3.1. Global Open Access Repositories

Since the development of DSpace and E-Prints in 2002, two major pieces of repository software, the construction of repositories has progressed in earnest. DSpace was jointly developed by the Massachusetts Institute of Technology and the HP Research Institute in the US and it quickly became known to the public because Cornell University utilized it to create its own repository. E-prints, developed by Southampton University in the UK, contributed substantially to the establishment and stabilization of the repository at Oxford University. Subsequently, the number of universities and research institutes worldwide that are developing repositories has steadily increased.

Fig. 1 shows the repository growth worldwide. The total number of repositories in OpenDOAR showed a steady increase (except in the first year) from 128 in December 2005 to 3,502 in March 2018. While there are slight differences in magnitude among regions, this increment was consistent across them. The increase can be attributed to growing awareness of open access.

Fig. 1.

Overall growth of repositories in OpenDOAR from December 2005 to March 2018.

Fig. 2 shows the repository statistics by regions as of March 2018. Europe had the highest number of repositories, at 1,162 (46% of the total), followed by Asia (702, 20%), North America (615, 18%), and South America (309, 9%). Asia—centred on Japan, India, Turkey, Indonesia, Taiwan, and China—is showing rapid growth in the number of repositories, to the point where the number recently surpassed that for North America. Therefore Asia, along with Europe, is becoming a centre of global open access repositories.

Fig. 2.

Repository statistics by regions as of March 2018.

When examining repository type (Fig. 3), most repositories were classified as institutional repositories (accounting for 86% of the total), followed by disciplinary repositories (at only 9% of the total). The proportion of institutional repositories is slowly increasing, indicating that recognition of open access is spreading and the number of requests for establishing institutional repositories is growing.

Fig. 3.

Repository statistics by types as of March 2018.

Table 1 compares the repositories of representative institutions and countries that are obliged to deposit public research results. The National Institutes of Health (NIH) in the United States (Organisation for Economic Co-Operation and Development, 2015) is in charge of depositing and utilizing research papers produced by the NIH fund in line with the national public deposit policy, while the Spanish Foundation for Science and Technology (Fundación Española para la Ciencia y la Tecnología, FECYT) operates a national repository called the ‘Recolector de Ciencia Abierta’ (RECOLECTA), based on a connection with the Network of Spanish University Libraries (Red de Biblioteca Universitarias Españolas). The Chinese Academy of Sciences (CAS, 2014) operates a repository called the ‘CAS Institutional Repositories Grid’ (CAS IR Grid) that comprehensively deposits and manages papers produced with CAS funds, linking them to the repositories of CAS-affiliated institutions.

Table 1.

Comparison of repositories of representative institutions and countries that are obliged to deposit public research results

RECOLECTA, Recolector de Ciencia Abierta; CAS IR Grid, Chinese Academy of Sciences Institutional Repositories Grid; NIH, National Institutes of Health; FECYT, Fundación Española para la Ciencia y la Tecnología [Spanish Foundation for Science and Technology]; REBIUN, Red de Biblioteca Universitarias Españolas [Network of Spanish University Libraries]; CAS, Chinese Academy of Sciences; R&D, research and development.

In February 2000, PubMed Central (PMC) built an open repository managed by the National Library of Medicine (NLM) which collects and stores papers published in biomedical and life science journals according to the NLM’s legislative mandate for collecting and keeping biomedical papers. The academic journals fully participating in the PMC submit their papers to the PMC directly, and papers that are supported by NIH funds are directly deposited by the paper’s author(s). In addition to its role as a repository, the PMC makes it possible to store and cross-reference data from various sources using a common format. Using the PMC, it is possible to find all related materials by quickly searching the entire collection of full-text documents. The PMC also integrates literatures from different fields in order to improve the research and knowledge of experts such as scientists and clinicians. As of March 2018, about 4.7 million articles from approximately 7,000 journals are retained in the PMC, and the number of fully participating journals is 2,098.

In 2007, the Spanish government encouraged the establishment of an open access repository, announcing the ‘Draft of the National Law of Science.’ This law included a regulation whereby researchers who received public funds had to make their research results open access within six months. Article 37 of Spanish Law 14/2011 on Science, Technology, and Innovation (named ‘Open Access Dissemination’) established a national standard stipulating that the outcomes of research activities supported by the state must be deposited in open repositories. Furthermore, since 2007, Spain’s FECYT and Red de Biblioteca Universitarias Españolas have sought to build a national infrastructure for open repositories; accordingly, through steady collaboration, these two organizations conceived RECOLECTA, an open platform that links all institutional open repositories in Spain and provides services for repository managers, researchers, and decision makers. RECOLECTA has promoted and coordinated a national infrastructure for interoperable digital science repositories utilizing standards adopted by communities worldwide and was designed to promote research development and the adoption of open policy. Specifically, the RECOLECTA provides easy, free access to all scientific research outcomes stored in Spain’s repositories, as well as seeking to build, maintain, support, and improve the national repository infrastructure. Specifically, it provides users with support services, enhances the national open community, and offers statistical data on repositories.

