redistributive implications of open access

This article addresses the virtues of gold open access (OA) from the perspective of its impact on social science scholarly associations and their members. OA has clear and obvious virtues, including redistribution downward and outward of research findings. But it also has the potential for upward redistribution or narrowing of the realm of publication, which this author finds troubling. A central question is who will cover article processing charges. The article identifies five potential sources of the necessary funds or ways to reduce the funds that are necessary, and discusses problems with each in terms of likely gainers and losers. It also identifies two potential substantive concerns about the kinds of social science scholarship most amenable to OA. It concludes by observing that, as is often the case, an apparently narrow technological innovation opens large issues – organizationally, substantively, and even morally.

L ike many other scholars, this author is eager to see widespread gold open access (OA) to her own and others' publications. The benefits seem clear in terms of the ability to have rapid and broad availability of each others' publications and to use them freely. Implementation of OA does, however, entail difficult decisions from the vantage point of professional associations that are simultaneously committed to fostering scholarship, providing services to members, giving special assistance to relatively vulnerable members or subject areas, and providing academic public goods. This is evidenced in the analysis of the issue by the largest political science association in the world, the American Political Science Association (APSA). APSA's Publications Planning Ad Hoc Committee (of which this author was Chair) was instructed in 2013 to 'look broadly at the needs, opportunities and issues for the future of scholarly publications and communications of the association, and to make recommendations for new initiatives or changes in approach'. The charge included examination of 'evolving technologies and formats for electronic media and open access'. 1 The committee of twelve was enthused about the potential of OA, but also noted that it 'raises broad and deep issues, leading one expert to claim that "a comprehensive OA paradigm is pretty much a frontal assault on professional societies"'. 2 Expressing concern that 'under current financial arrangements … open access can be cost prohibitive', the Ad Hoc Committee's report limited itself to the somewhat unhelpful call for 'a much more developed analysis of the financial, legal, and intellectual implications of open access … before concrete recommendations are made'. It urged the APSA to examine the 'opportunities and risks attendant on a regime of open access, especially in light of the other worthy demands on APSA resources'. In short, the whole issue was punted to a committee to be named later. 3 This article does not address the concrete financial, legal, and intellectual details of how a scholarly society in the social sciences should engage with innovations leading to gold OA. Instead, it approaches the issue at a slightly higher level of abstractionmostly assuming its benefits and looking more systematically at its costs. In particular, it considers some of the deeper, even moral, conundrums of redistribution and democratic control that might be associated with a move toward a broad programme of gold OA among social scientists.

UPWARD AND DOWNWARD REDISTRIBUTION
OA has evident benefits for writers, researchers, students, the interested lay person, policy makers, and perhaps others. OA journals can promote greater equality among both individuals and ideas; they may be especially valuable to younger scholars or those whose reputation is not (yet) so broad that readers will seek out their work in print or gated publications. When combined with electroniconly publishing and other changes in the ecology of publication, gold OA can contribute to both breaking the logjam created by highly selective print journals and reducing the fragmentation of too many subfield-specific journals. OA, that is, gives more scholars access to a wide audience and gives more readers access to the fruits of the resources enjoyed by scholars in wealthy universities and research centres. Thus it 'aligns directly with what is typically a core goal [for scholarly societies]the dissemination of knowledge '. 4 However, this commentator continues, OA 'potentially threatens their [scholarly societies'] financial viability and, therefore, their ability to fund other important activities, such as outreach, support for young and early career researchers, advocacy, and so on'. Just as the benefits of OA can be downwardly redistributive, so can its threats. After all, scholarly societies facilitate professional advancement especially for junior scholars, help scholars in teaching-oriented settings to remain professionally engaged, collect and disseminate data that can be used to promote diversifying a discipline, and provide arenas for academics who feel marginalized to connect with likeminded others in their discipline; if they must divert funds to OA, scholarly societies may be forced to perform fewer of these services.
'Just as the benefits of open access can be downwardly redistributive, so can its threats'.
Thus it seems unduly cavalier to say, as one European Community funder is reported to have held, 'If learned societies are a casualty of the move to OA, then so be it'. 5 The APSA is by no means a perfect institution, but it fosters the discipline of political science and the well-being of political scientists in w ays that few other institutions can. Therefore, if scholars want to keep professional associations solvent and to facilitate OA either to new or to traditional journals, they must consider alternative sources of publication funds beyond conventional paid subscriptions to print journals. The literature on OA suggests five strategies for covering costs: • Payment to publish a specific article by the author, the author's university, or an external funder such as a grant-making foundation. • Underwriting a journal, an issue, a scholar's research programme, or some other bundle of publications by a government agency, a foundation, or corporate or political paid advertisements. • Lowering publication expenses by reducing support for editorial direction, peer review, copy editing and page setting, electronic handling of manuscripts, and so on. • Having professional associations pay the costs that otherwise would have been covered by journal subscriptions. • Charging high fees to libraries and other organizations that serve as a conduit from a journal or publisher to scholars, students, and other readers.
