The Goods in the Scholarly Marketplace

This a set of notes for my talk at Duke University this week. It draws on the Political Economy of Publishing series as well as other work I’ve been involved with by Jason Potts at RMIT amongst others. The title of the talk is “Sustainable Futures for Research Communication” and you can find the abstract at the Duke event page.

The video is now available along with the slides. The lecture capture didn’t get such a clear view of the slides so you may want to bring both up and play along.

Sustainability is the big discussion point in Research Communication. Will journal’s survive? Are APCs the only credible route to sustainability? Or will they lead to inevitable destruction of journals? Will monographs survive? Scholarly societies? What, as is asked over and over again, is the future of the library? Of the institution? This week, frankly, of the nation state?

Stuff costs money. If it's free, it will either fold in the future or it will have to start charging.

— Stefano Tonzani (@tonzani) September 29, 2016

Yes: cost breakdown is fascinating & refutes myth that publishing is free. But APCs are not a sustainable model IMO. https://t.co/uBrDcvAeQM

— Elizabeth Yates (@LibraryWriteHer) September 29, 2016

Figure: Two responses to the same tweet. Differing views on what “sustainability” is all about.

And into this space we see perspectives from economics and political economics start to filter in. Martin Eve and David Golumbia are concerned about taking labour – and implicitly – labour theory seriously. From within the publishing industry the perspective of consultants – mostly financial analysts rather than economists – looks to trends and balance sheet calculations. Scientific Reports up, PLOS ONE down, build a story out of the data. As someone who has seen the inside those stories are almost always roughly right on the trends and almost entirely wrong on the underlying reasons for them.

Financial analysis is important but it is a weak grounding to understand Research Communications. Here I agree with Golumbia and Eve, as well as Jason Potts amongst others, that “OA advocacy” but also industry advocacy “[…]lacks robust critical grounding for its propositions in credible Marxist or socialist economic theory” or indeed in any but the most facile theories of market operation. There are sound political reasons for this. Articulating a clear dichotomy between corporatist, self-enriching encumbent players, and the selfless and public spirited intent of scholars to communicate has been an affective mean of driving the debate.

The nature of “scholarly goods”

This dichotomy is implicitly framed in economic terms. The private goods of publishers are growing at the expense of the creation of public goods by scholars. Knowledge, must surely be a public good. The Jefferson quote we often use could hardly have been shaped better to make the case for which quadrant of Ostrom’s diagram knowledge belongs in. But its also obviously more complicated than that. Jefferson fails to mention to McPherson in his letter that you need your own candle to accept his flame, or at least a taper or lamp. That you can only receive that light by being close to the candle. The flame may be non-rivalrous but it is still exclusive – there remain hurdles to gaining access.

He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening me.

Thomas Jefferson, Letter to Issac McPherson, “No Patents on Ideas,” 13 August 1813.

As with the candle flame, so with knowledge. The aim of Open Access is not to simply nationalise the private goods of the artist formerly known as the publisher. To convert private goods to public goods. History and economics tell us this is not a generally successful path. Our goal, and indeed I would argue that a core goal of scholarship itself, is to make the exclusive goods made within small groups of scholars more public. Since Robert Boyle railed at alchemists, the importance of communicating scholarship has been central to science, and since Plato wrote down Socrates’ complaints about the dangers of writing scholars have debated how new technology affects that communication. In that sense the idea of “public making” and the tension that immediately arises from the question of “to which publics” has always been the central issue of scholarship.

Groups make knowledge

It is easy to fall down the rabbit hole of philosophy of knowledge. So I’m not going to seek to define it here. I will however assert that private experience is not knowledge; that knowledge is a characteristic of groups. In that sense it is born exclusive, as it is held within the group. The process of public-making is both an investment – it requires effort – and a risk – it reduces exclusion. There must therefore be a process of exchange, the group receives something back in return for this public-making. The process of exchange, as knowledge is absorbed and tested and refined by other groups, creates collective (or more public-like) goods that benefit a wider range of groups. New ideas unlock problems, new tools are built that can be generally used, research is applied by others to create wider benefits.

However classical economics, in particular the work of Mancur Olson, tells us that the value of collective goods is not a sufficient incentive for their creation. In large systems it is more in the interest of contributors to try to free-load on collective goods, and in a classic tragedy of the commons, the collective good is created at a lower than optimal level, if at all. It is not at all an accident that knowledge making groups make an exclusive good. Buchannan’s work on the 60s identifies goods with these characteristics as the ones that make groups (specifically clubs providing facilities in his work) attractive to members, and therefore sustainable. We therefore have two high level questions to answer if we want to address sustainability: how to sustain the groups that make knowledge, and how to make it in their interest to invest in making their exclusive “club” goods more public-like.

