount could be around $200 million. Marcus responds that this is a false separation and diverting a "relatively small amount" of funding for publication should be seen as part of the same overall research process rather than a separate project altogether.
I tend to agree with Marcus in the abstract. I find an appealing symmetry in the notion that the same sources that pay for the test tubes and reagents and animals and research assistants should pay for the dissemination of the results. And on a per grant basis, the additional costs are indeed relatively small.
But then I think about the practicality of discussing these issues with the key decision makers at my institution -- one that is very focused on NIH-funded research and extremely concerned about the flattening of the NIH budget. Even if Marty is only half right and it would only take $100 million to fund publication, I can see the faces of my VP for Research or Dean of Medicine getting more pinched. They're thinking of the young investigators who are submitting their first renewal, or even their first grant as PI, and worrying about what's going to happen in the face of increasing competition for a shrinking pool of funds. Why should those promising careers be jeopardized in pursuit of an open access goal that they're not convinced is warranted? Marcus is right about the need to make hard choices.
Many librarians, on the other hand, have become very dismissive of the Biomed Central membership model. They don't like the idea that the funding should come out of their budgets. "It's nothing more than another subscription!" Since few librarians are competing for NIH grants, the notion of taking a percent or two out of that pool to fund publication is eminently sensible. But taking it out of library budgets is just wrong.
Peter Suber, in his comment on Marty's editorial, dismisses the Cornell study as "discredited". I think that's a little harsh. Whether one agrees with the specific numbers in the Cornell model or not, it seems pretty clear to me that an institution like mine, given the number of articles we publish each year, would pay more in publication costs, if it had to fund them, than it currently pays in subscriptions (or license fees) through the libraries. The money has to come from somewhere.
I tend to think that Marty exaggerates the negative impact of using NIH funds to cover publication costs. But I also think that it's quite reasonable (again, in the abstract) for an institution to divert funds from library acquisitions budgets to cover those costs. In fact, from an institutional standpoint, it probably makes more sense to do that. After all, if an institution has been spending a million or three from the library to fund publication via subscriptions, why doesn't it make sense to use that same money to fund publication directly via open access?
Librarians who think that open access is intended to reduce the pressure on their budgets won't like that, of course.