The CAS is a core pioneering organization in the field of Chinese technology and natural sciences, consisting of a comprehensive R&D network, higher education system, and outcome-based academic society. In China, open access began in 2003 through participation of Chinese scholars in open access and academic publishing seminars. Open access only became standardized the following year when the CAS and National Natural Science Foundation of China signed the Berlin Declaration. Since then, China has been constantly working on open-access-related activities, such as establishing a CAS institutional repository system (CAS IR Grid) on a trial basis in 2007 and opening access to China’s information portal in 2008. The CAS IR Grid contains 114 institutional repositories as of March 2018. When a researcher deposits his or her paper into a CAS-affiliated institutional repository, the paper becomes available in the CAS IR Grid. If an institution does not yet have a repository, papers must be deposited in the repository operated by the National Science Library of the CAS. Since 2012, the annual number of papers registered in the CAS IR Grid has ranged from 40,000 to 1,500,000, of which more than 70% have the original text available. As of March 2018, there are about 820,000 registered papers, of which about 75% provide the original text.

Japan began its Cyber Science Infrastructure Program in 2005. Under this program, institutions (including universities) have begun devoting some effort to establishing their own repositories. In fact, by 2010 nearly 200 institutions across Japan had established an institutional repository. Research institutions, including universities, independently established these repositories using software such as DSpace. However, many institutions, while expressing a willingness to build a repository, found it difficult to afford or hesitated because of the expected burden of operating the repository after establishing it.

Based on past experiences of establishing a repository, the National Institute of Informatics of Japan promoted the introduction of the JAIRO Cloud in 2011 in order to promote wider development of repositories. The JAIRO Cloud, established in 2012, is a computing service based on an SaaS system that was created by the National Institute of Informatics to improve the operation of institutional repositories. Initially it targeted universities without an institutional repository, but since January 2014 it has begun accepting institutions with an existing repository. In May of that same year, starting with the transition of Tulips-R (the institutional repository of the University of Tsukuba), established institutional repositories throughout Japan began transferring to JAIRO Cloud. Fig. 4 shows the state of establishment of institutional repositories in Japan, and it is evident that rapid growth has been achieved as a result of introduction of the JAIRO Cloud, with many institutions transferring from their own repository to the JAIRO Cloud.

Fig. 4.

Number of institutional repositories in Japan as of August 2017. IR, institutional repository.

3.2. Korean Open Repositories

The establishment of institutional repositories in South Korea began with the Korea Advanced Institute of Science and Technology (KAIST) Open Access Self-Archiving System (KOASAS). In 2007, KAIST allocated some of its own budget to develop and operate a repository for managing, preserving, and distributing research outcomes obtained by the professors and researchers of the university. KOASAS utilizes the same model operated by the libraries. Later, in February 2012, KAIST made active use of its institutional repository by establishing the Researcher Information Management System, a performance evaluation system of researchers in KAIST, and connected it with KOASAS. As KOASAS holds more than 200,000 academic articles, including papers published in 2018, it is a valuable resource for researchers in South Korea and abroad.

The central library of Seoul National University officially launched its institutional repository S-Space in December 2008.2 This repository was developed by benchmarking with DSpace and KOASAS, upgrading and customizing for the convenience of its members. More than 98,000 materials have been registered in S-Space as of March 2018, including research papers published in academic journals, papers presented at academic conferences, and dissertations issued by various academic societies and institutions affiliated with Seoul National University. In 2017 only, there were over 6 million downloads.

The full-scale implementation of a domestic open access repository was the OAK Project. The OAK Project, which was promoted by KISTI in March 2009, sought to build knowledge cooperation to promote open access for domestic academic information by adhering to the following steps: repository development and dissemination, open access journal publication support, open access portal (OAK Central) establishment, and open access governance system establishment. Following the replacement of the host organization by the National Library of Korea in 2014, the development and dissemination of the Korean OAK repository began.

