Academic Journal Publishing and the Open Access Movement
(1)
Faculty of Law, Ruhr University of Bochum, Bochum, Germany
3.1 The Academic Journal Publishing Market
In this section we elaborate on the general structure of the academic publishing market. In fact, it is important to understand how the academic publishing market and especially the journal publishing market has evolved over time. First, we will generally look at the principles of academic publishing, creating a basic understanding on the different models for scholarly publication and the players involved. Second, we will elaborate on the specific market characteristics. Facts and figures on the publishing model provide with a comprehensive overview on the immanent structure and changes of market characteristics in historical reflection. The implications drawn from this industry analysis lay the foundations when later reflecting on the effects of certain legislative or policy changes.
3.1.1 The Principles of Academic Publishing
Academic Publishing Models: An Overview
In general, three ways of distributing academic works may be distinguished: (1) Closed Access Journals, (2) Open Access Journals and (3) Self-Archiving or Repositories (Fig. 3.1).
Fig. 3.1
Business models in academic publishing: an overview
First, closed access journals generally revert to the basic principles in academic journal publishing. That is, the publishing process involves a peer-review for quality selection, but also other services like typesetting and editing tasks. In general, “closed access” means that some form of restriction on the access and use of the journal content is imposed on readers. Publishers provide access to their journals subject to the payment of an individual or institutional subscription price. In the digital era the supply of so-called “Big Deals” has become common practice in the academic journal market, i.e. institutional subscribers (e.g. university libraries) subscribe to online aggregations of multiple journal titles through consortial or site licensing agreements (Houghton and Oppenheim 2010, p. 42). Journals are run by both commercial publishers as well as scientific associations. Second, open access journals revert to the same general principles of journal publishing. That is, the nature of academic publishing remains the same as a peer-review process shall guarantee a minimum quality and other publishing services are provided. In contrast, “open access” means that access to the journal content is not subject to individual or institutional subscription, but is provided free of charge to readers. The publishing costs are typically born by the author (author fee) or funding agencies. In particular, two forms of OA journals can be distinguished: (a) hybrid models and (b) pure or true OA models. While the latter follows the OA principles outlined in the Budapest Open Access Initiative (2002),1 hybrid OA does not fully satisfy the conditions laid down in the 2002 report. Several hybrid models exist where journal publishers provide with OA to only parts of their set of journals (including only particular works or snippets of a paper), or only in retrospective as well as with a delay of 6, 12 or even 24 months after the publishing date (Bernius et al. 2009, p. 106). A very frequently applied method is Open Choice. With Open Choice authors can make their articles freely available in a separate OA issue upon payment of an author fee.2 Third, the emergence of the internet has provided with an additional business model for disseminating academic works—self-archiving or repositories. In contrast to journal publishing, self-archiving does not involve any peer-review or other services generally associated with the publishing process. Instead authors are enabled to deposit their works on internet platforms (repositories), where they can make their works available to anyone with internet access and free of any subscription and/or author fee. The requirements for article quality typically revert to a simple academic character threshold, i.e. the paper has to exhibit scientific attributes. As a result, self-archiving can primarily be seen as a platform for disseminating pre-print versions of recent papers and hence as an additional opportunity for communication and discussion.3
Academic Journal Publishing
The traditional journal publishing model involves basically three players: (a) scholars who are typically producers (authors) and consumers (readers) of academic works at the same time, (b) publishers who act as intermediaries between authors and readers, and organize the peer-review process as well as the editorial services in bundling different papers to an issue of a journal publication, whereas the access to academic works is facilitated by (c) libraries who subscribe to sets of journals and bargain on behalf of the group of researchers and (university) students with the journal publishers (Bernius et al. 2009, p. 104). The legal framework that enables interaction between the players is the copyright system. The author of an academic work typically transfers an exclusive commercial right to the publisher. In the end, the relationship between the players in academic publishing are affected by three different market mechanisms.
First, the mechanism characterizing the relationship between a scholar and the publisher is reputation. With a publication of a paper in a particular journal, a scholar receives a quality or reputation signal. This signal reflects the standing of a particular journal in a certain discipline and is hence historically based. As outlined before, the measurement typically referred to for estimating the level of reputation of a particular journal is its impact factor. The impact factor was devised by Eugene Garfield (see Garfield 1955) and measures the average number of citation of a journal in a particular year or period.4 The impact factor is calculated on a yearly basis for journals listed in the Journal Citation Reports (JCR) by Thomson Reuters.5 The JCR allows the ranking of journals by discipline to not only compare different journals between disciplines, but also within a discipline by means of their impact factor.
Second, actions between publishers and the library are coordinated by a price mechanism. Libraries subscribe to journals or sets of journals for a certain time period—typically multi-year agreements. The subscription price paid by a library is the result of a bargaining process between the two parties. In order to coordinate and bundle similar demands and reach a better bargaining position, libraries often form consortia. With the advent of online publishing, however, “big deal” contracts have become common practice. In this context, libraries purchase their journal subscriptions from large publishers in form of bundled site licenses that allow electronic access to nearly every journal of that publisher. The price for such “big deal” subscriptions depends on the historical expenditures for print subscriptions from the respective publisher.6 While before the online publishing era institutional subscription prices for universities were identical and independent from previous holdings, the “big deals” offer a contractual strategy for a diversified pricing scheme. Moreover, big deal contracts often contain confidentiality clauses that prohibit that information on prices and contract details are shared between libraries. Finally, the decision on the composition of journals in the “library’s basket” depends not only on the price, but is closely linked to the previous reputation mechanism. Thus, journals with a high impact factor form the core in the library’s holdings. In an “ideal world” the choice of journals could be simplified to subscriptions to all journals above a certain impact factor threshold. Nevertheless, the prevailing practice of “big deals” and hence the bundling of fixed sets of journals by publishers (journals with high and low reputation) sets the boundaries in the freedom of the libraries’ decision making process.
