Bernd Pulverer 
Bernd Pulverer
Chief Editor, The EMBO Journal 

We all want Open Access – authors, readers, funders and indeed publishers alike. EMBO is keenly interested in the OA question both from a policy viewpoint and as a funder and publisher of high-level biomedical research. EMBO Press publishes four journals, two of which are fully OA and one of which, Molecular Systems Biology, was one of the first OA journals to be founded. MSB is currently the highest Impact Factor OA journal (for whatever that metric is worth). The other two journals offer OA as an option at publication and as a default after one year.

The advantages of OA have been discussed at length and center on equitable access to research supported by public funds, with the immediate benefit for journals being the resulting larger readership and increased web access. As article level metrics become more relevant, these results alone pique the interest of publishers. On the other hand, it is less than clear whether this translates into greater citations to articles, the hard currency of research assessment. For example, a randomized study found no significant citation effect despite a clear boost in web access to OA articles (BMJ 2008;337:a568), which is in line with our own data based on 91 ‘EMBO Open’ papers compared with 510 non-OA papers published in The EMBO Journal between 2010 and 2012. An explanation for this discrepancy is that the ‘professional’ users who are in a position to cite work are likely to have access to a paper irrespective of its OA status. What is more surprising, though, is that at least within the constraints of the small sample we considered at The EMBO Journal, that OA also had no significant effect on social media activity around a paper and, notably, on access from less well-heeled areas such as South America or Africa (it is worth pointing out that the papers in question are mostly of interest to those engaged in basic research).

Be that as it may, almost everyone in principle favors OA - but at what cost? One fundamental characteristic of Article Publication Charge (APC)-based models is that a journal’s income is directly proportional to the number of papers it publishes. This is not a big issue for journals with limited editorial and production overheads and high acceptance rates, but it is a serious issue for highly selective journals (the four EMBO Press publications all have acceptance rates around 10%). The editorial overhead to administer the efficient, fair and informed editorial process cultivated at EMBO Press is very high, yet this cost has to be shouldered by only one tenth of the submitting authors – i.e. those we publish. The result is that OA at EMBO has to cost much more than the 2,000 US$ limit currently considered reasonable by most researchers, institutions and funders.

How can we close the gap on this cost differential? Cutting corners is not an option for us. In fact, we are heavily investing in the editorial process to allow us to publish papers fit for an online world – papers that are proper research tools, not merely online copies of printed papers. Furthermore, we are investing in systematic pre-publication ethics screens to fulfill what we see as an obligation for journals to resolve or flag up issues at the last checkpoint before publication.

Some publishers are taking a leaf out of the Public Library of Science’s (PLOS) playbook, the publisher of PLoSONE, a high revenue journal, despite being OA, to launch ‘low threshold’ journals that generate sufficient profit to support selective journals. Like others, I feel sure, we have discussed similar journal ecosystems at EMBO. Indeed, we are keen to publish all papers that meet our high standards for experimental quality and scholarly discourse in one or another of the EMBO Press journals without the need for de novo re-review. What we want to avoid at all costs, however, is to ‘open the gates’ to half-baked or low quality science: over one million papers are added to PubMed every year, published in well over 15,000 Journals, and we have no intention of adding unnecessarily to this mountain. Information overload is one thing, but more importantly, bad science has a toxic effect – it misleads other researchers who rely on published information to direct their investment of time and funds.

Our current approach, which we hope will allow us to remain selective while publishing all valuable research, is to establish highly efficient manuscript transfer arrangements between the four EMBO Press journals and with outside partners including eLife and the BMC journals. As a second route to providing income to highly selective journals, we would propose that in the event of a successful transfer, the transferring journal would receive an ‘APC rebate’ from the receiving journal as a legitimate means to account for editorial overheads.

A third way would be to levy submission charges. This may ultimately be the only fair means of equitably distributing editorial costs between all the authors interested in publishing in a given journal (in our model, the submission costs for successful papers would be counted towards the APC upon publication). Such charges remain rather unpopular, so journals shy away from them for fear that they may dampen submissions. The jury is out on the question of whether such charges would really discourage appropriate, high-quality submissions. In my view, however, we will require much more efficient online payment systems before we can contemplate this move.

In terms of the additional costs arising from enhanced production processes, such as the typesetting of supplementary information figures (which we call ‘Expanded View’) and Source Data files, we may ultimately replace ‘APCs’ with ‘Article Rendering Charges’ (ARC’s), which would allow authors to opt in to these additional services as they wish.

At the end of the day, the real cost of publishing a paper will not change, whether or not it’s published OA. The main bottleneck to the wholesale conversion to OA is that the funds for publishing are currently sequestered in library budgets, which are still earmarked for subscription fulfillment, while APCs are shouldered by authors. Paying APCs can involve time consuming invoicing, and authors sometimes have to dip into funding that would otherwise be used for more interesting research. When the APC bill comes in, researchers suddenly face a rather significant new cost, and it feels to them like yet another charge they have to shoulder just to share their own work. To add insult to injury, this charge comes from publishers with names heretofore associated with inflated profit margins.

The dichotomy of financial channels (library vs. researcher) has led to a real increase in cost to research organizations (though this increase should of course be transient). As might have been predicted, libraries are increasingly tasked with paying APCs on behalf of authors. This is fine, but we should bear in mind that it bundles payment power, leading to strong incentives for discounts that will then not reflect the actual cost of publishing a paper (OA at high quality journals barely breaks even at present). Everyone is under budget pressures, but it would be  unfortunate if powerful library consortia drove APC discounts, as independent OA journals would have a hard time surviving without compromising on quality. Furthermore, making authors who do not have access to such consortia pay the financial shortfall would be unfair, especially as such authors will often hail from less well-to-do research environments.

EMBO Press and many others are keen and poised to go fully OA, but we need clear signals from funders and governments that they are committed to making this method of publication financially realistic and sustainable. The costs will need to be covered, and at high-level journals they will be higher than a couple of thousand dollars, as has always been the case.