Institutionalising the scholar

These may not seem like the questions that I suggested I would answer at the beginning but that is deliberate. These are the foundational questions we have to address. Olson and Buchannan as well as Elinor Ostrom offer us some solutions. Olson notes that one solution to the challenge of collective action is to make it compulsory. He is thinking of taxation systems, but the effective compulsion for professional scholars to publish provides a similar mechanism. Ostrom notes that while in many situations facile economic analysis shows coordination is “impossible” – Hardin’s tragedy of the commons – that in the real world institutions can evolve so as to help communities solve collective action problems.

The institution of the research university in which a scholar has resources to work with, gains membership to a set of interlocking “clubs”, and in return is expected to publish is an example of this kind of evolved institution. By “institution” I don’t just mean university, but more broadly Ostrom’s sense of: “the prescriptions that humans use to organize all forms of repetitive and structured interactions”. Merton’s norms are an institution. But so are Mitroff’s anti-norms, the observation that the success of the research enterprise is in part dependent on behaviours exactly opposite to the ones that Merton prescribes: particularism, individualism, self-interestedness and organised dogmatism. To the extent that we recite standards of behaviour, and to the extent that we follow it, or indeed something else, culture is also an institution. This is where the economics intersects with the politics.

This is also where we finally get to publishing (and indeed to libraries). The publishing system is also an institution (or set of institutions) that function as infrastructure. The publishing system provides collective goods: access (albeit to a traditionally exclusive collective), discovery, archiving. The purpose of the institution that is the university is in part to provide a mechanism for compelling scholars to engage with that publishing system so as to provide that collective good. Not by any means the whole mechanism, nor is that its only purpose, but its a part.

Institutionalizing the publisher

The obvious follow-on question is what incentive do publishers have to engage with this collective good production? The obvious answer we would traditionally give is: to extract rents on monopoly rights to generate private goods. But that hasn’t always been the answer. In fact for most of the history of scholarly publishing its been a loss-making enterprise. We need to understand a bit more about that history and the trends that underly it before giving a definitive answer. I am borrowing a lot here from a workshop I went to which is reported in this set of papers in Notes and Records of the Royal Society.

Journal publishing arose out of clubs. First the clubs of emerging national academies like the Royal Society and then out of an increasing set of scholarly (including disciplinary) societies. The term “journal” actually dates from the 19th century and originally applied to serial publications that mixed both research and public interest material as a way to break even. Some of these were money making enterprises for the researchers involved but almost universally pure research and scholarly society journals lost money. They engaged publishers, or rather printers, by paying them for their services and the community subsidized the cost, both directly and through personal and institutional subscriptions.

In the mid to late 19thC relatively few researchers were paid by universities. Only those with endowed chairs were free to focus on research. Many active researchers were independently wealthy, some had wealthy patrons. Many, famously including Charles Darwin, interacted widely with correspondents from beyond the professional research sphere, including the full range of what we might now call citizen scientists. But at the same time professionalization was growing, and membership of the appropriate scholarly clubs, was only growing in importance. The gentleman scholar had the resources to engage, and more often than not the social status to demand attention. The growing class of professional researchers needed to gain access to the club, and publishing in the right place, being read in the right societies was a key part of that.

A journal is a club

My emphasis on clubs is deliberate. When journals (and to some extent the same is true of the University or Scholarly Press) were small, and focussed on a specific, recognizable community then the losses could be absorbed. The collective goods of the journal were created and individual members (subscribers, authors, community members) gained through association and identity. This continues to this day. The enterprise of creating a new scholarly discipline, a new club, invariably involves the creation of its cognate (or is that eponymous?) journal, even if that “journal” does not take a traditional form. See Chris Kelty’s fascinating “This is not an article” for a parallel example of this outside the journal world. The sustainability of these “clubs” with their various tiers of membership and benefits can be understood within the framework of Buchannan’s analysis of club economics.