Third and following from the character of a library as an intermediary between publishers and readers, the use of library collections by their scholars (and students) characterizes the last market mechanism. Libraries shall provide with an adequate supply to information for their researchers and (university) students. Accordingly, access to certain journals within the library’s collection can be seen as a club good. Club membership is subject to employment (researchers) or university enrollment (students).
In an universal OA regime, however, the role of libraries becomes somewhat obsolete.7 Scholars, as authors and readers, interact directly with a publisher. The market relationship between the two parties is characterized by both market mechanisms: reputation and price. An author of an academic work pays an author fee to the publisher.8 In return he receives the reputation signal for a publication in the respective journal. In contrast, the usage of works is not limited by means of individual or institutional subscription. Journal content is freely available via the public internet without any legal, financial or technical barrier other than internet access itself.9 Figure 3.2 summarizes our findings in comparing the principles of journal publishing between both worlds—the CA and OA mode—and illustrates the players involved and the market mechanism characterizing their relationships.
3.1.2 The Journal Publishing Market: An Industry Analysis
A Brief History of the Journal Publishing Market
The first scientific journals appeared in the late seventeenth and early eighteenth century in central Europe. A period of intense transformation: the Enlightenment. The aim to challenge ideologies that were grounded in tradition and faith and to foster the dissemination of knowledge throughout the society by means of scientific methods led to the launch of the first academic journals. Early academic publishing modes were primarily organized and coordinated by academic associations. The first scientific journal in Europe was the Journal des Scavans (later renamed to the Journal des Savants), which was first published on January 5, 1665, by Denis de Sallo. In England, the Royal Society of London launched the Philosophical Transactions on 6 March 1665.
Most interesting is the relationship that the early academic journals had to the emerging copyright system. Despite the fact that both developments have their origin at about the same time period, there is little evidence of causality. In contrast, Ramello (2010) highlights that “the first two scholarly journals [Journal des Scavans and the Philosophical Transactions] had a somewhat problematic relationship with copyright and its ancestors (Ramello 2010, p. 12). Both journals eventually owe much of their early success and prominence to the presence of pirated copies that were widely used and distributed in France and England. For a long time the relationship between academic journals and copyright remained merely occasional, since academic journals were primarily published by learned societies and then academic institutions (Ramello 2010, p. 13). Even though copyright did not play a vital role in the relationship between publishers and authors at this time, licensing (implicitly or explicitly) became the common framework in shaping the rules. The consequences of this still prevailing contractual framework only became evident when commercial publishers started to enter the market in the 1960s and 1970s. The market entry of commercial publishers did not only happen by launching new titles or by filling niches in the academic journal market, but also on behalf of scientific associations and hence by acquiring existing titles. Over the last years this development has led to a significant concentration in the academic publishing industry (Edlin and Rubinfeld 2004; McCabe 2002; Nicita and Ramello 2007; Ramello 2008, 2010).10 The consequences of which will be discussed in the next section.
The Distinctive Features of the Journal Publishing Market
The distinct features of the journal publishing market concern all players of the academic journal market. Accordingly, we will first focus on the market characteristics for publishers of academic journals. We will also look at the specific market conditions from the perspectives of researchers and the libraries, before we conclude what follows for our analytical framework.
Publishers
A closer look at the publisher’s side of the academic journal market reveals significant differences between commercial and non-commercial publishers as well as between closed and open access journal publishers.11 The trend towards a higher concentration throughout the last years has led to a situation where basically three commercial publishers dominate the journal market: Reed Elsevier, Springer and Wiley Blackwell (McGuigan and Russel 2008). In fact, OA journals seem to play hardly any role when comparing journal impact factors between open and closed access journals. Accordingly, journals by established closed access journal publishers, like Reed Elsevier, Springer and Wiley Blackwell, show significantly higher impact factors than those of OA journal publishers.12 A closer look at the distribution of impact factors among the major publishers of open and closed access journals supports this “reputation advantage” of closed access journals. Figure 3.3 captures the distribution of the impact factors (x-axis) for CA (only the big three) versus OA journals. The y-axis reflects the relative number of CA or OA journals with an impact factor that is within the boundaries of the interval.
Fig. 3.3
Impact factor distribution: CA journals vs. OA journals
Testing the impact factor advantage of CA journals over OA journals shows that CA journals exhibit in fact significantly higher impact factors. Comparing the distribution of all OA journals with all CA journals reveals significant differences in the impact factor distribution. Figure 3.4 shows that OA journals show by far higher frequencies in lower impact factor classes.13 While 456 and hence about 55 % of all OA journals have an impact factor of no higher than 1.0, it is 2,844 or about 38 % of all CA journals. CA journals instead reveal significantly higher impact factors in higher impact factor classes. Considering that CA journals have a general market advantage by combining 90 % of all ISI listed journals, the existing market dominance is even more evident. That is, only 10 % of all ISI listed journals are considered to fulfill the conditions of an OA journal (DOAJ definition).14 The Pearson 
approves this impression of an impact factor advantage of CA journals.15
approves this impression of an impact factor advantage of CA journals.15Fig. 3.4
The impact factor advantage of CA journals
The market dominance of only few commercial publishers becomes even more evident when reverting to market shares as the number of closed versus open access journals with an impact factor of greater or equal to 1.0 (2.0). In this regard, the three giants (Reed Elsevier, Springer and Wiley) manage to combine more than 36 % of all journals with an impact factor greater or equal to 1.0. When considering only journals with an impact factor of at least 2.0, the market share even amounts to almost 37 %. In contrast, the four main OA publishers (including HINDAWI, BioMed, PLoS and Medknow) only publish 2.61 (1.75) % of all ISI listed journals with an impact factor of at least 1.0 (2.0). The average impact factor of an OA journal is 1.44. Looking only at the major four OA publishers, however, journals show on average an impact factor of 2.64, ranging from 0.93 for a journal published by Medknow to an average of 8.51 for the seven journals published by the Public Library of Science (PLoS). However, a maximum impact factor of 16.269 for PLoS Medicine shows that there are outliers for journals of particular disciplines.16
Consequently, especially PLoS but also BioMed have established competitive publishing outlets. This fact is also reflected when looking at the relevance of OA journals by discipline. Accordingly, OA journals may primarily be considered as a sufficient publishing outlet in “Biology & Life Sciences”, “Health Sciences” as well as in “Physics and Astronomy”, whereas it seems not to play any role in fields like “Economics & Business” or “Law & Political Sciences”, where not even a single ISI listed OA journal exists.
