Opening Access to Research

Transcription

1 Opening Access to Research Mark Armstrong Oxford Naples: October 2014 Armstrong () Open Access Naples: October / 22

2 Overview Two claims: many publishers charging libraries too much for content everyone should be able to read published research for free, especially if research was publicly funded First is surely a major problem House of Commons report (2004): Mounting concern that the financial benefits from the Government s substantial investment in research are being diverted to an excessive degree into the pockets of publisher s shareholders Both issues may be overcome by an open access policy: gold : regulated authors must publish in journal with immediate and free access to readers green : regulated authors must provide inferior version (working papers online, publish in journal which gives free access after embargo) Armstrong () Open Access Naples: October / 22

3 Some journal prices Subscription funded: Small: subscription $5256 bundle of seven AER journals: subscription $735 average research-oriented university spends about $6m on journal subscriptions Open access: to read Jealously in dogs in PLOS ONE costs nothing, but author paid $1350 [PLOS ONE published 26,000 articles in 2012] Delayed open access: PNAS: free access after six months [subscription is $2800 for research university] Hybrid: Economic Journal: subscription 477, but author can opt to pay 1500 to make article open access Armstrong () Open Access Naples: October / 22

4 Traditional journals and their alternatives Journals traditionally offer a number of benefits to authors and readers: certifying article s novelty via journal s reputation distributing article to those (and only those) permitted to read it provide definitive typeset version of article improving manuscript with suggestions from editors and referees First of these is freely available to readers even in traditional reader-pays model Fourth is freely available if journal permits author to post accepted version Armstrong () Open Access Naples: October / 22

5 Authors value certification others working in same area know a researcher s skill, regardless of journal but author cares about other people s evaluation too (tenure committees, job offers, Nobel Prize) Readers value certification an article is an experience good readers value ex ante information about quality before deciding whether to read a paper But now other signals in addition to journal s name Google scholar citations etc increasingly used in committees (and even by journal editors deciding what to publish?) wisdom of crowds vs. a couple of referees (who probably disagree anyway...) but danger of ineffi cient herding? Armstrong () Open Access Naples: October / 22

6 PLOS ONE s editorial policy: too often a journal s decision to publish is dominated by what the editor thinks is interesting and will gain greater readership PLOS ONE will rigorously peer-review your submission and publish all papers that are judged to be technical sound judgments about the importance of any particular paper are then made after publication by the readership (who are the most qualified to determine what is of interest to them Journal accepts about 70% of submissions and (currently) has impact factor of 3.7 Streamlined refereeing/decisions has many benefits any decent paper refereed only once, so more time for research reduced delays to publication benefit authors and readers current system in economics out of hand? [50 years ago the R&R option was rarely used, papers one-third of the length] Armstrong () Open Access Naples: October / 22

7 Authors value readership but care mostly about other researchers (who generate citations) not the wider public But some benefit to the wider public seeing latest research researchers in poorer countries/institutions vast public interest in medical research Putting content behind a paywall does make a difference: 36 traditional journals allowed Davis (2011) to randomly select articles for open access treatment for the subsequent year their downloads were about double those of regular papers but no significant increase in citations three years later consistent with view that the readers who generate citations have easy access to subscription content in any case [but how to control for (i) readers finding self-archived versions of regular papers and/or (ii) readers taking label open access as a cue for higher quality?] Armstrong () Open Access Naples: October / 22

8 Internet allows authors to bypass journal distribution: author posts paper on own webpage, institution s repository or subject-specific repository (e.g., Repec) can be done with or without permission of publisher (e.g., in last issue of the Economic Journal, three authors have illegally posted the published pdf on their webpage) authors have incentive to be careless about copyright (they do not get the subscription income themselves, and benefit when more people see their article) if posted version is the accepted version, and posted immediately, there is de facto full open access copyright policies vary widely across publishers (e.g., Elsevier allows immediate posting of accepted version on author s webpage, while for economics Wiley and OUP require authors to wait 2 years to do this) Armstrong () Open Access Naples: October / 22

9 Propensity to self-archive varies across subjects lots in economics, physics, maths, computer science almost never for books in humanities (since authors lose out on royalties) very little in life sciences, where medical journal such as Cell do not allow submission of papers which have previously been posted on the web (the Ingelfinger rule ) even when allowed to do so, many authors do not bother to self-archive, so may need element of compulsion in effective green open access policy Armstrong () Open Access Naples: October / 22

10 Costs and revenues Traditional journal tasks cost money in old days, many prestigious journals had financial problems But electronic technology means many costs have fallen: distribution and storage is free cheap editorial software [Editorial Express costs just $2000 per year] means journals might not need admin staff or premises Publishers can now better extract surplus from libraries monitor download activity bundling means large publisher can more accurately predict an institution s willingness-to-pay for its collection like first degree price discrimination, more surplus is extracted yet more libraries are served In sum, technological change has led to dramatic imbalance between costs and prices Armstrong () Open Access Naples: October / 22

11 One necessary input is not paid for: the article itself unlike books, the norm for journal publishers is not to pay authors for their work (why?) means high revenues from readers are not fully passed on to authors, but retained as super-normal profits Commercial journal publishers are indeed highly profitable: e.g., Elsevier in 2012 made profit of 780m on revenues of 2063m [ 3 revenue per download] Sometimes journal profits given to learned societies who use revenue for conferences, scholarships, plush headquarters etc e.g., American Astronomical Society gets annual $4.8 million from its three journals many such societies lobby against open access regulations Armstrong () Open Access Naples: October / 22