Buchannan is a classical economist so his models make some assumptions. Specifically that all the club goods are ultimately exchangeable for money and that sustainability means an achievable equilibrium where positive externalities balance negative externalities. We are dealing with only partly exchangeable goods and equilibrium is a long way distant. Nonetheless there are useful lessons to be drawn about possible steady states. In particular Buchannan’s model turns on the question of what happens as the club changes in size. He shows that, where there is friction in access to the club goods – where it is not perfectly non-rivalrous – that this places limits on club size. In the print journal world friction arises in access for authors to the pages of the journal, access to the attention of readers and to the expert criticism of the community.

These trade-offs work when the community is small. Buchannan, Olson and Ostrom all tell us that negotiating these kinds of subsidies to support collective benefits can work when communities are small, homogeneous and have common goals. And to varying extents this was true of the various research communities up until WWII. But then they grew, an explosion of funding and an expansion of the institution of the university both geographically but also within traditional national centres of western scholarship, lead to a massive expansion. And arguably the system has been broken ever since.

The problem of scaling

While the set of readers and authors and funders of a journal are highly similar then there is a common body of culture and knowledge. Amongst other things this commonality reduces a number of costs, specifically of defining whether any given contribution is “acceptable”. Because there are common conceptions of what is within scope the internal sense of “quality” is cheap to determine. The “journal club” is contiguous with a “knowledge club”. This helps to define why society run journals are generally cheaper than commercially run ones, despite the claim that a market economist would make that the commercial run operation should be more competitive and efficient.

As we scale up the system a number of coping strategies emerge. One is specialisation, creating more and more specific communities and journals. This works and is a large part of the history of 20th century scholarly publishing. The price is paid in the creation of silos, an opportunity cost that is hard to delineate in which not only does cross fertilization not take place, but the modes of knowing and testing, the individual culture of each community becomes more and more incompatible with that of others.

If one approach is scale down, then the other is scale out. Scale up the journal to encompass more than one knowledge community. This also has potential to work, it grows the subscriber base and it can create prestige for the journal, attracting authors. But there’s a problem. Because the knowledge community is no longer singular or coherent the cost of determining inclusion becomes externalized. Because the club is no longer capable of determining a single standard a standard needs to be imposed. It turns out that PLOS ONE and Nature actually have the same problem. At PLOS ONE (and Scientific Reports) a small army of people are required to do the process of checking whether articles meet the standards. The most successful areas within PLOS ONE are those where there is a community (paleontology, neuroscience, some parts of genetics) that can do this policing, but across the broad remit of the journal its not possible to rely on academic editors and referees to get across everything and staff need to be brought on to manage that.

At Nature a small army of people (different job titles, different qualifications, but nonetheless) do exactly the same job. And again, they appear on the balance sheet. The complaint most frequently leveled at Nature is that “it doesn’t publish work of type X”. What’s actually happening is that individual editors within Nature have their own views, they create a community, one that is fundamentally predicated on prestige and excitement of being associated with the masthead. But the gaps, and indeed the failures of judgement, illustrate the exact same costs and challenges as the occasional dodgy papers that slip through the PLOS ONE system.

Separating equilibria and the luxury APC market

Another thing happens as the journals spread their scope. The attraction for membership shifts from being membership of the knowledge club, a form of identity politics if you like, to being prestige. Where there is high consistency between knowledge styles, a member of the club can reach an opinion about the quality of a given article. But the author of an astrophysics article in Nature cannot judge an article on neuroscience in the same journal, nor can the astrophysics editor judge that same article, at least not by any means that the neuroscience community would recognize. They can only judge the fact that it appears in that journal. The masthead is a proxy for something but it cannot possibly be a consistent characteristic of the work described in the article.

This might not be a problem in and of itself. In our manuscript on the issues around “Excellence”, we start by advancing the argument that the fact that the word is meaningless might not matter as long as it serves a useful political purpose in aligning the research community. But as we go on to show in that context the concept of “excellence” is actively harmful. It leads to hyper competition, and based on well established findings from behavioral psychology of rewards, leads to performance of excellence, not the characteristics of research we want.

Even this might not be a problem from an economic perspective. Markets need signals and can operate just fine with proxy signals that don’t actually mean anything. These markets tend to be volatile but that in and of itself is not a problem. But, as with excellence, it means that the market focusses on the proxy. In the subscription world this creates the “must-have” journals that authors seek to get into and strengthens their hold on the market. It concentrates power and helps to make big deals work. It focusses the publisher’s attention on maintaining prestige rather than quality of the process (for instance it took Science seven years to implement the same level of QA on statistical soundness as PLOS ONE).