The boxplot of Fig. 3.5 reveals a bulk of journals in the fields of “Biology & Life Sciences” as well as “Health Sciences” with an impact factor ranging from about 1 (first quartile) to about 2.5 (third quartile), whereas outliers reaching impact factors of above 16 or almost 18 are present in “Health Sciences” and “Physics and Astronomy”, respectively.17 An impression on the distribution of OA journals and ISI listed OA journals by discipline is also readable from the legend of the x-axis. Here, the columns represent the numbers of OA journals and ISI listed OA journals by discipline, where the first row of each column shows the absolute number of OA journals listed by Thomson Reuter’s Web of Science (ISI listed), row two the absolute number of all OA journals in the respective discipline and row three the discipline.18 Obviously, in the fields of “Business & Economics”, “History & Archeology” and “Law and Political Sciences” there is not a single journal which is listed by the Web of Science database.
Fig. 3.5
Boxplot—impact factors of OA journals by discipline (Source: Mueller-Langer and Scheufen (2013, p. 368))
All of the above clearly depicts reality in the academic publishing market and provides evidence to the fact that the prevailing coexistence of closed and open access models is far from reaching a “fair and reasonable” level of competition. Despite the vast increase in the number of OA journals in the past decade, OA publishers still lack in signaling a sufficient level of reputation to attract both readers and authors.
Researchers
The previous section clearly reveals a competitive advantage of established and highly ranked CA journals. There is a young branch of literature that has been investigating the attitude of academics for publishing in either of the two regimes. A survey of 481 researchers from different disciplines by Mann et al. (2008) emphasizes that authors tend to adopt a “wait-and-see” attitude in making use of OA. The authors show that two main variables tend to influence the decision of the publishing outlet: the expected performance of OA and the peer use. Eger et al. (2013) run a survey in Germany for both universities and research institutes. The sample of 2,151 respondents clearly reveals large difference in the attitude to and the experience with both OA models, i.e. OA journals and self-archiving. The pillars in Figs. 3.6 and 3.7 display the experience of researchers from different disciplines in publishing papers in OA journals (light grey) and/or with self-archiving platforms (dark grey) for researchers from both universities (Fig. 3.6) and research institutions (Fig. 3.7). Obviously, OA journals are frequently used as a publishing outlet especially in “Biology & Life Science” and “Health Sciences”. Whereas self-archiving is a common model in “Mathematics & Statistics”, “Physics & Astronomy” as well as “Business & Economics”. Interesting in this respect is, that self-archiving is hardly ever used in “Biology & Life Science” and “Health Sciences” which showed significantly higher levels of experience with the gold road of OA publishing. Moreover, the authors introduce two indices to mirror both the personal awareness with the concept of OA and the relevance of OA journals within the respondent’s discipline, where 1 = very low, 2 = low, 3 = middle, 4 = high and 5 = very high. That is, the higher the average number reflecting the personal awareness of a respondent or the general relevance of OA journals in one’s discipline, the higher was the level of awareness or relevance of OA, respectively. The average count for each discipline is captured in the awareness and relevance factor as an index for knowledge and attitude. The secondary axis in Figs. 3.6 and 3.7 refer to this awareness/relevance factor ranging from 1 = very low to 5 = very high. Not surprisingly, Eger et al. (2013) observe a higher index in such disciplines that exhibited a rather low degree of OA publishing and vice versa.
Accordingly, the immanent rules may explain why OA publishing does not play a vital role in most of the disciplines. The reward system in science and especially the “publish or perish”—environment may somehow force (particularly young) researchers to publish their papers in top-tier journals and hence in journals with high impact. As a consequence, OA journals may be less valuable for the researcher’s career concerns.
Libraries
Libraries have been facing considerable changes over the last decade in the way they can negotiate the terms and conditions for journal subscriptions with publishers (“big deals”). As a result, serial expenditures have been steadily increasing over the last 20–25 years. As already discussed, Ramello (2010) provides evidence for increases in serial expenditures by 273 % and serial unit costs by 188 % from 1986 to 2004 in the US, as compared to an increase in the consumer price index of 73 %. Edlin and Rubinfeld (2004) even provide evidence for increases of more than 600 for the time from 1984 to 2001, in disciplines like physics. Similar finding for Europe is provided by Dewatripont et al. (2006). McCabe (2002) shows that the vast increase in subscription prices was at least partially caused by significant mergers between large commercial publishers. A comprehensive overview on the development of journal prices and contract conditions is provided by Ted Bergstrom and R. Preston McAffee, who have been gathering data and information on journal prices by evaluating university subscription contracts.19 Accordingly, average journal prices for 2011 were ranging from 109.13 USD in law to 1,486.37 USD in physics for non-commercial publishers, and 713.03 USD in history to 3,174.48 USD in physics for commercial publishers (Bergstrom and McAffee 2013).
Seeking for a solution to the prevailing serial crisis, Parks (2001) points out that librarians may have no incentives to revolutionize academic publishing by moving towards a new business model, only to keep serial costs in line with their budgets. He even argues that some librarians will be motivated to maintain the traditional subscription model to legitimate their employment.