The OAK repository was built using DSpace, and customized to the domestic environment. It was distributed through the help of an OAK repository operation consultative group, consisting of KISTI, repository system developers (KISTI’s partners), and the OAK repository operation organization. This consultative group selects target institutions for new repository establishment through a public contest and shares the trends and know-how in operation through training and seminars with the selected institutions. About five institutions are selected annually based on their applications for establishing the repository; as of March 2018, a total of 38 OAK repositories have been established. KISTI’s partners played a role in spreading OAK repositories to various institutions that wished to install it. Furthermore, the OAK repository is continually updated, in accordance with updates to DSpace.

In addition, institutions that installed the OAK repository identified new requirements through operation of the repository, thus helping the consultative group to improve the repository and its operation methods. The contents of all OAK repositories can be retrieved through integrated search services by both domestic and foreign users through the OAK portal, which is operated by the National Library of Korea. Interested users can access the original texts of content via the repository portal. This has helped increase web traffic to the OAK repository.

However, not all institutional repositories in South Korea are smoothly operated. While there are about 24 unique repositories that hold academic papers, only a few—such as KOASAS and S-Space—are actively operating. Considering the number of domestic universities and public institutions, this figure indicates exceedingly poor performance compared with Europe and Japan.

Open access is being pushed in various directions all over the world. For example, gold open access is being implemented through such policies as the OA2020 (led by the Max Planck Digital Library), SCOAP3 (centred on European Council for Nuclear Research), and the Big Deal models of various European countries, whereas green open access is being implemented through the establishment of open access repositories. Although South Korea is making a considerable effort to keep up with this trend, its achievements are comparatively limited because of problems such as peculiar characteristics in the domestic academic ecosystem, limited participating institutions, lack of government policy support, and low awareness among researchers of open access (Hwang, 2017).

As part of an effort to overcome this problem, the OAK Project is seeking to promote collaboration among institutions in the development of repositories through OAK Central. However, there is still a need to build a national repository such as Japan’s JAIRO Cloud and Spain’s RECOLECTA. The current situation in South Korea is similar to that in Japan before the introduction of JAIRO Cloud. In particular, while a number of institutions have established an institutional repository, some are in name only, as the institutions are incapable of maintaining their operation. It is therefore necessary to implement a nationally integrated repository, as well as to build up personalized institutional repositories for institutions which desire to build repositories, but lack the capability as well as the necessary technology to do so.

4.1. Model Outline

It is difficult to calculate the potential impact of implementing an open access repository, and doing so can cause a considerable degree of controversy. Nevertheless, to assess potential impacts and use them for future reference, Houghton developed a model using the Solow-Swan model (for further detail, refer to Houghton et al., 2009).

The basic Solow-Swan model (Solow, 1957) is represented in the following production function:

$Y=A^{\eta }{K}^{\beta }{L}^{\alpha }$

where A is an index of technology, K is the capital stock, and L is the supply of labour. Both K and L are taken to be fully employed by virtue of the competitive markets assumption. Solow further developed this model, proposing that once we exclude the impacts of capital and labour, what is left is the impact of technology. He subsequently studied the impact of technological development on overall production. He also applied the model to estimate the rate of return to R&D.

This model is based on several major assumptions. The first assumption is that all R&D creates useful knowledge in economic or social terms (the efficiency of R&D). The second assumption is that all created knowledge is equally accessible to anyone who wants to use it for productive activities (accessibility of knowledge).

However, in the real world there are numerous barriers or limitations to accessing and utilizing knowledge. Based on this, Houghton (2009) demonstrated that it is possible to calculate the impact on return to R&D by improving the accessibility and efficiency of knowledge and reducing friction. In this modified Solow-Swan model, accessibility and efficiency are considered ‘friction variables.’ He proposed the following formula:

$\frac{\partial y}{\partial R}={\rm Y}\frac{Y}{R}(1+{\delta }_{\Phi })(1+{\delta }_{ѕ})$

Where δ_Φ (1 + δ_ѕ) is the percentage change in efficiency (accessibility), Y represents the contribution ratio of the rate of growth of R&D knowledge stock to output growth as a factor of production (i.e., the elasticity), and R indicates the stock of R&D knowledge, which can be calculated as follows:

$R_t=(1-\delta )R_{t-1}+R\&D_{t-1}$

where δ is the rate of obsolescence of the knowledge stock.

4.2. Operationalizing the Model

The main parameters for applying the modified Solow-Swan model are rate of return to R&D, accessibility, and efficiency. Research on the economic impact of R&D at the firm, industry, and national levels has been increasing. However, the claimed variation in the rate of return to R&D differs widely among researchers. For example, Salter and Martin (2001) found that the rates ranged from 10% to 150%. Hall, Mairesse, and Mohnen (2010) found similar degrees of variation depending on the researcher and analysis level. When all results of these studies are combined, a conservative estimate puts the rate of return to R&D at between 10% and 20%.