In a subscription world this is just another set of perverse incentives, a lack of alignment between what would provide the greatest collective benefit, and the interests of monopoly rent seekers. This is neither surprising nor wholly avoidable. But something much worse happens with the shift to an APC model. There’s a lot of pretty naive economic analyses of this shift. Most recently you can see the interchange between David Schulenberger in a paper commissioned by the US Association of Research Libraries and Rick Anderson on Scholarly Kitchen.

In what should by now be a fairly familiar one-two the disagreement between the two centres on the power of authors in the market. Schulenberger argues that authors have zero market power, because they have no choice on where to publish, and therefore an effective author-side market can not emerge. Anderson argues that there is sufficient choice by listing prestigious journals in a number of fields, essentially showing there is more than one for any given field.

This form of simplistic analysis, which to be fair I’ve been guilty of in the past, however is unhelpful. The answer one gets tends to depend on how you frame the question. And that in turn is usually determined by how sloppy we are in defining what the market is. We need to understand how the author-side and reader-side markets differ and what is being purchased. Then we need to understand how those markets can break down. There’s a lot of work to be done here.

I want to cut to the chase here though and point out that both Schulenberger and Anderson can be right. Authors can have both substantial choice and buying power and prices can still rise. This happens in markets with specific properties. Markets where prestige sits at the centre. Markets where there is a disconnect between product quality and price. And worst of all, markets where price becomes a proxy for prestige. The APC market that we’ve built has all of these characteristics.

Luxury goods markets operate in situations where there are customers driven by prestige and some proportion are “rich”. Prices rise because it is in the interest of sellers to identify who is “rich”. Think of all those shops in airports, or most recently for me a Tesla dealership on a pedestrian mall in Sydney. Those shops are not for us. They are for people who can afford to walk in and just buy something. Sellers price otherwise unexceptionable goods at high prices to identify purchasers who will pay those high prices. If price can be set as a proxy for prestige (think cars, watches, handbags, haute couture) then the seller wins.

The rational economic move is for the rich person to not play the game. But in the APC world we have perfect conditions for a run away luxury goods market. Prestigious institutions demand prestigious scholars publish in prestigious venues to build up prestige. In the past they paid through the nose for this through subscriptions. But bringing authors into the loop and allowing the concretion of both traditional prestige markers and price as a proxy is potentially toxic. And without delving into the details, the situation seems even worse in the monograph space.

A brief segue on data

It’s worth taking a short sidestep into considering data and data sharing. For better or for worse data publishing doesn’t have the same prestige factor in play. We don’t fight to get our data into Zenodo and when rebuffed head to Figshare or vice versa. This is good – the same run away luxury goods market isn’t emerging – and bad – the institutions that motivate data sharing aren’t there for many disciplines. Where they are they are tied to formal article publication.

Data “publishers” perform much the same economic function as article publishers. They provide an arbitrage between the needs of readers and those of authors. They identify and define knowledge communities. They manage the quality assurance processes that the community defines. The objects are somewhat different but there is no reason to expect the economics to be fundamentally different. Differences in their funding will be signals of historical differences or differences in the value that a community places on articles vs data.

The issue of prestige we have already noted. This means, at least in part, that data sharing infrastructures must be more coupled to knowledge clubs, a factor that is clear in the disciplinary focus of many mature data sharing services. Adoption of broad-based data sharing services is driven more by either direct author benefits to sharing or by mandates. If a judgement were forced it might be argued that sharing through Figshare or Zenodo gains less prestige or community membership than submissions to PDB, Genbank or model organism databases. This actually mirrors community perceptions of general journals like Nature in the early part of 20th century. Publishing in disciplinary journals was viewed as more important – the idea that something of general interest is of more value is actually quite recent.

The funding of data infrastructures is also a challenge, and more so because the lack of prestige reduces the opportunities for creating commercial interest through monopoly rent opportunities. There are some examples, but they are generally residual monopolies from the print world. The Chemical Abstracts Service of the U.S. American Chemical Society stands out as an example. Other resources have gained the status of being two important to fail. The World Wide Protein Data Bank is an example of an infrastructure that is effectively top-sliced funding, albeit through the somewhat unstable mechanism of uncoordinated grants from a range of national funding agencies. Europe Pubmed Central is supported by a consortium of funders through a formal mechanism brokered by the UK’s Wellcome Trust.