3.1.3 Implications
All of the above shows that the traditional closed access or copyright model is still the dominant business model in place. Basically three big publishers (Elsevier, Springer, Wiley) dominate the market. The OA model plays only a minor role, with some exceptions in disciplines like Biology, Physics and Health Sciences. The reason why researchers still seem to be rather reluctant towards OA is easy to grasp and has its origin in the inherent reward system.20 Pursuing a career in research requires researchers to receive credit in form of publications in as highly ranked journals as possible. Accordingly, the impact factor is an important signal for a researcher in deciding where to publish her works. Especially young researchers tend to be locked-in, as the competitive environment for tenure track does not allow any deviation from the traditional publishing model. Eger et al. (2013) show evidence for a non-linear relationship between age or profession and the disposition towards OA publishing. Thus, the likelihood function for publishing OA reveals an inverted u-shape when the age of a researcher is taken into account.21 Several scholars (Bjoerk 2004; Cavaleri et al. 2009; Megheli and Ramello 2013; Shavell 2010) have drawn the picture of a chicken-egg problem, where newly launched open access journals will be restricted in accumulating reputation and hence in creating a certain level of demand. The problem for an OA journal then is as follows: In order to build up a certain level of reputation the journal will have to assure readers to read the works and authors to submit high quality works. Readers, however, will prefer particularly highly ranked journals to minimize information cost, while authors are forced to publish in highly ranked journals to attract readers. Thus, in a co-existing system of open and non open access journals researchers tend to be locked-in to the weak Nash equilibrium.22
As a consequence, if we believe OA to be superior, we will have to evaluate the instruments that may foster an evolutionary process towards promoting OA publishing. Before we further investigate the costs and benefits of OA and discuss whether an abolishment or other instruments may be a desirable road for shaping the future of academic publishing, we will have a closer look at the OA movement for understanding its principles and development.
3.2 The Open Access Movement
3.2.1 The Open Access Principles
The concept of OA as an initiative or movement to provide free and unrestricted access to scientific knowledge was first to be defined by the Budapest Open Access Initiative (BOAI). Accordingly, open access to scientific works is defined by means of a “free availability on the public internet, permitting any user to read, download, copy, distribute, print, search or link to the full text of these articles, crawl them for indexing, pass them as data to software, or use them for any other lawful purpose, without financial, legal, or technical barriers other than those inseparable from gaining access to the Internet itself.” (BOAI 2002). In this context, OA to scholarly journal literature can be achieved by means of two complementary strategies: (1) Self-Archiving (“the green road”) and (2) OA journals (“the gold road”).23
Self-archiving offers scholars the tools and assistance to deposit their published or unpublished works in so-called archives or repositories.24 Authors may decide to deposit their works to either subject-based and/or institutional repositories. While a subject-based repository bundles research results of specific research fields, institutional repositories provide the option for self-archiving of research output by institution. In practice, several examples for both types of self-archiving exist. Famous for subject-based repositories are ArXiv.org and PubMed. The pre-print server ArXiv.org currently offers open access to more than 860,000 e-prints in physics, mathematics, computer science, quantitative biology, quantitative finance and statistics.25 PubMed comprises over 21 million citations for biomedical literature from MEDLINE, life science journals and online books, covering the fields of biomedicine and health, life sciences, behavioral sciences, chemical sciences and bioengineering.26 In economics, the “Research Papers in Economics” (RePEc) is the world largest collection of working papers, journal articles and software components. The RePEc database currently provides free online access to more than 1.4 million research papers from more than 1,700 journals and 3,700 working paper series. Currently more than 35,000 authors are registered at RePEc.27 In contrast, institutional repositories are mostly run by the libraries belonging to an institution. The creation, location and growth of OA institutional repositories and their contents is indexed by the “Registry of Open Access Repositories” (ROAR).28 More than 2,200 institutional and 250 cross-institutional repositories have been registered in ROAR, where the majority (1,236) is located in Europe. The use of institutional repositories has especially gained momentum since 2004. While first institutional repositories were already set up in the 1990s, the usage of self-archiving jumped up to more than 150 new repositories per year since 2004 and peaking in 2010 with almost 500.29 In this context, the University of Southhampton (UK) and the Lund University (Sweden) abound as the most famous examples for institutional repositories.
The “gold road” to OA adapts the general principles of scholarly publishing by offering a peer-review of submitted papers for quality selection. The OA movement promotes both the launch of a new generation of journals committed to OA and the transition of existing journal titles to OA. OA journals are consistent to the traditional publishing model in so far, as they shall provide with the same publishing services and particularly exercise quality control on submitted papers through an editor, editorial board and/or a peer-review system. However, OA journals do not invoke copyright to restrict access. Instead, copyright and other tools shall ensure immediate and continuous open access to journal content. In this regard, journals generally revert to Creative Commons30 licenses which can also be used to specify usage rights. The “Public Library of Science” (PLoS) and BioMed Central as the leading OA publishers, for example, apply the so-called Creative Commons Attribution License (CCAL). The CCAL allows authors to retain their copyright, but allows anyone to download, reuse, reprint, modify, distribute and/or copy articles from the respective journal. Springer Open Choice,31 as a commercial publisher, uses the Creative Commons Attribution Non-Commercial License, which allows readers to read, copy and distribute a work and to create derivative works for non-commercial purposes. As OA journals are by definition accessible without paying any subscription or access fee, journal publishers have to turn to other forms for covering their expenses. The BOAIlists “many alternative sources of funds for this purpose, including the foundations and governments that fund research, the universities and laboratories that employ researchers, endowments set up by discipline or institution, friends of the cause of open access, profits from the sale of add-ons to the basic texts, funds freed up by the demise or cancelation of journals charging traditional subscription or access fees, or even contributions from the researchers themselves.” (BOAI 2002). As seen before, there are several types or different degrees of OA journals. The “pure” or “true” OA model only considers journals as OA if they follow the lines of the BOAI definition. These “pure” OA journals are listed by the “Directory of Open Access Journals” (DOAJ), currently listing more than 9,900 OA journals (September 2014).32 Much more common are so-called hybrid models, where publishers and libraries apply only a weak form of OA by either providing optional, retrospective, delayed or partial OA. However, Bernius et al. (2009) argue that in contrast to the green and gold roads of OA these hybrid models do not fully satisfy the intended purpose of the OA declarations.