12 Estimates of cost of publishing a paper vary widely Depends in large part on a journal s rejection rate Nature claims its cost is in the range $10,000 30,000 highly selective PLOS Biology author fee is $2900, while less selective PLOS ONE author fee is $1350 Costs of running a pure open access journal lower than a hybrid or traditional journal no team needed to negotiate contracts with libraries, or to monitor copyright compliance Armstrong () Open Access Naples: October / 22

13 Pros and cons of open access: distribution Like other diffi cult markets, journal market is a competitive bottleneck Key asymmetry between authors and readers authors can publish paper in only a single journal ( single-homing ) each article is a mini-monopoly readers have to subscribe to full range of journals to be sure to get an article they need ( multi-homing ) outcome often that authors can publish for free, while readers pay high prices see also: call termination, bank accounts, credit cards, newspapers, search engines Armstrong () Open Access Naples: October / 22

14 Opening editorial to leading open access journal PLoS Biology: each journal has monopoly on a resource vital to scientists the unique collection of articles it has published. Anyone who depends on the information in a specific article has no choice but to pay whatever price the publisher asks Outcome would be different in alternative world where a reader subscribed to just one journal, and authors had to place their work in several journals to reach a wide audience akin to newspaper market, where each reader buys single newspaper, and advertisers must place the same ad in several papers to reach enough eyeballs Armstrong () Open Access Naples: October / 22

15 Example Initially two parties: author and reader author values reader seeing article at $10, reader values seeing article at $10 it costs the authors some small amount to deliver article to reader (say, posting on internet) joint surplus of interaction is about $20 Now third party, a journal, comes along makes offer to author to deliver article for free, which author has small incentive to accept journal charges reader $10 for access joint surplus of author and reader is now $10, with the rest extracted by journal [if journals could pay the author, competitive outcome is that they offer about $10 to the author, and sell to reader for $10; joint surplus is then $20, but all goes to author] Armstrong () Open Access Naples: October / 22

16 Gold open access policy Regulated author forced to publish in journal which makes article available immediately for free to readers entirely overcomes monopoly pricing problem library spend reduced, wider public has access But authors will likely have to pay to publish even with research funds, there s an opportunity cost evidence many people pay out of their own pocket supply of articles will contract [extreme case: supply of new economics textbooks would fall if authors had to pay rather than be paid] article not being published even worse for readers than high prices For now, top-5 economics journals don t have open access option Armstrong () Open Access Naples: October / 22

17 Green open access policy Regulated author must self-archive accepted version or publish in journal with (not too) delayed open access inferior version made freely available gives libraries an outside option, so forces journal to charge less for published version may still leave enough revenue from subscriptions to allow authors to publish for free supply of articles not reduced, so can be better than Gold policy Armstrong () Open Access Naples: October / 22

18 Pros and cons of open access: certification With gold policy, reader price cannot adjust for value-added supplied by journal pretty formatting, valuable comments from referees, certification usually, when a seller improves quality, it can set a higher price but when price is fixed (at zero), incentive to boost quality is lower Might mean that there is less certification in equilibrium? harms both readers (who want to be guided in what to read) and (good) authors who want to be seen to publish a good article Armstrong () Open Access Naples: October / 22

19 Example Author might exogenously have good or bad paper author knows quality has intrinsic benefit from publishing in journal with higher average article quality reader cannot observe quality without effort of reading or journal certification journal can sell paper known to be good for $10k, cannot make any revenue from paper known to be bad Journals come in two forms (many of each) a non-discriminating journal publishes anything, incurs no costs a discriminating journal inspects each submission, which costs $5k each time (but no other costs), and publishes only good papers whether a journal is discriminating is common knowledge (reputation etc) Armstrong () Open Access Naples: October / 22

20 Example Without regulation: discriminating journals compete for good papers good author publishes for free, readers pay $10k for access to discriminating journal journal name is perfect signal of quality bad papers go unread (but may still be published for vanity reasons) With gold policy: journals must recover cost of peer-review from authors to publish good paper in discriminating journal costs author $5k only do so if intrinsic benefit from publishing well is greater than $5k or author has generous funds Armstrong () Open Access Naples: October / 22

21 Some good papers will appear in non-discriminating journals author either has insuffi cient funds or does not care enough if fraction of good papers in non-discriminating journals is low, readers will not bother sifting through the dross and good papers go unread if fraction higher, readers will read non-discrirminating journal, and incur disutility from having to read bad papers Model assumes journal name only signal of quality what about Google scholar citations etc? but early readers must rely only on journal name but if good papers go unread by early readers, they can never take off so journal certification may still be important Green policy again may allow authors to publish for free, and certification continues to be effective Armstrong () Open Access Naples: October / 22

22 Conclusions Green policy seems superior to no regulation in many subjects, little propensity to self-archive indirect impact on the prices publishers can charge compatible with authors publishing for free Gold policy more radical and disruptive Big unknown: to what extent does wisdom of crowds replace journal name as effi cient certification device? if it can, then PLOS ONE type journals may become dominant, with low-key peer review and low author fees if not, then (costly) journal certification remains important but published authors should not pay entire cost eg, allow regulated authors to publish in cheap but not necessary free journals, and/or have submission rather than publication fees Armstrong () Open Access Naples: October / 22