Olson’s three options and the story of Crossref

These different stories map onto the paths that Olson described in the 60s as the means by which groups could solve collective action problems. Above a certain size he saw only three options. First is a kind of technicality, where the community is structured so that a small number of players dominate the space. Europe Pubmed Central is an example of this, the Wellcome Trust and a few other players, originally restricted to the UK could get around a table and just decide to act.

The second path is one where there is a non-collective benefit that contributors receive in exchange for supporting the collective benefit. Subscriptions are the traditional version of this, but membership models of all sorts generally involve a benefit. Sometimes, as I would suggest is the case for the Open Library of Humanities, that benefit is being seen to provide leadership, demonstrating a progressive stance. There will be more models like this in the future but there are questions about how they can scale – its not a cool kids club when everyone is a member. The Cambridge Crystallographic Data Centre is an example that tries the square the circle, providing some subscription access benefits while making most data available in some form.

The final approach is compulsion. At the nation state level general taxation funds collective goods and we (or at least most of us) don’t get a choice about contributing. Access to the Chemical Abstracts Service is pretty much a requirement of running a serious chemistry operation, one that the ACS protects by trying very hard to ensure its also a requirement for certification for chemistry courses. Regulatory capture is also a good strategy, as regulated professions like medicine and the law can attest. Standards, pace Malamud, operate in a similar manner.

The story of Crossref, the provider of article DOIs and de facto home of scholarly bibliographic metadata is instructive. Crossref was started essentially by fiat by a small number of publishers. In practice between five and nine publishers dominate the market. The heads of houses (they have been known to call themselves that on occasion) will meet and agree certain things. The setup and early funding of Crossref was one of those. This is Olson’s first option, effective oligopoly.

Once setup, Crossref offered a membership benefit. The growing infrastructure of DOI referrals and the improvements in article discovery led to traffic to publisher websites. And those numbers could be sold back to subscribing libraries. Both the ability to assign DOIs and the traffic are non-collective benefits of membership, and ones that increase in value as the collective good, the growing pool of public metadata and infrastructure to use it, grows in turn.

Finally today, membership of Crossref is effectively compulsory for any serious publisher of scholarly articles in STM, and increasingly for book and HSS publishers. It’s just part of the costs of doing business. A compulsory, tax-like, part of the system. Now compare the growth of ORCID. ORCID was also started largely because publishers decided it was a good idea. A few funders came on board but it was really publishers who put the money in up front. Gradually the funders have come in (again, really there are maybe 20 really important funders at a global scale), but the institutions? Institutions have been rubbish, despite actually being the place where the greatest benefits probably accrue. There are just too many to get past the collective action problem. We’re starting to see progress now, not because the universities have got their act together but because national level coordination is starting to kick in.

This story repeats itself pretty much every time the scholarly communications infrastructure needs an upgrade. Publishers move first, not because they have the most capital. The institutions do. Harvard could buy Elsevier and Wiley and T&F and SpringerNature and have cash to spare to throw in Calvariate or whatever they’re calling the rebranded ISI and PLOS for good measure. And not because the publishers have more freedom of movement. Publishers have share holders, or at least owners, who want a return. Many funders are in a much better place to call shots. The publishers move because they can, because there are few of them.

The future of the library

Which brings us to the university library and its unique place. What can we learn here? Maybe the IR infrastructure is a counter example? What do these models tell us about the success of institutional repositories? Well it suggests that those that started in a department (small, homogenous) will work better than those started at scale (Harvard and Southampton spring to mind). That those which provide a non-collective benefit back may be able to scale (Minho provides a great example of this with its competition with cash prizes for the best department). And that where those two do not hold that success will depend on effective compulsion (Liege, and the genius of Bernard Rentier obviously).

The library itself is a club good, membership of the university provides access to resources and people. The digital resources are non-rivalrous, but exclusive – even if that is exclusion is no access control. But Sci-hub and its inevitable successors blow an enormous hole in that. Privileged access to digital content will not survive as a sufficient value. Access to people, to expertise, now that’s different because that’s rivalrous. Proximity still matters. But services can also be bought elsewhere and the artist formerly known as the publisher is rapidly moving into that space. Neutral, non-biased advice in the interest of the researcher or broader university member is a good differential to start with, but focussing on how membership of the university is linked to access to the right kind of relevant expertise is key.

The institution itself, and the library within it, is pretty much at the worst of all worlds scale. Too large for collective action, too small and internally fractious to act unilaterally. Too many to solve the combined collective action problem. Collectives, thematic and geographic might provide solutions, and some actions the institutional leadership can take unilaterally and internally. But as currently disposed the university system globally is almost designed to prevent effective action.