3.2.2 The Open Access Movement: A Brief History
SPARC and the Open Access Movement
The historical origins of the OA movement can be traced back as far as to the 1960s.33 As a first milestone in the history of OA, the literature frequently refers to the launch of the “Educational Resources Information Center” (ERIC)34 in 1966. Before the 1990s, journal articles and working papers were primarily disseminated by use of electronic mailing lists (Laasko et al. 2011, p. 1). Only with the advent of the internet and the spread of digital technologies, the OA model was more frequently adopted in the scientific community. The first free scientific online archive has become known as ArXiv.org at Los Alamos National Library and was launched on August 16, 1991. In 1993, the statistician Gene Glass followed by launching the first peer-reviewed OA journal—the “Education Policy Analysis Archive” (Willinsky 2009, p. 53). Despite these “early shots”, the number of OA journals as well as platforms for self-archiving increased only slowly in the 1990s (Laasko et al. 2011).
It was not until 1998 that the OA movement began to gain momentum, especially due to the efforts of the “Scholarly Publishing and Academic Resource Coalition” (SPARC). In 2001, a group of researchers circulated an open letter to establish “an online public library that would provide the full contents of the published record of research and scholarly discourse in medicine and the life sciences in a freely accessible, fully searchable, interlinked form” (PLoS 2001). The letter was signed by 34,000 scholars around the world and finally led to the launch of the “Public Library of Science” (PLoS). The OA movement gained political weight with a number of public statements. The three most important statements are the “Budapest Open Access Initiative” (2002), the “Bethesda Statement on Open Access Publishing” (2003) and the “Berlin Declaration on OA to Knowledge in the Sciences and Humanities” (2003). The BOAI was launched by the Open Society Institute (OSI) on February 14, 2002. The primary contribution of the BOAI was to eventually provide with a first definition of the OA concept. The initiative gained a broad consensus in the scientific community and was signed by more than 4,000 individuals as well as 365 institutions (Opderbeck 2007, p. 108; Kuhlen 2008, p. 477). The Bethesda statement was released in June 2003 and can be seen as the first step for encouraging faculty and grant recipients of signing institutions to utilize the instruments of OA publishing. Extending on the Bethesda statement, finally, the Berlin declaration of October 22, 2003, established a binding commitment among the Max Planck Institute in Germany and other leading research organizations in Germany, France and Switzerland to encourage researchers and grant recipients of signing institutions to publish their research output by following the OA principles (Opderbeck 2007, pp. 108 et seq.; Schirmbacher 2007, p. 25). Afterwards it was signed by 371 organizations, mostly from Europe, but also including the “Library and Informations Association of South Africa” (since October 6, 2011), the “Indian National Science Academy” (since April 5, 2004) and the “National Natural Science Foundation of China” (since May 24, 2004). The “BBB-definition (Budapest, Bethesda, Berlin) for open access” (Suber 2004) created not only the foundation for all subsequent conferences, but has been put into action by several private grant funders around the world. Consequently, the OA principles were subsequently encoded into public policy and even positive law in Europe, the United States and in the international IPR arena (Opderbeck 2007, pp. 109 et seq.).
Open Access in National Legislation and Public Policy
United States
The United States of America can be seen as the pioneer and natural origin of the OA movement. As a matter of fact, there had been first steps towards an unrestricted and free sharing of research output even before the birth and spread of the internet. With the launch of the “Educational Resources Information Center” (ERIC) by the U.S. Department of Education, the Office of Educational Research and Improvement as well as the National Library of Education, and MEDLINE by the National Library of Medicine in 1966, the US initiated the first OA projects of the world. Both ERIC and MEDLINE are still online. Moreover, the US Department of Defense followed with its launch of the Advanced Research Projects Agency Network (ARPANET)—as the direct ancestor of the internet—primarily the purpose of easing the access and sharing of research output (Suber 2006).35
The implementation of OA policies especially gained momentum throughout the launch and endeavour of two OA advocacy organizations: (1) the Scholarly Publishing and Academic Resource Coalition (SPARC) and (2) Public Knowledge (PK).36 SPARC was founded by Rick Johnson in 1998 and is a coalition of currently nearly 800 institutions in North America, Europe, Japan, China, and Australia.37 Even though its primer purpose was to enhance competition in the market for academic journals, SPARC became an active advocate for OA since the Budapest Open Access Initiative in February 2002. Ever since SPARC has been actively promoting OA publishing, especially by means of campaigns (e.g. the Publisher Assistance Program and the Publisher Partner Program) and guidelines for stakeholders (e.g. the Authors Addendum (for authors), the Directory of Open Access Programs (for librarians and administrators) or the OA Sponsorship Guide (for journal publishers)). Moreover, SPARC initiated the SPARC Open Access Forum which was launched in July 2003 and is moderated by Peter Suber.38 PK was founded in 2001 and speaks for the public interest in information policy (Suber 2006, p. 8). PK initiated an OA project in 2003 that is directed at promoting both the gold and green road by informing policy makers and the public about the critical role of OA publishing inside the US and internationally. Suber (2006) points out that both SPARC and PK had been active in promoting OA publishing before the US congress asked the National Institute of Health (NIH) for OA policies in 2004.