All of these strategies are feasible, subjects of current discussions, or actually in place. They can be combined or subdivided. Leaving aside the (important) details of design and implementation permits one to focus on the central concern of this article: the implications of each strategy for redistribution of publication possibilities and content.
User fees (article processing charges, or APCs, in this arena) are common in many areas of life, perhaps increasingly so in the public arena as financially pressed local governments seek to raise revenue without raising taxes. User fees have the standard benefits of participation in a market; they are efficient, targeted, well-understood, transparent, flexible in response to changes of supply or demand. They are effective in calibrating and responding to the potential user's level of desire; if one wants to publish an accepted article badly enough, one will find the funds to pay for its publication.
But user fees also have the standard flaws of participation in a market: any more than a trivial charge is much more costly for the poor than the rich. Estimates of the charges for publishing in an OA journal range from under $200 to over $5,000, 'with the lowest prices charged by journals published in developing countries and the highest by journals with highimpact factors from major international publishers'. 6 These charges can be burdensome for a graduate student or junior faculty member, a person with heavy family commitments, scholars in resourcepoor colleges and universities, scholars in disciplines or disciplinary subfields with little or no external funding ('woe to the scholar of Chaucer or Prester John') 7 and individuals in poor countries that lack foundations or other underwriting organizations. It is easier for well-established scholars to find a sponsor for their new article than for those not yet widely recognized (a nastier version of this point is the prospect of 'vanity publishing' from wellheeled pedants matching up with 'predatory' OA journals that will publish anything for a sufficient fee). It is easier for scholars with research grants, which are much more common in empirical than in philosophical subdisciplines, to cover APCs. Perhaps the publisher or scholarly society can subsidize or waive fees for people or european political science: 15 2016 redistributive implications of open access institutions or countries with financial stressbut that adds a layer of complicated and potentially politicized bureaucracy, and simply exaggerates the costs for those just above the threshold of subsidy. User fees, in short, risk encouraging publication by the well-off or well-placed and discouraging it by the badly off, which surely violates all of the norms of meritocracy, equity, or innovation cherished by scholarly societies and by the OA movement itself.
A system by which government agencies, foundations, or sponsors of ads underwrite a publication bundle avoids the direct upwardly redistributive implications of author charges. Particularly if the foundation etc. supports a journal, or even a specific issue or substantive topic area, scholars without access to resources are not disadvantaged. But there are other potential hazards.
At least at the margin, a foundation or government granting agency might need to shift funds from direct research support to dissemination support, whether by eliminating some grants or requiring a successful applicant to divert research funds into publication costs. 8 That could be problematic; as the president of the Council of Scientific Society Presidents puts it, 'In the worst case, by diversion of research funds from research to publication, if less research is done … if fewer papers are published, if fewer students are trained, clearly competitiveness and innovation take a hit '. 9 It could be argued that the harms of diverting funds from research to publication are not so clear; it would be interesting to determine if the last grant given or the last tranche of money spent on a particular project adds more value to scholarship than what would be gained from a subsidized gold OA journal. But to this author's knowledge no one has made that determination, and the difficulties of doing so convincingly are obvious.
Perhaps a greater danger is that foundations, governments, or ad sponsors could have undue influence on scholars' research agenda or substantive findings, or on publishers' and journal editors' choices of what to publish. One need not imagine corruption or manipulation. It is simply the case that over time people tend to hunt where the ducks are, and if scholars, editors, and publishers know that some research will have publication subsidy and other research will not, a tendency to drift in the former direction seems likely. And inevitably, decisions at the margin will be contentious: 'in the middle of the year 2012-2013 it seemed that reduced money for APCs meant that university managers might have to be the people who would determine what articles got published, a very threatening move to academics, but an equally terrifying one to universities' (Wickham and Vincent, 2013: 8).
Finally, a foundation, government agency, or ad sponsor might change its priorities or focus after some period of time. That could be desirable; there is no reason to assume that a journal or research agenda that was highly valuable X years ago will continue to be equally valuable years or decades later. But it will be unsettling, and perhaps unjust, for scholars, publishers, or journal editors to learn that the dissemination support they had long counted on is being withdrawn or diverted to a new agenda. In short, if markets have one set of flaws, institutional subsidies have a different and perhaps worse set. 10 These anxieties would be lessened if the costs of publishing an article in an OA journal were lowered, say, from several thousand to several hundred dollars. As some OA journals have shown, it is not hard to reduce expenses. The front-end costs of handling manuscripts could be cut by, for example, reducing the number of peer reviews or expending less effort to find the most appropriate reviewers, reducing checks to confirm the submission's originality, desk-rejecting many more manuscripts without an editor spending time to explain why, or using off-the-shelf rather than tailored database systems or software for managing the flow of manuscripts. The back-end costs of handling accepted manuscripts could be cut by, for example, reducing the level of copy-editing or fact-checking, lowering the quality of artwork or requiring authors to cover its costs, or eliminating proof reading at the journal. Editors could receive a smaller subvention and fewer assistants or staff; face-to-face meetings could become tele-conferences. Or journals could change their review criteria, perhaps to match PLoS ONE's focus on only methodological rigour rather than substantive importance, and thereby raise acceptance rates dramaticallythus increasing the number of scholars available for author charges. There are probably other possibilities too.