The ways forward

The thing that the scholarly community has, if it can regain control of the institution is capital. I know it’s not proper to turn into a Marxist in middle age but this is fundamentally a question of control over deployment of capital and the position of labour. The labour sits almost entirely within the university and the scholarly community. Between libraries, scholarly societies, scholars and the institution.

But capital, or at least free capital, has been deployed largely by corporate interests and publishers in particular. Except that there is much more capital sitting with the funders and more importantly the institutions. Maybe it is time to challenge what the endowments and pension funds of institutions are supposed to be for. Maybe a trillion dollars is tied up, not being put to the use of the mission of these – often legally speaking charitable – institutions.

If you accept my argument that the institutional purpose of a university is to compel the publication of research outputs, as part of the contract of being a professional scholar, then the true Marxist would necessarily argue that the institutional capital, both financial and non-financial, must be liberated into the hands of labour. Ok, that may be a little radical. Particularly here in the Triangle.

There are some more practical goals. Ostrom shows us that institutions can grow in which collectives form collectives. If we focus on good community design and on shared principles of behavior, on building strong culture then our local communities, where we can solve the collective action problem, can band together at a higher level to solve the bigger problems. Stop seeing the university as a company with a CEO and see it as a collective of communities with a Community Manager. Leadership is not a dirty word in communities: leading from the front, leading from the middle, or leading from behind can all be successful. But leading by fiat will not be.

Still too radical? These communities are growing. Support them. The rise of collectives, OLH, Knowledge Unlatched, but also new forms of scholarly societies: the Research Data Alliance, ArXiv, FORCE11. Help these to grow, support them financially where you can, and with moral support where you can’t. We can work towards service definition. Let’s take the publisher’s at their word. They want to be a service operation: lets collectively define those services and our requirements. When I talk to the “big bad wolves” they are just as keen to have the space where they should bring their commercial sensitivities mapped out. But they can’t wait forever for us to make up our mind.

We need to reduce costs. I haven’t really talked about the structural aspects of costs in scholarly communications and how they are changing but there are opportunities for orders of magnitude reductions. To achieve that we need shared infrastructures. But for those we need funding models, which will necessarily be compulsory, and for those we need some form of shared governance. No taxation without representation. What does that mean in our space? And how can that be equitable. One of the other things Olson shows is that in delivering collective benefits the small will exploit the large. Time for the big institutional players to suck it up and play that role.

Finally we need to do everything in our power to sever the connection between prestige and price. By whatever means are necessary that link needs to be stomped on, disavowed or destroyed. Frankly, this one seems harder than the Marxist revolution. But if we fail at this we will reap the whirlwind. APC based models are growing. Separating equilibrium gives us two things, run away luxury prices, and a commodity product market for those that can’t afford the costs of luxury.

This isn’t just inequitable, and it doesn’t just damage the diversity of views on which scholarship depends. It is unaffordable at both ends. First because the luxury goods market will rise until it gets smaller and smaller, bankrupting the publishers that need to chase that extra differentiating piece of leather trim first, and then the research communities that are paying for it. But at the other end the commodity market is being supported by the rise, not of the predatory journal, which are frankly small fry in the scheme of things, but the predatory author.

The industrialization of fraud and plagiarism is already leading to an arms race which will rapidly raise prices. Just as we’ve seen at PLOS ONE and Scientific Reports, the potential savings from technology development will be swallowed up in layers of checks for author and referee identities, validating figures, checking for ever more sophisticated plagiarism. That’s the real story of the price rises – that the drop in the quality of submissions leads to an increase in the necessity for professional checks. And that is also unaffordable.

The optimal answer probably lies in smaller journals, and smaller, community based, data repositories, built on a common infrastructure that makes these cheap to run. An efficient market in services where innovation isn’t driven entirely by the possibility of a 100x sell out, but where commercial competition also plays its role. I don’t know if its possible to get there from here. But if we can then the big shift will be the role of knowledge communities, taking control of their own processes, within a shared framework that makes that easy to do.

To return to my original point. We want communities to create knowledge, and we want them to invest in sharing that to wider audiences in a way that is both principled and makes economic sense. The best way to do that is to provide platforms and infrastructures that make that investment both cheap, and provides good returns. What does the community, the club, get back in return? We will need new forms of economics to figure this out, the work on the economics of clubs and of commons, is still only beginning. But this will be the crucial knowledge we need to design new and improved institutions for the future.