In July 2004, the US congress instructed the NIH to develop an OA policy that would require NIH grantees to deposit a copy of their (NIH funded) works on PubMed Central (PMC) 6 months after their publication in a peer-reviewed journal. However, the final version of May 2005 fell short of this objective in two respects. First, the requirement was substituted by a request. Second, the permissible delay was extended to 12 months after the publication date (Suber 2006, p. 10). In NIH (2005) the policy “requests and strongly encourages that authors specify posting of their final manuscripts for public accessibility as soon as possible (and within 12 months of the publisher’s official date of final publication)” (NIH 2005). Despite the weaknesses of the NIH policy, the NIH was the first agency to actively ask for OA archiving of their funded research results. Suber (2006) highlights that the NIH was also a good agency to do this first step, since “the NIH is the world’s largest funder of medical research, and its 2005 budget, at $28 billion, was larger than the gross domestic product of 142 nations. The NIH policy simply applies to more literature than any other single initiative is ever likely to cover – about 5,500 peer reviewed journal articles per month.” (Suber 2006, p. 10). Subsequent discussions—also influenced by the American Center for Cures Act (introduced in the US Senate in 2005) and the Federal Research Public Access Act (introduced to the US Senate in May 2006)—induced several revisions of the original NIH policy of May 2005. In January 2008 the NIH released a revised OA policy, now making OA archiving of NIH funded research mandatory. The revision “shall require that all investigators funded by the NIH submit or have submitted for them to the National Library of Medicine’s PubMed Central an electronic version of their final, peer-reviewed manuscripts upon acceptance for publication, to be made publicly available no later than 12 months after the official date of publication” (NIH 2008). More revisions or add ons of the NIH policy followed in 2008 (7 March and 23 September), 2009 (19 March, 12 August and 30 October), on 16 November 2012 and on 9 January 2013.
Europe
In Europe the history of OA policy developments initially began to gain momentum with the Berlin declaration. The first country to encode the OA principles was the United Kingdom (UK). In 2004, the UK Parliament’s “Select Committee on Science and Technology” (SCST) issued a report that promotes the implementation of an OA publishing regime (Opderbeck 2007, p. 109). The report lists 83 conclusions and recommendations that greatly highlight the effectiveness of the OA model. In particular, the committee suggests that the Research Councils should require authors to deposit copies of publicly funded research in institutional repositories and that funds should be made available to cover publication costs. Furthermore, the committee recognizes the need for international coordination and recommends that the UK Government should “act as a proponent for change on the international stage and lead by example” (SCST 2004, p. 97). In 2005, one of the greatest research funders—the Research Councils UK (RCUK)—issued a mandate that required authors who received funding by the Research Councils to deposit their works in an OA archive. In October 2006, also the Wellcome Trust—a UK-based charitable foundation and one of the world’s largest research funders—implemented an OA policy that requires “electronic copies of any research papers that have been accepted for publication in a peer-reviewed journal, and are supported in whole or in part by Wellcome Trust funding, to be made available through PubMed Central (PMC) and UK PubMed Central (UKPMC) as soon as possible and in any event within 6 months of the journal publisher’s official date of final publication” (Wellcome Trust, Open Access Policy).39 In Germany, similar OA initiatives were issued after the Berlin declaration, particularly driven by support from the “German National Scholarship Foundation” (Deutsche Forschungsgemeinschaft, DFG). In January 2006, the DFG adopted a policy in which it requires that DFG funded research results should be made available digitally and on the Internet by means of OA.40 The implementation process was fostered by the “Alliance of German Science Organisations”. The members, e.g. the “Max Planck Gesellschaft” (MPG), committed to implement a policy to require research faculty members to deposit a copy of all their published works in an OA repository and encouraged their researchers to publish their research results in OA journals wherever a suitable OA journal exists. In reaching this goal, e.g. the MPG and the Humboldt University of Berlin created a document and publication server (edoc) to offer the necessary organisational and technical framework to all staff members for providing OA to their works.41
The EU Commission started to take considerable action in implementing OA policies for academic publishing in 2006. The efforts at the European level followed a report by the EU Commission in January 2006, in which the commission highlights the need to promote the OA model for taking action against anticompetitive bundling practices by journal publishers. In particular, the report suggests to (i) establish a European policy mandating published articles arising from EC-funded research to be available after a given time period in open access archives, and (ii) explore with Member States and with European research and academic associations whether and how such policies and open repositories could be implemented (ECReport 2006, p. 11). The developments at the European level especially began to gain momentum in 2007 and 2008, when the EU Commission launched two OA initiatives.42 In December 2007, the ERC Scientific Council published its “Guidelines for Open Access” as a follow up on the EC Report. In August 2008, the EC Commission complemented this initiative with the launch of the “Open Access Pilot in FP7”. Both initiatives followed the statement made in the EC Report to require that researchers provide OA to every article that has been the result of EC funded research.43 In the following years the OA principles were implemented in all EU member states.44 In this context, especially the DRIVER project45 helped to establish and create OA repositories in each of the member states and stimulated OA archiving by promoting policy developments on the national level. In addition, several initiatives—like the OpenAIRE initiative in cooperation with stakeholders like SPARC Europe, COAR or LIBER—supported the further expansion of OA repositories and created awareness on the various OA possibilities among researchers and research faculties. Overall, the various OA initiatives in Europe show a strong tradition of OA models in academic publishing and highlight the efforts in establishing both an infrastructure of OA repositories as well as OA journals.