Accessibility can be defined as the proportion of the stock of knowledge generated by R&D accessible to those who would use it productively. Houghton et al. (2010) suggested measuring the increment in accessibility by combining the degree of access to desired academic information (access gaps), the degree at which academic information was cited (citation), and the degree of variation at which academic information was downloaded. Although the degree differs according to the characteristics of the repository to be built, the results of existing studies suggest that accessibility can be increased by as much as 4.5% (as a conservative estimate).

Efficiency can be defined as the proportion of R&D spending that generates useful knowledge; it can have a number of dimensions relating to wasteful, inefficient, and/or poorly directed research expenditures. Houghton et al. (2010) suggested using scenario-based measurement tools for efficiency, such as wasteful expenditure, number of new opportunities, and time saving for research.

Various other parameters must also be defined. First, a project to establish a research repository can be considered a kind of ‘information system business.’ Considering the lifecycle of a system in South Korea, the analysis period of the main information system is generally four to seven years. Therefore, we conducted a study spanning five years, which falls in the middle of this range.

There is a time lag between research spending and the social and economic impact of research results. In some fields, this lag can range from 2 to 30 years or more, whereas in others, the lag is no more than 1 to 2 years. According to Mansfield (1991, 1998), the average lag in US firms between publication of academic research and the timing of a related commercial innovation was around 7 years (which fell to 6.2 in the later study). Adding the time for publication, the lag was about 10 years, but it can be assumed that the time was shortened when considering the difference from the time when the research results were announced. Accordingly, this study assumed a lag time of 7 years.

Since the cost and benefits of a business manifest over a long period of time, it is necessary to compare them by converting all the costs and benefits that will occur in the future to their present value. This conversion process means discounting the current value, and the interest rate applied at this time is called the discount rate. Since the task of estimating the appropriate social discount rate is exceedingly complicated, domestic studies in general use the 5.5% social discount rate presented by the Korea Development Institute.

Fig. 5 shows the results of estimating domestic R&D expenditure from 2016 to 2024 via a regression analysis, drawing on information of domestic R&D expenditure from 2010 to 2015. While the structure of the R&D expenditure is decided by policy, its coefficient of determination (adjusted R²) was nevertheless very high; thus, it can be considered a valuable estimate. The results indicated the estimated R&D expenditure in 2019 and 2024 would be 22.9 trillion won and 28 trillion won, respectively, with an estimated annual growth rate of 4.4%.

Fig. 5.

Scale and prediction of research and development expenditure.

Table 2 shows the estimates of the impacts of a national open access repository. For illustrative purposes, we expanded the range of rate of returns on R&D from 10% to 50% so that it could be applied to various scenarios. We also examined the increases in accessibility and efficiency by 1% to 10%.

Table 2.

Estimates of the impact of investment in a national open repository

R&D, research and development.

With a 20% return to R&D expenditure on 22.9 trillion won in 2019, an increase of about 1% in accessibility and efficiency yields a return to R&D of about 63 billion won. This is a discounted amount based on 2019, taking into account the 7 years of time-lag between expenditure and impact. Overall, it is evident that the increase in R&D expenditure leads to an increase in impact. The increasing rate of return on R&D is beyond the rate of increase in R&D expenditure. The increase in accessibility and efficiency also appears to have a strong influence on the impact.

Table 2 shows the increasing impacts over the years, as well as the result of converting them into the present value (2018) for basic economic analysis. As mentioned above, this refers to the value calculated for every 5 years based on the economic lifecycle; it could be much more effective if the lifecycle were longer than 5 years. Therefore, the values presented in Table 2 show the social and economic impacts of the establishment of a national open repository. This provides a guideline for investment in the establishment of a national open access repository. When focusing on the most conservative situation, if the rate of return on R&D is 10% and the rate of increase in accessibility and efficiency is 1%, there is still an impact of about 147 billion won. By contrast, in the moderate situation (30% rate of return to R&D, 5% increase in accessibility and efficiency), the impact is over 2.2 trillion won.

In this study, we exclude the costs of establishing and operating the national open repository. This is because these would differ considerably according to the architecture and scope of application of the to-be-established system, and estimating without reliable information on system design is foolhardy at best. Nevertheless, the information in Table 3 provides a rough guideline for national open repository investment.

Table 3.

NPV of estimates of the impact of investment in a national open repository

NPV, net present value; R&D, research and development.