Some of these cost-cutting measures could spark useful innovations. Perhaps authors (or their institutions) can efficiently be held responsible for copy-editing and proofing; a department, university, or scholarly society can hire a professional editor so that the expense of improving an article would be widely distributed among authors or even among all faculty or APSA members. More interestingly, what is gained by substituting online post-publication peer review for pre-publication anonymous peer reviews 11 seen only by the editors and author? Perhaps post-publication exchanges and citation counts do more to adjudicate the quality and value of an article than our traditional criterion of journal impact factors (Curry, 2013: 62); perhaps comments are written more carefully than some manuscript reviews are. A (moderated) comments section in an electronic OA journal can attract readers and observations from around the world, thereby becoming itself a lively and valuable contribution to OA scholarship.
Cutting publication costs also, however, entails losses. The best (or at least, busiest) scholars are likely to be unwilling to serve as journal editors if their subvention, released time, and staff and student assistance are cut. Peer reviews of submitted manuscripts improve their quality much if not all of the time; post-publication comments can be erratic, politicized, or merely unhelpful. Loss of editing, copyediting and proofing, as well as increased reliance on off-the-shelf electronic editorial management tools, disadvantages those for whom the language in which the journal is published is not a native language. Increasing the number of desk rejects to lower transaction costs risks losing articles that are innovative, idiosyncratic, ideologically or epistemologically different from the tastes of the editor or the journal's traditionor may simply be bad guesses. If traditional biases continue to obtain, scholars who are young, female, non-white, or at teaching colleges or other non-university settings are especially likely to be harmed by high levels of desk rejects and the loss of blind peer review. In short, cutting publication costs in order to reduce the authors' costs of publishing in a journal without a conventional subscription base may harm the quality of articles and narrow the range of likely acceptances.
Perhaps scholarly societies, rather than individuals or organizations, will bear the costs of shifting from subscription-based journals that provide revenue to the society toward gold OA journals that require payment from the author or a proxy. Since the APSA runs on a tight '… what is gained by substituting online postpublication peer review for pre-publication anonymous peer reviews seen only by the editors and author?' european political science: 15 2016 redistributive implications of open access budget, this new expense might be covered by cutting costs elsewhere. Again, confronting a new expenditure could spur innovationmore tele-conferences rather than face-to-face meetings, making in-house newsletters entirely electronic, seeking a new pool of members among, for example, political scientists not employed in colleges and universities, and so on. Directors of some learned societies, including the APSA, are even developing business models in which a new gold OA journal might increase rather than decrease association membership, on the grounds that the new journal would disseminate the valuable products of the discipline more widely in the public arena and lure new (especially younger) members into the association.
But until that new model proves itself, most scholarly associations will face difficult decisions. After all, 'learned societies are disproportionately … dependent on journal subscriptions; and their very considerable contribution to the academic ecosystem in the form of scholarships, travel grants and the like is thus itself dependent on people and institutions continuing to buy journals, or at least pay (if Gold open access continues to be relevant) for the articles contained in them' (Wickham and Vincent, 2013: 9). What might the APSA do, if gold or even green OA leads to a loss of dues-paying members? Should it reduce support for graduate students to attend the annual conference, reduce the size of the governing Council, or eliminate some member-based committees so as to cut travel and meeting costs? Should a professional association jettison smaller conferences that serve a minority of members and do not quite meet their budget, such as a conference on teaching and learning? Should it cut back on discipline-wide data collection and dissemination, or eliminate one staff position? Raise dues and accept the risks of losing some members and disproportionately disadvantaging students and assistant professors? All are possible, none are desirableand how to weigh the intellectual downward redistribution of OA against the likely upward redistribution attendant on reducing services provided by scholarly societies is an open question.
A final strategy for meeting the expenses of a non-subscription OA journal is for academic publishers or scholarly societies to increase the prices to libraries and other organizations for buying monographs, print journals, or bundled licenses for electronic rights. (As another blogger put it, 'do people who use subsidized child care at conferences know that libraries are paying for it?' through higher subscription payments for scholarly society journals than are needed to cover publication costs.) Prices for journals and books from commercial publishers have risen faster than library budgets, especially in some disciplines, and there is no reason to think that the trend will stop. Library budgets at even wealthy universities such as Harvard and Yale are pressed, and smaller or less wellendowed colleges and universities are cutting collections deeply.