International IPR Policy
At the international level it was primarily through the efforts of the “Organization for Economic Co-operation and Development” (OECD) that the OA principles became recognized. In 2004, in total 35 OECD member states, including Germany, France, the UK and the US signed the “Declaration on Access to Research Data from Public Funding”46 in which they declared their commitment to work towards the establishment of an OA regime for digital research data from publicly funded research activities. A first move towards encoding the OA principle into the international policy making framework was the “Access to Knowledge” (A2K)47 treaty in 2004. The draft treaty emerged throughout a call from Brazil and Argentina for a development agenda for the WIPO and was primarily intended to ease the transfer of knowledge to the developing world (Opderbeck 2007, pp. 113 et seq.). During a series of meetings in 2005 a draft of the treaty was prepared by representatives from developing countries as well as representatives from the UK and the US, including a catalogue of exceptions to copyright that essentially mirror the “fair use” or “fair dealing” concepts. In particular, the “A2K” treaty would generally limit copyright law akin to existing compulsory licensing provisions (Opderbeck 2007, p. 115; Helberger 2005). A concept that is already included in the “Agreement on Trade Related Aspects of Intellectual Property Rights” (TRIPS),48 but has seldom been used in the past.49 Furthermore, the “A2K” draft treaty includes sections on the limitation of digital right management (DRM) systems (article 3-6), copyright term extension (article 3-9) and compulsory licensing of copyrighted works in developing countries (article 3-12). Part 5 of the draft further specifies ways for “expanding and enhancing the knowledge commons”. Accordingly, any work “resulting from government-funded research shall be publicly available at no charge within a reasonable time frame, subject to reasonable exceptions, for example, for classified military research, for patentable discoveries, and for works that generate revenue for the author, such as books.” (article 5.2(a)). In addition, a knowledge commons committee (KCC) shall “promote cooperation and investment in databases, open access journals and other open knowledge projects that expand the knowledge commons.” (article 5-1). Within two meetings in February and June 2006, finally, the committee chair proposed to move forward on proposals that had received consensus support. This proposal was rejected by developing countries that claimed significant IPR reforms. In fact, the proposal followed primarily interests of the US and the European states which was considered as a back-room maneuver by developing countries (Opderbeck 2007, pp. 116 et seq.). Consequently, the draft of the proposed treaty has ever since been in debate.50
Especially with respect to developing countries the OA movement has still affected the international policy through four programmes run under the patronage of “Research4Life” and initiated by the “World Health Organization” (WHO), the “Food and Agriculture Organization” (FAO), the “United Nations Environment Programme” (UNEP) and the “World Intellectual Property Organization” (WIPO). The “Research4Life” programmes generally seek to provide free or reduced fee access to research in health, agriculture and the environment for all eligible countries in the developing world. Institutions in eligible countries are universities, colleges, research institutes, professional schools, extension centres, government offices, local non-governmental organizations (NGOs), hospitals and national libraries. The history of “Research4Life” started in 2002 when the WHO, the Yale University and six major publishers51 launched the “InterNetwork Access to Research Initiative” (HINARI),52 providing access to peer-reviewed journals covering medicine, nursing and related health and social sciences. In October 2003 a similar model was employed for research in food and agriculture with the “United Nation’s “Access to Global Online Research in Agriculture” (AGORA),53 which was initiated by the FAO of the United Nations and nine founding publishers.54 Accordingly, AGORA provides with free or low-cost access to peer-reviewed journals in agriculture and related fields to public institutions in developing countries. The “Research4Life” initiative was finally supplemented by the “Online Access to Research in the Environment” (OARE)55 initiative in 2006, where OARE is directed at providing access to scholarly literature in the area of environmental research. OARE was launched by UNEP, the Yale University and leading science and technology publishers.56 Only recently, the WIPO together with its partners in the publishing industry57 launched the “Access to Research for Development and Innovation” (ARDI)58 initiative. Starting in 2009 and joining “Research4Life” on August 23, 2011, this initiative has been providing free online access to major scientific and technical journals to local, not-for-profit institutions in least-developed countries and low-cost access to industrial property offices in developing countries across the world. We will elaborate on these initiatives in Sect. 4.2 by analyzing the impact of OA as a means to assist developing countries in bridging the knowledge gap and to involve all nations in science. Before, we will have a closer look at the attempts to actually create OA journals in the academic publishing market.
3.2.3 The Rise of Open Access Journals: Some Descriptive Statistics
The development of OA journals can be investigated by using data from the “Directory of Open Access Journals” (DOAJ), listing only journals that follow the lines of the BOAI definition of OA.59 Accordingly, the DOAJ defines OA journals as journals that are freely available via the internet and provide the reader with the right to read, download, copy, distribute, print, search, or link to the full texts of articles.60 By the time of writing this section in January 2013, the DOAJ listed more than 8,600 OA journals, where more than 3,700 journals are searchable at article level and almost 776,000 articles are made freely available. The growth rate of DOAJ has been steadily increasing and shows that on average more than three titles were added to DOAJ per day in 2011. Figure 3.8 shows the development of the number of OA journals added to DOAJ from 2002 until 2012.61 The pillars refer to the number of newly launched OA journals by year. The line graph shows the development of the aggregated number of OA journals over time, i.e. the added number of all journals minus the number of eliminated titles in the respective year. The pillars revert to the primary axis, while the line graph reverts to the numbers as displayed on the secondary axis.
Fig. 3.8
Development of OA journals: 2002–2012
Obviously, the total number of OA journals increased vastly from around 33 in 2002 to almost 8,500 journals by the end of 2012. While there was a steady increase of journals added to DOAJ by more than 400 on average in the period between 2003 and 2007, the number of newly launched OA journals almost doubled for the years 2008 and 2009, reaching on average almost 800 newly launched OA journals per year. The number doubled again for the period after 2009, reaching an absolute yearly growth rate of more than 1,400 OA journals, peaking in 2011 with 1,538 newly launched OA journals. The number of launched newly OA journals by the time of data extraction was 1,005 for the year 2012.62
Interesting in this respect is the development of OA journals by country.63 Figure 3.9 clearly shows that the United States has been most active with 1,211 OA journals listed in the DOAJ by October 2012. Among the pioneers of OA publishing in 2002 were the United States with 19 journals (57 %), the United Kingdom with 5 journals (15 %) and Germany with 3 journals (9 %).64 However, with Brazil and India also two developing countries belong to the top 5 of the most active countries in launching OA journals. Accordingly, in 2012 Brazil and India ran a total number of 801 and 463 OA journals, respectively. Brazil and India are closely followed by Egypt on rank 6 with 350 journals in 2012. The percentage distribution of OA journal by country has changed considerably from 2002 to 2012. In this regard, the dominance of the USA has been steadily decreasing over the last years, with currently about 14 % of all OA journals published by US publishers. Nevertheless, it becomes evident that there are regions, especially in the developing world, that hardly contribute to the pool of OA journals. Here, the Sub-Saharan African countries are far from publishing 1 % of all OA journals in present.65
Fig. 3.9
OA journals by country: from 2002 until 2012 (Source: DOAJ)
Also striking is that the disposition to the idea of OA publishing differs considerably between different fields of research. In this context, Health Sciences is most active in OA publishing with 2,011 or 24.3 % of all OA journals. This is followed by the fields of social sciences with 1,471 (17.8 %), technology and engineering with 771 (9.3 %) and biology and life sciences with 629 (7.6 %) OA journals. As such, these four subjects combine almost 60 % of the total amount of OA journals published until 2012.66 In business and economics in total 418 journals are published that follow the definition as stated by the DOAJ, with 241 journals assigned to business management subjects and 177 journals in economics. Consequently, this provides evidence for the fact that not all disciplines pay equal attention to the OA model. A standardized GINI-coefficient of 0.502 (where 
) endorses this impression of a fairly high concentration rate in the distribution of OA journals by discipline. Figure 3.10 summarizes the facts on the distribution of OA journals by discipline, where n i = absolute number of journals in discipline i, q i = relative number of journals in discipline i, v i = cumulative percentage of journals (i.e. 
