--- title: The Problems of Unit Costs Per Article layout: post image: feature: oa.png --- Every five minutes or so, someone tries to come up with a cost-per-article figure for academic publishing. In the past, I've tried to do it too. But more and more I find myself wanting to resist the temptation. Not only because the data collection takes forever, but because the figures that I would produce, from my organisation, would likely not be cross-applicable to another organisation. Yet, Plan S and other organisations keep floating the idea of some kind of transparency on a per-article basis. Here are some of the problems with the approach that I have been trying to think through. This is very much a 'thinking aloud' approach here and all of these aspects are up for debate. 1. It encourages the wrong mentality around support in the consortial model that we run. It encourages institutions to think of their contributions to OLH as though they were payments for their own authors, rather than a re-pooling mechanism so that everyone has access to scholarship. This in turn leads to metric demands for "how many of our authors have published in OLH?", which becomes very difficult to accurately answer when we sponsor material across different platforms. 2. APC caps favour larger, extant organisations with capital reserves. If you set a cap of $600 per article (which has been suggested) and you want your publisher to have someone working on it full-time as an employee on a salary of $30,000 (which will cost you, as an employer, at least in the UK, closer to $40,000 with on-costs), with typesetting costs of £150 per article, adding on memberships to Crossref, CLOCKSS, COPE, COUNTER, and having a pro platform in place, then you need to publish about 100 articles per year. If you do any kind of in-house copyediting, or even if you don't, that's a lot of articles to get out the door, while running an organisation that also has to a.) collect revenue; b.) perform organisational overhead activities (in our case, legal charitable filings and so forth); c.) market the organisation so that you can actually get some revenue/submissions; d.) have a surplus so that you don't go bankrupt if there are revenue fluctuations. I would also add that it's pretty tough to go from zero to 100 articles in year one if you have any kind of quality control. Most publisher estimates are that it takes eight years to establish whether a journal is viable. This cap level seems to me unrealistic because it assumes a flexibility of salary levels that just doesn't exist (i.e. that it's OK for people to only take a portion of a salary while they build up to the next viable point at which they could reasonably hire someone). It also favours large extant organisations because they will already have divisions of labour and an article/submission volume that could support these staff, hence allowing these organisations to differentiate salary levels (even as they may also accrue dis-economies of scale through coordination overhead). 3. The obsession with unit costs forces us into anti-distributional economic patterns. I've written quite a lot about this before, but APCs concentrate costs on points of the system where funding may not be available. For instance, the average book processing charge for academic monographs (back of the envelope figure here but loosely [based on Frances Pinter's landscape study](https://doi.org/10.3998/3336451.0021.101)) is around £10,000. This is more than the entire book purchasing budget for a year of my department. But if two hundred universities worldwide contributed £50 per book -- as per the Knowledge Unlatched model etc. -- then suddenly it is affordable once more. If you think that you are buying a service or an object, we have a problem in flipping to open-access economic models in the world outside of project funding. 4. Of course, lots of for-profit academic publishers, to be blunt, have taken the piss with exorbitant profit margins for decades in the 30% region. Some not-for-profit actors have brought in obscene surpluses to the world's wealthiest institutions, usually through textbook sales. Some Learned Society publishers use library budgets as their own personal intra-disciplinary sources of funding. In many ways, it is these entities' faults that we have ended up in a situation of near-total trust breakdown between many academic libraries and academic publishers. And why should there be any trust there? These are some of the reasons that I believe that we need trustworthy infrastructural academic publishers, funded by academic libraries, at affordable rates, but that are not constantly evaluated on a strict unit-cost basis. (Getting the metrics for evaluation is, itself, a _job_ that is a total pain.) Certainly, it will become obvious very quickly if those entities are extremely inefficient and the unit cost is high, because we should require some degree of transparency from them. A really good example of this is [Open Book Publishers](http://blogs.openbookpublishers.com/introducing-data-to-the-open-access-debate-obps-business-model-part-three/). They say how much money they got. They say how they spent it. And they come out with a unit cost figure if you want it. But note that other publishers will have different costs. There will be inherent variability here according to organisation size/the established nature of the publisher and so forth. And this presents another problem with APC caps. If you set the level to an upper bound (let's say $1,000) then everyone will charge that, because nobody wants their organisation to go bust. So there is an incentive, from the payer perspective, to drive down APCs. But this, again, does not necessarily allow support for the types of trustworthy, honest, mission-driven organisations that we might want. The distributed 'market' of academic publishing and [the cries of academic freedom](https://eve.gd/2018/10/23/jottings-on-academic-freedom-and-plan-s/) also create problems for this type of infrastructural, as opposed to goods or service -driven approach, even as the academic publication environment behaves [about as dysfunctionally as might be possible](http://books.eve.gd/oahums), fuelled by the symbolic economy of journal titles and Impact Factors tied to academic hiring procedures. The whole thing is a painful mess. Plan S's idea of funder-infrastructure-publishers could be a great fix for this. But they seem limited in scope to those funded by that funder, and they are reluctant to dictate that academics _must_ publish in such venues anyway. I assume they might, also, levy APCs on those outside of the funding ecology, too, thereby not solving the whole dilemma. These are a few of my not entirely assembled thoughts. But I find that putting this stuff out on Twitter and then muting incoherent respondents on paper helps me to formulate better my own ideas around such topics, and the delicate balances between transparency/accountability and a system founded on trust, with trustworthy actors, that could allow for greater imaginative potential futures for our scholarly communication infrastructures.