Again, innovations are possible and welcome in the face of financial pressures; many, such as purchasing consortia, efficient interlibrary loan systems, or green OA are underway. But libraries are to some degree a captive audience since their faculty and students insist on almost immediate access to the version of record for publications in their field, and publishers know it. What categories of individuals and institutions are benefited and harmed by libraries' subsidies of the costs of journal publication ('people who use 'What might the APSA do, if gold or even green open access leads to a loss of dues-paying members?' subsidized day care at conferences')and whether the downward redistribution of OA is worth the upward redistribution of library collectionsis a question worth careful study. In short, OA journals or OA to extant print journals have the effect of redistributing the dissemination and the receipt of information downward and outward. In this author's view that is just what individuals, universities, scholarly societies, libraries, and publishers should be in the business of doing. But the expenses entailed in providing such access run the risk of redistributing upward, by making it easier for wealthier or more powerful individuals and institutions to avoid payment for what they will gain through OA. How to balance the upward and downward redistributionsor how to minimize the former and maximize the latterremains unclear.

SHAPING WHAT INFORMATION IS OPENLY ACCESSIBLE
Like many such discussions, my discussion of gold OA focuses on articles of the type that normally appear in scholarly journals. Since it will be a long time, if ever, before books are also freely accessible to all, OA journals raise another issue of redistributionsubstantive this time, not financial. If people around the world can freely and easily read articles on their computer but must buy or borrow printed books, readership is likely to shift decisively toward the style of research best encompassed by an article. That has implications for what information is presented, how much is presented, and how it is presented. In some social science disciplines (e.g., economics and psychology) there will be little disruption, since the coin of the realm has long been articles or even unpublished working papers. In other social sciences such as sociology, cultural studies, and political science, however, a growing dominance of articles over books could substantially change the nature of the discipline (Darley et al, 2014: 27). In political science, for example, shifting attention to articles would advantage some subfields or topics, and disadvantage others, relatively or even absolutely. Just as gold OA has progressed further and along different financial and intellectual lines in the natural sciences than in the social sciences, so it has found more enthusiasts in some than in other subfields of political science. The more 'science-y' end of the disciplinescholars who study politics through precisely defined methodologies such as aggregate data analysis or formal models, and who mostly write short, highly focused empirical articlestend to see more benefits in OA than do scholars on the more humanities-oriented end of political science, such as political philosophers, political historians, or those using qualitative methods such as ethnography or intensive interviews. 12 Whether the benefits of greater availability of articles outweighs the costs of losing the breadth of scope, epistemological frameworks, and type of information available most often in a book seems worth considering, even if it is impossible to decide. 13 A final way in which the technology of OA could redistribute the content of scholarship has to do with the hoary question of a possible trade-off between quality and quantity. The standard assumption is that OA is or will be closely linked to electronic publishing, so that the old constraints of hard copy publication will vanish. But by a different logic, OA, non-subscription based journals could actually introduce new constraints on the number and type of articles published. That is, if foundations, government agencies, departments or universities, scholarly societies, or even individuals have to shift funds from direct research support to support for publication and dissemination, at european political science: 15 2016 redistributive implications of open access the margin fewer research projects will receive funding, at least in the short run, or individuals will think twice before incurring the costs of submitting another manuscript to a journal. Whether commercial or nonprofit publishers are as eager to begin and maintain journals once their revenue streams from subscriptions decline, or whether new publishers will emerge, are also open questions.
Fewer and more carefully chosen articles might not be a bad thing, given that the modal political science article receives barely one citation. 14 A lot of words are published that have no impact except possibly on the career of the writer. But the idea that fewer and better articles might be preferable has more than a hint of elitism, which is antithetical to the moral and intellectual impulse behind OA. And there is no guarantee that 'fewer' will be associated with 'better'. Therefore the more likely and perhaps more desirable outcome is an expansion in the number of published articles, independent of quality which brings us back to the question of 'who pays?' with which this argument began.
In sum, OA is an apparently technical, financial, and organizational issue that in fact raises questions of redistribution and democratization in disciplines and scholarly societies. Neither onlookers, scholars nor experts can predict with any certainty what will happen in this arena. In a personal communication about this article, a friend wrote, 'I asked XX what her publishing company would be like in twenty-five years. She basically said it was hard to know what it would be like in five yearsand that's a publisher of hundreds of academic and educational journals'. The central point here is that evaluating how OA should be shaped may be even more difficult to determine than what it will be like.
'Fewer and more carefully chosen articles might not be a bad thing, given that the modal political science article receives just over one citation'.