), h i = percentage distribution over disciplines and H i = cumulative percentage of the distribution over discipline (i.e. 
).
) endorses this impression of a fairly high concentration rate in the distribution of OA journals by discipline. Figure 3.10 summarizes the facts on the distribution of OA journals by discipline, where n i = absolute number of journals in discipline i, q i = relative number of journals in discipline i, v i = cumulative percentage of journals (i.e. ), h i = percentage distribution over disciplines and H i = cumulative percentage of the distribution over discipline (i.e. ).Fig. 3.10
Distribution of OA journals by discipline: GINI coefficient
As a shift towards the OA model comes along with a change from the “reader pays” principle to the “author pays” principle, it may also be interesting to see how the publishers of OA journals have coped with this challenge. In October 2012, in total 2,335 OA journals (28.24 %) raised author fees for accepted journal articles. The majority of 5,510 (66.63 %) had no author fees.67 An amount of 210 (2.54 %) had conditional author fees. No information was provided for 214 (2.59 %) OA journals. In this respect, Chang (2003) shows that open access publishers have to consider different income sources to recoup their first copy costs. First and foremost, public grants, endowments and subscriptions to print versions emerge as promising options in addition to a charge of a publication fee per article. Figure 3.11 provides an overview on the largest OA publishers and their income sources.
Obviously, there is evidence for additional income sources than charging an authors’ fee per article. In particular, the Public Library of Science (PLoS)68 received a $9 million start-up grant from the Gordon and Betty Moore Foundation as well as financial support by other grantors.69 The Hindawi Publishing Cooperation70—a commercial publishers of more than 300 OA journals—charges in addition to a publication fee per article—amounting up to $1,500—from the author(s) of a work also print subscriptions which range from $195 to $1,895 (2012 rates).71 In contrast, Medkow Publications72 does not charge authors for paper submissions, but realizes income streams by means of print and online advertisements as well as reprint purchases for distribution.
All of the above reveals the increasing relevance and prevalence of OA publishing in the market for science, with currently more than 9,900 OA journals listed by the DOAJ (numbers of September 2014). Meanwhile, all types of market actors—commercial, non-commercial and scientific societies73—have been engaged in the OA movement. While some publishers still only provide with hybrid models of OA, which do not fully satisfy the purpose of intended OA declarations, also a large number of “true” OA examples abound in practice. Especially in recent times several new steps have been taken for promoting and strengthening the OA principles in academic publishing. The following section shall highlight some of these recent developments.
3.2.4 Recent Developments
There have been several new developments in the last view years. As a review of all recent initiatives, policies and movements is way beyond the scope of this subsection, we would like to elaborate on two general perspectives and some developments that have gained most attention in the public debate.74
First, there have been recent steps by policymakers in setting a legal framework and introducing policies for the promotion of OA publishing of publicly funded research. In this regard, the new OA policy by the “Research Councils UK (RCUK)” has gained most attention. In its new OA policy from 16 July 2012 the RCUK follows the recommendations of the “Working Group on Expanding Access to Published Research Findings” (the Finch Report). Most importantly, the report recommended a clear policy direction in the UK towards fostering OA journal publishing. The OA policy issued on 16 July 2012 clearly states that all academic works that “result from research that is wholly or partially funded by the RCUK (1) must be published in journals which are compliant with Research Council policy on Open Access, and; (2) must include details of the funding that supported the research, and a statement on how the underlying research materials such as data, samples or models can be accessed.”(RCUK 2012b, p. 1, paragraph 3). Regarding the first section, the paragraph 4 in RCUK (2012b) specifies the compliance with journals. Accordingly, a journal is compliant with the RCUK OA policy if (1) the journal itself provides immediate and unrestricted access to the final paper version via the journal’s website; or if (2) the journal allows to deposit a final version of the paper (including all changes that result from peer review) in other repositories, without any restriction on non-commercial re-use and within an embargo period of 12 months. The payment of an “Article Processing Charge (APC)” is ensured through block grants to universities and eligible research organizations.75 The receipt of funding is further specified in paragraph 5 of RCUK (2012b). Paragraph 6, finally, specifies the rules on implementation and compliance. Thus, the policy applies to all RCUK funded research papers that are submitted for publication from 1 April 2013. In a workshop on 13 November 2012 the Research Councils further discussed and agreed upon rules for the monitoring of the OA policy.76 In Germany, the recent developments show a two sided approach towards OA publishing. Besides the implementation of OA policies from funding agencies and the “alliance of science organizations”, a discussion on the implementation of a so-called inalienable right of secondary publication for authors of academic works has been induced at the governmental level. On 10 April 2013 the German government introduced a bill for a reform of §38 UrhG, covering not only the implementation of an authors right for secondary publication, but also issues related to the orphan works problem.77 The bill shall provide the author of an academic work with the right to re-use her publication for non-commercial purposes after an embargo period of 12 months after publication. In particular, the new bill shall enable the researcher to deposit a copy of her publication on a repository. On 3 May 2013, the German Federal Council (Bundesrat) even strengthened the position of academic authors by reducing the embargo period from 12 months to only 6 months after publication.78 The bill is still pending (as of July 2013), but has already released a storm of controversy not only among publishers but also within the group of researchers themselves.79